AE-11 Spring Design Manual Prepared under the auspices of the SAE Spring Committee Published by: Society of Automotive Engineers, Inc. < 400 Commonwealth Drive Warrendale, PA 15096-0001 sanao) Preface It was in 1678 that an English scientist, Robert Hooke, stated that, within certain limits, deflection is proportional toload. His work on flexible members is still the basis for spring design today. The wide use of springs to store and release energy can easily be observed by looking at the tools, appliances, and vehicles, that we use in our daily lives. While there are several basic types of spring forms, the variations within each ‘category are endless. The processes which are used to manufacture these items may also vary, further ‘compounding and enlarging the subject. In 1943, at the behest of the U.S. Ordinance Department, the first spring manual was published under the itle “Manual on Design and Application of Helical Springs for Ordinance.” The manual found wide distribution, and revised editions were issued in 1945, 1958, 1962, 1973, 1980, and 1989. In 1944 the manual “Design and Application of Leaf Springs” was published with revised editions being issued in 1962, 1970, and 1982. The manual “Design and Manufacture of Torsion Bar Springs” was first published in 1947 and revisions were made in 1966 and 1981. The manual “Design and Manufacture of Coned Disk Springs and Spring Washers” was first published in 1950 with revisions in 1955 and 1989. The newest manual “Incorporating Pneumatic Springsin Vehicle Suspension Designs” was published in 1989. The publication of this manual on pneumatic springs, along with all the recent revisions to the other manuals, reflect the addition of new technologies and industry practices for automotive applications. ‘The incorporation ofthese five manuals into one volume represents the most comprehensive reference work available today. The Spring Committee wishes to note that these manuals should not be regarded as a compilation of design or manufacturing specifications; instead, they should be considered as reference ‘works which contain essential information which may be helpful to the engineer and designer on a broad range of topics — material selection, tolerances, end configurations, fatigue life, load and tress calculations, and processing information. All of the manuals employ SI Units in accordance with SAE 916, “Rules For Use of SI (Metric) Units.” ‘The Spring Committee recognizes the generous measure of time, effort, and dedication which the respective Subcommittees put forth in the preparation of these manuals for publication, EH. JuddTable of Contents (Part 1 ‘Design and Application of Leaf Springs SAE HS 788 Part 2 Design and Application of Helical and Spiral Springs} SAE HS 795 Pant 3 Design and Manufacture of Torsion Bar Springs SAE HS 796 Part 4 Incorporating Pneumatic Springs in Vehicle Suspension Designs SAE HS 1576 Part 5 Design and Manufacture of Coned Disk Springs (Belleville Springs) and Spring Washers! ‘SAE HS 1582Part 1 Design and Application of Leaf Springs SAE HS 788TABLE OF CONTENTS Chapter 1__ GENERAL DATA| 1. Introduction 2. General Characteristics of Leaf Springs 3. Leaf Springs for Vehicle Suspension Chapter 2 NOMENCLATURE AND SPECIFICATIONS I. Nomenclature - 2. Specification Requirements 3. Spring Eye Tolerances Chapter 3 DESIGN ELEMENTS T. Leaf Sections 2. Leaf Ends 3. Spring Eyes and Spring Ends 4. Spring Bye Bearings 5. Shackles: 6. Center Bolt and Cup Center 7. Center Clamp 8. Alignment Clips 9. Rebound Leaves .. 0. Variable Rate Springs Chapter 4 GEOMETRY 1. Detection Theory 2. Cantilever Spring . 3, Semi-Elliptc Spring 4. Center Link Extension Method... 5. Two-Point Deflection Method... 6 Layouts and Nomenclature Chapter 5 DESIGN CALCULATIONS]. T. Rate, Load and Stress 2. Stiffening Factor 3. Preliminary Calculations 4. Stress Distribution 5. Sample Calculation 6. Variable or Progressive Rate Springs 7. Strength of Spring Eyes Chapter 6 INSTALLATION EFFECTS| 1. Characteristics of Shackles 2, Windup of Springs 3. Twist of Springs ..... Chapter 7 INTERLEAF FRICTION 1. Characteristics 2. Measurement ...- 3. Control Chapter 8 OPERATING STRESS AND FATIGUE LIFE|.. 1. Operating Stress 2. Fatigue Life 3. Evaluation of Fatigue Test Results . 228 12 2 1s 1s 16 16 "7 19 22 24 4 7 2 n 8 231 37 37 39 3 38 6 a B n al 81 al - 8 33 83 85Chapter 9 MATERIAL AND PROCESSING] ......... 1. Steel . FBpsbonco00 Mechanical Properties ‘Surface Decarburization ... |. Mechanical Prestressing . .. 5. Surface Finishes and Protecting Coatings 2. 3. 4 5. Chapter 10 DESIGN DATA FOR SINGLE LEAF SPRINGS Single Leaf Types Rate Calculations Rate Factors Stress Calculations Practical Details ....... Camber of Single Leaf Spring Sample Calculations ‘Appendix A Conversion Table ‘Appendix B Derivation of Formulae for the Tabulated Values in Tables 5.2 and 5.3). 97 297 297 101 e108 --101 108 1108 3 cd 14 121 12Chapter 1 General Data 1, Introduction This Manual is written asa guide for the designer of leaf spring installations. It contains information which will make it possible to caleulate the space required for a leaf spring, to provide suitable attachments, and to determine the elastic and geometric properties of the assembly. The detail design of the spring itself also is described, but it was not the intention of the Committee to lay down fixed rules for this. The choice of leaf lengths, leaf thick- nesses, and leaf curvatures depends upon the type of in- stallation and upon the kind of service. Only an ex- perienced spring engineer can make the best choice of these factors. It is therefore recommended that the de- signer of a leaf spring installation consult a spring maker before the design is finalized. For standards and practices not covered in this Manual, see the current SAE Handbook. No attempt has been made to investigate or consider patents which may apply to subject matter presented in this Manual. Those who intend to use any of the construc- tions described herein should make their own investiga- (© tions and arrangements in order to avoid liability for in- fringements ‘The term multiceaf has generally been applied to springs of constant width and with stepped leaves, each of | ‘constant thickness, except where leaf ends may be tapered in thickness. More recently, the term has been extended to include an assembly of stacked “single” leaves, each of which is characterized by tapering either in width or in thickness or by a combination of both. Chapter 10 includes design data for single leaf springs which may be of variable width and constant thickness, constant width and variable thickness, or a combination of variable width and variable thickness. 2. General Characteristics of Leaf Springs The leaf spring, like all other springs, serves to absorb and store energy and then to release it, During this cycle the stress in the spring must not exceed a certain maxi- ‘mum in order to avoid settling or premature failure. This consideration limits the amount of energy which can be stored in any spring. For leaf springs based on a maximum stress of 1100 MPa, the energy listed in Table 1.1 may be stored in the Active part of the spring. If consideration of the inactive (© partof the spring required for axle anchorage, spring eyes, 14 ENERGY STORAGE OF STEEL SPRINGS AT 1100 MPa es Saf aoa nares tg ft eect et |e | ete ty | saat wrod 038 % eee esse Porett:) “For dszrpton of Type sae Caper 10 etc is included, the energy per kg ofthe total spring mass will be less than shown. For comparison, the stored energy inthe active material ‘ofa helical spring of round bar section is S10 J/kg at 1100 MPa, and for a torsion bar of round section is 390 J/kg at 965 MPa. This comparison shows that a leaf spring is heavier in mass than other types of springs. Balancing this disadvantage of mass, the leaf spring possesses the advantage that it can also be used as an attaching linkage or structural member. In order to be economically competitive, the leaf spring must therefore bbe so designed that this advantage is fully utilized. ‘Also, a leaf spring made entirely of full length leaves of constant thickness (see type F-1) is very much heavier and less efficient than a leaf spring made of properly stepped leaves (see type F-2) or single leaf springs (see types F-4, P22, T-1, and T-2) The maximum permissible leaf thickness for a given deflection is proportional to the square of the spring length. By choosing too short a length, the designer often ‘makes it impractical for the spring maker to build a satis- factory spring, although the requirements for normal load, deflection, and stress can be fulilled For example: A cube of steel, weighing 44 kg and mea- suring about 178 mm on each side, can be made into a spring carrying a load of 16 000 N at 125 mm deflection with a stress of 480 MPa. If 1500 mm is allowed for the length, the spring will look like Fig. 1.1. It will consist of 10 leaves, each 75.0 mm wide and 10.00 mm thick. fonly 750 mm is allowed for the length, the spring will look like Fig. 1.2. It will consist of 80 leaves, each 75.0 mm wide and only 2.50 mm thick. ‘When springs are made with stepped leaf lengths of type F-2, itis desirable to choose a length so thatthe spring will hhave no less than three leaves. Springs with many leavesA 1500 mm. Fig. K1—-Leat spring of Fengin — 760 mm — En 250mm 80 Leaves Fig. 1.2—Leat spring of type F-2: impractical design with inade- ‘quate length ‘te sometimes used for heavy loads, but they are economi« only where the shortening of the spring leads to def nite savings in the supporting structure. In addition, als Jowance will have to be made for increased spring rate and Breater eye stress, assuming the same load and width are used. In most installations the spring is also subject to windup loads. A typical example is that of the suspension spring (ina vehicle with Hotchkiss drive) which must withstand both driving and braking torque. The stresses under such loads are inversely proportional to the spring length; and the windup stifiness is proportional to the square of the length for the spring of given load rate (see Chapter 6) This is another reason why itis important to make the spring long enough and to check the resulting stresses and deflections. When a leaf springiis used as an attaching linkage, it will tend to guide the supported members in a certain geomet- tical path (see Chapter 4). Ifno other guiding members are used, the desired geometry must be obtained by properly placing the supporting parts on the structure which car. Fics the spring. If other guiding members are used, theit ‘eometry must fit that of the spring, or forces may be set up that will cause failure, ~ Leaf Springs for Vehicle Suspension Leaf springs are most frequently used in suspensions. This Manval, therefore, contains information which is 12 10.00 mm __j 10 LEAVES ‘type F-2:Practical design with adequate most useful in the design of suspension springs, bu it is also applicable to leaf springs for other installations. The characteristic of a spring suspension are affected chiely by the spring rate and the static deflection of the spring. The rate oa spring is the change of load per unit of deflection (N/mm). This is not the same amount at all Positions of the spring, and is diferent for the spring a8 installed. Static deflection of a spring equals the static load divided by the rate at static load it determines the stifines of the suspension and the ride frequency ofthe vehicle. In ‘Most cases the static deffection differs from the actual deflection of the spring between zero load and static load, due to influences of spring camber and shackle effet, AA soft ride generally requires a large static deflection of the suspension, There are, however, other considerations and limits, among them the following: 1. A more flexible spring will have a larger total deflection and will be heavier. 2. In mest applications a more flexible spring will cause more severe striking through or will require a larger “ride clearance” (the spring travel on the vehicle from the design load positon to the metal-to-metal contact pos tion), disregarding rubber bumpers. 3. The change of standing height of the vehicle due to variation of load is larger with a more flexible spring, The static deflection to be used also depends upon the available ride clearance. Further, the permissible static defection depends upon the sizeof the vehicle because of considerations of stability in braking, accelerating, commer- ing, ete. Table 12 shows typical static deflections and ride clear- ances for various types of vehicles. These values are ap. roximate and are meant to be used only as a general indication of current practice in suspension system design, The mass ofa spring subject toa given maximum stress is determined by the energy which isto be stored (sce Table 1.1). This energy is represented by the area under the load-deflection diagram, which therefore is also. mes. sure of the required spring mass. The following considera. tion will indicate what effect some changes in either rate or clearance will have on the required spring mass and therefore on the load-defection diagramNM | | i | ssangr asoncoies ‘0-0 res ‘neengoe) ser conene voo0 sours ‘atre ao) Fern to B08 faves Concerning changes in rate, Fig. 1.3 shows a theoretical oad-deflection diagram of a stiff (high rate) spring, and Fig. 1.4 that of a very flexible (low rate) spring, both for the same design load and clearance. The energy stored in zach, when fully deflected, is the same (1125 J), and the two springs will have almost the same mass if made of the same kind of material 15 000|N LOAD 100mm 60mm 0 50mm 100mm Set camer | Fig. 13-Theoretlealload-dfiecton dlagram of soring igh rate In the case of the stiff spring, energy and mass will be decreased by making the spring more flexible. In the case of the very flexible spring, energy and mass will be de- creased by making the spring stiffer. The dividing point ‘between these two cases is defined by "static deflector clearance”, The load-deflection diagram of this * ‘mum energy” or “minimum mass” spring is shown by a dashed line in both figures; it indicates a stored energy of 1000 J Concerning changes in clearance, Figs. 1.3 and 1.4 bring out the fact that a change in clearance by a given distance will affect the stored energy and therefore the required mass of the stif spring, much more than that of the very flexible spring. ? 1 100 m0 01mm 100% [STATIC DEFLECTION—-}CLEARANCE} Fig. 14 Theoretical load-detlection diagram of 2 low rate spring 13Chapter 2 Nomenclature And Specifications - 1, Nomenclature ‘The following terms are recommended for use on draw: ings and in specifications to avoid misunderstandings. The terms apply mainly to semi-lliptic springs. Datum Line—Most of the dimensions defined here refer to a datum line. In Figs, 2.1 and 2.2 (where the springs are shown inverted as in a machine for load and rate checking) itis shown as the line X-X. On springs with eyes, the datum line passes through the centers of the eyes. On other springs it passes through the points where the load is applied near the ends of the spring. These points ‘must be indicated on the drawing Seat Angle Base Line—(see Figs. 2.1 and 2.2.) Refer- ence line drawn through the terminal points of the active spring length at each eye, taken along the tension surface of the main leaf. For a Berlin type eye (sce Fig. 3.3 E, F, G in Chapter 3), the terminal point is the intersection of an extension to the contour of the tension surface with a perpendicular line through the center of the eye. On springs without eyes, the seat angle base line is coincident with the datum line. Loaded Length—(see Figs. 2.4-2.8,) Distance between he spring eye centers when the spring is deflected 10 the specified load position. On springs without eyes, it is the distance between the lines where load is applied under the specified conditions. Tolerance, + 3.0 mm. Loaded Fixed End Length—(see Figs. 2.4-2.8.) Dis- tance from the center of the fixed end eye to the projection on the datum line of the point where the centerline of the ‘center bolt intersects the spring surface in contact with the spring seat. Tolerance, + 1.5 mm. Straight Length—Distance between spring eye centers when the tension surface of the main leaf at the center bolt centerline is in the plane of the seat angle base line. The distance is measured parallel to the seat angle base line. Tolerance, + 3.0 mm. Seat Length— Length of spring that isin actual engage- ment with the spring seat when installed on a vehicle at design height. Its always greater than the inactive length. Inactive Length—Length of the spring rendered inac- tive by the action of the U-bolts or clamping bolts. For metal-to-metal type spring seats, this length is usually assumed to be equal to the distance between the insides of | the U-bolts, except for some curved seats where itis apt to be slightly shorter. For soft seats (using rubber type isolation, as in many passenger car installations) the inac- tive length may approach zero Ne Seat Angle—(see Figs. 2.1 and 2.2.) Angle between the tangent to the center of the spring seat and the seat angle 15 base line. When the spring is viewed with the fixed end of the spring to the left as shown, and the load is applied to the shortest leaf from above, the seat angle may be spe- cified as either positive (counterclockwise) or negative (clockwise), depending upon the angular direction in Which the tangent to the center of the spring seat is dis- posed from the seat angle base line Consequently, with the spring in normal vehicle posi tion so that the load is applied to the shortest leaf from below as shown in Figs. 2.4, 2.6, 2.7, 2.8, and 2.9, and again with the fixed end of the spring to the left of the ‘drawing, the seat angle is defined as positive when that tangent is disposed clockwise, and as negative when the tangent is disposed counterclockwise For suspension layout purposes, the seat angle is usually established with themain leaf straigh (see Straight Length). In this position (in which the center of the spring seat lies ‘on the seat angle baseline), the suspension layout specifies the contour of the main leaf, namely either: + flat, so that the seat angle is zero; or else + with front and rear segments being approximately circular arcs tangent to each other at the spring seat (“S- shaped” main leaf). This is then the tangent ofa seat angle defined as other-than-zero. Tolerance is usually held within + 0.5 deg, of as re quired for a particular application. For production checking purposes itis sometimes con- venient to deflect the spring to the position specified for load checking, and there measure the angle between the tangent to the spring seat and the datum line (instead of the seat angle baseline). For correct angle evaluation, the following relations between this “checked angle” and the seat angle in the straight-main-leaf (or any other) position must be considered: . 1. the two angles differ according to the distance by which the spring is deflected between any two positions, and according to the spring control (®, see Chapter 4). 2. The angles also differ according to the distance at cach end between the seat angle base line and the datum line. With the spring inverted as shown in Fig. 2.1, the terminal point of the seat angle base line is: 1) Higher by ID/2 than the datum line of the up- turned eye. ’b) Lower by 1/2 than the datum line of the Berlin eye. ©) Lower by (t + 1D/2) than the datum line of the downtured eye. 4) Identical with the datum line when there are no eyes. Finished Width—Width to which the spring leaves are ground or milled to give the edges a flat bearing surface.‘SHACKLED END FIXED END SEAT ANGLE Positive (POSITIVE AS SHOWN) OVERALL HEIGHT REBOUND LEAF HO. 1 DATUM LINE SEAT ANGLE BASE LINE OVERSLUNG SPRING WITH POSITIVE CAMBER Loa N02 ‘SEAT ANGLE No.1 (POSITIVE AS SHOWN) SEAT ANGLE BASE LINE POSITIVE OPENING DATUM LINE UNDERSLUNG SPRING WITH POSITIVE CAMBER FIXED END NEGATIVE Loa SHACKLED END OVERALL HEIGHT—4 SEAT ANGLE BASE LINE out Ra pres 10504, wo ap OB ay (NEGATIVE AS SHOWN) DATUM LINE TRESOUNO LEAF NO. 1 OVERSLUNG SPRING WITH NEGATIVE CAMBER FIXED END Loap SHACKLED END OATUM LINE Y SEAT ANGLE BASE LINE SEAT ANGLE (NEGATIVE AS SHOWN) CE vecative onenine esowo ver wo -———___] UNDERSLUNG SPRING WITH NEGATIVE CAMBER Fig. 2.1—Messurement of opening, overall height, and seat angle 16“TENSION SURFACE OF MAIN LEAF SEAT ANGLE BASE LINE ‘ALSO DATUM LINE WITH NO EVES JATUM LINE FOR BERLIN EYE CENTER OFMAINLEAF ‘or x BS INTEAF er SEAT ANGLE BASE LINE {LEAF THICKNESS ee © CENTER OF EVE (0 TERMINAL POINT FOR SEAT ANGLE BASE LINE x DATUM LINE FOR DOWNTURNED EVE =x ROE OF MAIN LEAF ‘SEAT ANGLE BASE LINE Fig. 2.2—Datum line and seat angle baee line for upturned, Ber- Tin downtumed, and ne eyen ees 17‘ScRIBE LINE ‘AS0UT 00 DEEP (ON EACH SIDE LOADING HEAD OF TEST MACHINE 1 OPTIONAL FOR aerate Fo. 23-tpng oad sk FIXED END SHACKLE END LOADED LENGTH #30 | ats — wor ff F : SEAT LENGTH ~ OPENING ‘CENTER BOLT SIZE qr STACK THICKNESS 1 ay 7 pL d —1'- 10. eve TL = = Lo. EVE 260 7 160 SPRING SHOWN UNDER = —=«NLOAD-— MATERIAL CLEARANCE mm HARDNESS RANGE RATE + N/mm LEAVES SHOT PEENED Fig. 24—Minimum specification requirements for undersiung “springs with positive opening 18t+ OVERALL FLAT LENGTH #3.2 LOADED LENGTH «3.9 —————___+| WIDTH c STACK THICKNESS _f OPENING —— SPRING SHOWN UNDER = + CLEARANCE = mm RATE + Nimm N LOAD j CENTER BOLT SIZE SEAT LENGTH MATERIAL HARDNESS RANGE LEAVES SHOT PEENED Fig. 25—Minimum specitication requirements for springs with, plain ends If the spring ends have a finished width, the required length of the finished edge must also be indicated (see distance A, Fig. 2.6.) The usual tolerances for finished width are: al Width Tolerance from "To and Nominal Width Over Including +000 ° 50 0.25 50 6 035 3 150 =050 Assembled Spring Width—Where more than one leaf constitutes a spring assembly, the overall width tolerance of the assembly within the spring seat length shall be as, follows: Tolerance from Nominal Width Ictuding 00, ° ey +25 63 100 30 100 125 $35 125 150 245 19 Stack Thickness—Aggregate of the nominal thick- nesses of all leaves of the spring including any liners and spacer plates which are part of the spring at the seat. Leaf Numbers—(see Fig. 2.1.) Leaves are designated by numbers, starting with the main leaf which is No. 1. The adjoining leaf is No. 2, and so on. If rebound leaves are used, the rebound leaf adjoining the main leaf is rebound leaf No. 1, the next one is rebound leaf No. 2, and so on (Rebound leaves are assembled adjacent to the main leaf on the side opposite. the load bearing leaves.) Helper springs are considered as separate units Opening And Overall Height—(see Fig, 2.1.) Distance from the datum line to the point where the center bolt centerline intersects the surface of the spring that isin contact with the spring seat I the surface in contact with the seat ison the main leat ‘or a rebound leaf (as on underslung springs), this distance is called “opening.” If the surface in contact with the seat is on the shortest leaf (as on overslung springs), this distance is called “over- all height.” “Opening” and “overall height” may be positive or neg- ative (see Fig. 2.1) They are specified dimensions not subject toa tolerance (see paragraph on Load in this chap- ter)FINISH WIDTH DISTANCE “A” {(IF REQUIRED) ! STOCK WIDTH LESS [ALLOWANCE FOR CLEARANCE = [WIDTH \ 7 : oy i #18 ——— -— i —____"____|_Loae LENGTH #3. BUSHING | +]-1.D. BUSHING | I FINISHED CENTER BOLT SIZE A —p- —-—-— fp f = | ll ' i —s— Nerack | ‘one natr oF FRoir poe | SEAT ernexe| |eve pushinco. 0. HEIGHT | "LENGTH eee |FOR TEST PURPOSES i 460+. 26.0 — = +60 __ 160... ——-i SPRING SHOWN UNDER = yioap ~~ MATERIAL CLEARANCE = mm HARDNESS RANGE RATE 9 N/mm LEAVES SHOT PEENED Fig, 26—Minimum spectication requirements for sprigs with ‘one eye and one plan end jWIOTH Tay of 1 FIXED END SHACKLE END — os LOADED LENGTH +3.0 ——| po + = SEAT LENGTH i CENTER BOLT SIZE 4 NEGATIVE OPENING rp] LD. EVE | i THICKNESS Lens6.0 — 26.0 J 7 260-0 46,0 --——- ——+ ‘SPRING SHOWN UNDER + NLOAD MATERIAL, CLEARANCE mm HARDNESS RANGE RATE = Nimm LEAVES SHOT PEENED INTERLINERS BETWEEN LEAVES Fig. 2.7—Minimum specication requirements for undersung ‘springs with negative opening 1.10FIXED END HELPER SPRING 260 LOADED LENGTH CENTER TO CENTER OF CONTACT PADS. STACK THICKNESS. HELPER SPRING SHACKLE END +30 WIDTH VERTICAL DISTANCE FROM CENTER OF FIXED SPRING EYE TO LOWEST POINT OF CONTACT PADS L 1D. EVE OVERALL HEIGHT MAIN SPRING MAIN SPRING roeve—| seacen as | 17 ZS neauine | cenren|pour size _stace esas 17, | + SEAT| LENGTH OF MAIN SPRING Joo *4 ao SPRINGSHOWNUNDER = ——~N LOAD HELPER TO CONTACT AT CLEARANCE ram RATEOFMAINSPRING = + N/mm RATEOFHELPER SPRING = N/mm Fig. 2.8—Minimum specification ‘commercial vehicle springs Clearance—Difference in opening, or overall height, between the design load position and the extreme position (of maximum stress) to which the spring can be deflected on the vehicle. Camber—Camber is not strictly defined and should therefore not be used in specifications, but it is sometimes convenient as a descriptive term. It is usually defined as the arc height of the main leaf. Camber is positive or negative analogous to opening, but this should not be confused with the fact that zero camber may be equivalent to either positive or negative opening, depending upon Whether the spring has upturned or downturned eyes. ‘Curvature—Curvature (1/R) is the reciprocal of radius (R). The curvature of a flat leaf is zero. Curvature is considered positive in the direction in which it increases With added load. Positive curvature corresponds with neg- ative camber. Load and Rate—Terms which are usually employed to describe the basic characteristics ofa leaf spring. As spe- «ified on the spring drawing, they refer to quantities mea- sured on the spring without center clamp and without N_ LOAD ON MAIN SPRING MATERIAL HARDNESS RANGE LEAVES SHOT PEENED quirements for overstung shackles. They are not the same as those of the installed spring. If it is necessary to specify load and rate as, clamped, this should be clearly shown on the drawing with full particulars of the clamp. Load is the force in newtons (N) exerted by the spring at the specified opening, or overall height. This force is greater during loading ofthe spring (“compression load”) than during unloading of the spring (“release load”). The specified load shall be the average of the compression load and the release load. For practical reasons, load as well as rate shall be measured in terms of compression loads only, but the compression loadin any positon shal be read only after the spring has been thoroughly rapped in that posi- tion with a plastic or soft metal hammer. ‘The tolerance on load at the specified overall height or ‘opening is usually expressed as a load range (N) which is equivalent 10 @ defection (mm) at the nominal rate (N/mm). This deflection may be as small as 6.0 mm for 4 passenger car spring and as large as 13.0 mm for a heavy truck spring Rate is the change of load per unit of spring deflection(N/mm). For leaf springs itis determined as one fiftieth (2%) of the difference between the loads measured 25 mm above and 25 mm below the specified position, unless otherwise specified (see Fig. 2.9.) The tolerance is usually hheld within -+ $% on low rate springs and within -+ 7% on high rate springs. ‘Measuring Methods—Instead of measuring loads at the specified position and 25 mm above and below, some users, measure loads at more than three positions during com- pression and release and plot a complete load-deflection diagram with a friction loop, similar to Fig. 7.1. This method requires more time but provides additional infor- ration, Such a diagram is preferably obtained by loading and unloading continuously and recording the data with an X-Y plotter. Load and rate are then obtained from the diagram. ‘When the load is measured, the spring ends shall be free to move in the direction of the datum line. The ends are usually mounted on carriages with rollers. The spring shall be supported on its ends, and the load shall be ap- plied from above to the shortest leaf It shall be transmitted from the testing machine head through a standard SAE loading block shown in Fig. 2.3 ‘The loading block shall be centered over the center bolt with the legs of the V resting on the spring. It is under- stood that the load specified on the spring drawing does rot inciude the force of gravity (usually called “weight” and equalling-mass times acceleration of gravity) on either the spring or the loading block. Just before the spring is checked for load or rate, it shall undergo a preloading operation. During the initial pre- loading by the spring maker, the spring shall be deflected at least to the position defined under the paragraph on Clearance. During any subsequent preloading, the spring shall be deflected only to and not beyond this “clearance position" in order to remove any temporary recovery from the set incurred during the initial preloading. After the spring has been preloaded, it shall be released to the free position before the load is applied for load and rate check- ing. Specification Requirements Minimum specification requirements are given in Figs. 24-29. They illustrate what information should be given to the spring maker for working out the detail design of| the spring ‘The spring shown in Fig, 24 is designed for underslung mounting. Therefore, the center bolt head is on the main leaf side and the height is dimensioned by specifying the opening. On an overslung spring, the center bolt head ‘would be on the opposite side and the overall height would be specified ‘The type of spring shown in Fig. 2.5 is often used on truck rear suspensions mounted in the position shown and with the center bolt head located as shown. In this case the ‘opening should be specified. Fig. 26 shows a spring which has a main leaf con- structed with an eye at one end and the other end plain; 4 construetion frequently used for truck suspensions. Fig. 2.7 shows the details of an underslung rear spring designed with considerable negative opening. The combination of main spring and helper spring shown in Fig. 2.8 is frequently used for truck rear suspen- sions and is mounted as an oversiung spring, Fig. 2.9 shows an overslung variable rate spring of the multistage type. The graph indicates the method of mea- suring rates of such springs, where rate (1) is usually measured at “curb load” (that i, at the load on the spring which is due to the mass of the vehicle without any pay- load), while rate 2) is measured at “design load” (that is, at the load on the spring which is due to the mass of the vehicle plus the payload) 3, Spring Eye Tolerances Spring Eyes and Bushings For round eyes with specified inside diameter, the size and roundness of the eye should be checked by means of 1 round plug gage from which two opposite seginents of 60 deg have been removed. The gage shall be tapered by 0.05 mm in diameter per 25.0 mm of length (see Fig. 2.10.) ‘The gage shall be inserted into the eye three times from cach side at angular positions differing by about 60 deg. ‘The eye is acceptable only if the gage reading on the side of the eye from which the gage is inserted is within the specified diametral limts at each of the six checks. ‘Also, the round eye should be checked with a round plug, GO/NO-GO gage, to determine if the eye is cone shaped or tapered. The GO diameter must pass com- pletely through the eye, and the NO-GO diameter must not enter the eye from either side. ‘The total tolerance shall be 19% of the nominal diameter of the eye, except for large diameter eyes (40 mm or more), where bushing retention may require a smaller tolerance of 0.75% of the nominal eye diameter, For eye diameters of less than 25 mm, the minimum tolerance is 0.25 mm. Where the ID of a bushing may have been affected by pressing into the spring eye, it should be checked with a round plug gage. Total tolerance is to be 0.13 mm unless otherwise specified Oval eyes (see Fig. 3.3H) consist of two lialf circle ends joined by flat sections. One method of cinecking their sizes is by using a GO/NO-GO plug gage system. This consists of: 1. An oblong GO gage to the minimum inside dimen- sions. 2. An-oblong NO-GO gage to the maximum inside dimensions for the half circle ends only, with the flat sides of the gage undercut 3.A rectangular NO-GO gage forthe inside dimensions between the fat sides only. 112LOAD AT METAL, ‘TO METAL POSITION (MEASURE RATE ‘OVER 25 mm DEFLECTION LOAD (2) RANGE SECOND STAGE LOAD IN) CLEARANCE LOAD (i) RANGE DEFLECTION FROM CURB LOAD (1) TO DESIGN LOAD (2) FIRST STAGE OVERALL HEIGHT {rmm) FIXED END MATERIAL HARDNESS RANGE LEAVES SHOT PEENED. CLEARANCE BEYOND DESIGN LOAD (2) mm CURBLOADI) = ¢ Nat OAH. DESIGNLOAD(2) + Nat OAH. FIRST STAGE RATE = + N/mm SECONDSTAGERATE = N/mm SHACKLE END 15 LOADED LENGTH AT DESIGN LOAD (2) T t t Cnr +! b-1D. EVE | = reese sracKes i 7 Ty THICKNESS-| a OVERALL FIRST STAGE HEIGHT ——= TeAveE 1 if CENTER BOLT SIZE SEAT LENGTH Fig. 2—Minimum specification requirements for variable r SECOND STAGE LEAVES (SHOWN HERE AT A LOAD BEFORE REACHING INITIAL CONTACT WITH FIRST STAGE LEAVES) oF progressive rate springs (oversiung type shown) 113‘THESE LINES TO BE 0.25 DEEP AND 0.25 WIDE AFTER FINISH GRIND ~__ OTHER LINES TO BE 0.13 DEEP AND 0.13 WIDE AFTER FINISH GRIND SO / >. = = 12.0 TYPICAL ~ | t "EEE EEE X x x x x x] 60° FLAT 4 I r FEE 4 6 | | | | 6.0 WALL | 60.0 o + 60.0 + 60.0 + 600 oie 60 0.05 PER 25.0 CONSTANT TAPER FOR THIS ~ DIST. 300.0 ————— 400.0 X-STAMP GAGE DIAMETERS AT THESE STATIONS MATERIAL: STEEL - G40270 (SAE 4027) OR EQUIVALENT PROCESS: CARBURIZED AND HARDENED; CASE DEPTH 0.50 MIN. SURFACE HARDNESS: Re 58 MIN. Fig. 2.10—Gage—teat soring eye plug Parallelism and Squareness of Spring Eyes Eyes of the main leafin the assembled spring, measured in the unloaded condition, shall be parallel to the surface at the spring seat, and square with a tangent to either edge of the main leaf at the spring seat, within + 1 deg.Chapter 3 Design Elements 1, Leaf Sections! For automotive springs, round edge flat steel was adopted as the SAE standard in 1938. The bars shall be of flat rolled steel having two flat surfaces and two rounded (convex) edges. The cross section tolerances per- mit the two flat surfaces to be slightly concave. When that ‘occurs, the radii of the arcs of the two concave surfaces, shall be of approximately equal length. The rounding of the convex edges shall be an arc with a radius of curvature that may vary from 65-85% of the thickness of the bar. Bars shall be substantially straight and free from physi- cal characteristics known as “kinks” or “twists” which render them unsatisfactory for spring manufacturing pur- poses. Distortions due to a bar being bent about either major axis of section shall be measured with the bar against a fiat checking surface so as to make contact with this surface near both bar ends. Gaps between the bar and the check- ing surface shall not exceed 4.0 mm/I m of bar length out of contact with the checking surface when this bar length is greater than 1 m, Also, a gap between the bar and a straight edge 1 m long applied along any portion of the surface or edge of the bar shall not exceed 4.0 mm. It is recommended that all leaf spring bars which have been cold straightened be identified by the steel mill so that the spring manufacturer can use them selectively ‘The bar sections, which are generally provided in alloy steel, shall be specified and rolled in the widths and thick- nesses shown in Table 3.1. These sections are subject tothe tolerances given in Table 3.2. TABLE 31 ine) With soo | soo} se] aco | i120 | teoo | Zao | aro Tables showing the mass per meter length and the ac- tual moment of inertia for each size of these bars are provided in Chapter 5. Ref, SAE 11123, ‘TABLE 32-€R0SS SECTION TOLERANCES inn) vam |ronee | meine, in Teknees™ ek 0 sia | atzo soo | tom |ou| ors | = | sos [oo | — seo | tors ore | ors | = | 00s | cos | = 300 | fias | ois | 020 | 030 | oos | oro | ors woo | +20 | - | ox | os | - ors | 02s “Phe mena lS ae te we at Sin terre rears rematch tna centr ct br ray etna an ne chnoe eae cacao tay ever ered eines ae ope. ‘Miura otoorc eines een eto 056 of ech Ba It is well known that fatigue failures in spring leaves usually start on the tension side of the leaf. Taking advan- tage of this fact, special sections shown in Fig 3.1. have been developed which place the neutral axis nearer to the FLAT SECTION (SEE SAE STANDARD) - TENSION SURFACE GROOVED SECTION = TENSION SURFACE PARABOLIC EDGE SECTION Fig. 3.1—Sections of spring steettension surfaces, Springs made of such sections are $-10% lower in mass than those made of conventional section. When grooved sections are used, special precautions should be taken to prevent corrosion caused by the mois- ture which tends to become trapped within the grooves. It is suggested that spring users interested in these special sections contact the manufacturers who produce such springs 2, Leaf Ends ‘Square End (Blunt End) (Fig. 3.2) This is the cheapest end to produce but i often unsatis- factory. It eauses concentration of interleaf pressure, re- sulting in more friction and galling than tapered ends. It isa very poor approximation of the theoretical triangular leaf uniform stress spring, and is therefére heavier thant necessary. Diamond Point (Spear End) (Fig. 3.2B ) This end makes a better approximation of the uniform stress spring by omitting excess material. The pressure distribution between leaves is slightly improved. Tapered End-(Fig. 3.2C) 7 This end can be formed to approximate very closely the {deal uniform stress shape. The plan view contour is con- ‘A. END SQUARE AS SHEARED trolled by trimming or edge squeezing as part ofthe taper- ing operation. Due to the flexibility of the leaf end, the pressure distribution in the bearing area is improved and interleaf friction is generally reduced. ‘Tapered And Trimmed End (Fig. 32D) This end is similar to Fig 3.2C, except that the plan view contour is controlled by trimming after the tapering oper- ation, and thus has the added advantage of the maximum obtainable area of contact. 3. Spring Eyes and Spring Ends Upturned Eye (Fig. 3.3A) This construction is most commonly used. If required, the second leaf can be extended to give support to the eye. Military Wrapper (Fig. 3.3B) In this design no attempt is made to use the second leaf ‘wrapper as an eye under design loads; but it may come into action on rebound and thus assist the main leaf. It also provides an emergency support if the main leaf breaks, The design has beer widely used on military vehi- cles and trucks where the service is severe. C. END TAPERED D. END TAPERED, THEN TRIMMED BURRS AWAY FROM BEARING SURFACE - = > Fig. 82—Leat ends 1.16ee sss @ ‘A. UPTURNED EYE —= o— €= D. DOWNTURNED EYE S 7 MELO G. WELDED EYE E. BERLIN EVE 8. MILITARY WRAPPER, ©. PLAIN END MOUNTING F. BERLIN EVE WITH MILITARY WRAPPER a. H. OVAL EVE Fig. 3.8-Spring eyes and spring ends Plain End Mounting (Fig. 3.3C ) This construction can be built as a flat leaf on a curved spring pad, or as a curved leaf on either a lat or a curved. spring pad. The leaf ends used with the rubber insulators shown in Fig. 3.11 are similar, but are sometimes provided with a T end or a circular hole for the transmission of lengthwise forces, Downturned Eye (Fig, 3.3D) This is sometimes used because it produces a desired spring geometry (or suspension motion) which may im= prove steering or axle control. If support to the eye by the second leaf is required, this construction is not recom- mended Berlin Eye (Fig. 3.3E) Longitudinal loads are applied centrally to the main leaf, thereby reducing the tendency of the eye to unwrap, Berlin Eye With Military Wrapper (Fig. 3.3F ) This construction is a variation of Figure 3.3B. Welded Eye (Fig. 3.36) This construction is used predominantly in applications such as torque rods where the horizontal force is high. The welding must be performed before heat treatment, using appropriate technique. Oval Eye (Fig. 33H) This eye construction permits the use of rubber bush- ings which have different rates in the vertical and horizon- tal directions. This eye was developed specifically to re- duce the magnitude of the horizontal force inputs in suspension applications 4. Spring Eye Bearings Threaded Bushings (Figs. 3.4, 3.5, and 3.6) This type of construction has the following advantages: It takes side thrust as well as vertical load, retains lubri- ‘cant, and excludes dirt better than a plain bushing, thus requiring less frequent lubrication. Spring eyes need not be finished in width, 447Fig. 2.4—Spring eye bearing: Threaded bushings in one piece C snackle Fig. 3.5Spring eye bearing: Threaded bushings and pins in fone piece Y shackle ‘The bushing has @ 60 deg V thread on the inside which fits the pin loosely. The outside is either plain or provided with a very flat angle thread. It is forced into the spring. eye or bracket ‘The bushings and pins are made of carbon or alloy steel, ‘carburized and hardened. | ‘Thread sizes generally used are M14X2 to M364 Load pressures up to 7.00 MPa on the projected area at normal load are used. Figs. 3.4 and 3.6 show seals. Self-Lubricated Bushings (Fig. 3.7) — Various designs and materials have been introduced on passenger cars and light trucks. They do not require lu- Fig. 3.7—Spring eye besring: Selt4ubricated bushings in double bolted shackle brication and are noiseless. The design shown in Fig. 3.7 takes side thrust on rounded citcular grooves and ridges. Spring eyes are not finished in width. Some types will stand pressures up to 8.40 MPa on the projected area at normal load. Plain Bushings (Fig. 3.8) This type bearing, usually bronze, is used on heavy trucks, It is simple (0 manufacture and service, and will sive satisfactory life if it is regularly lubricated. Side thrust is taken on the finished faces of the spring eyes. The wall thickness is usually 3.0 mm. Load pressures used are be- tween 3.50-7.00 MPa on the projected area at normal load. 118SECTION A-A Fig. 2.8—Spring eye bearing: Plain bushings, periodically lu- bricated, in double bolted shackle Rubber Bushings (Figs, 3.9 and 3.10) Various types are used successfully. Their flexibility is aan added insulation against noise, but the effect of the flexibility on road holding ability, steering control, and increase in spring rate must be considered. Rubber Cushion (“Shock Insulator”) (Fig. 3.11) This bearing is used on heavy vehicles. The design per- mits a limited amount of longitudinal motion of the spring ends. It is, therefore, successful only with fairly long springs which are approximately flat at design load. 5. Shackles ‘One Piece C-Shackle (Fig. 3.4) Used on passenger cars and light trucks. Fig. 3.8-Spring eye bearing: Rubber bushing in fixed eye pivot 119 One Piece Y-Shackle (Fig. 3.5) Has more load capacity than one piece C-shackle. ‘Taper Pin Shackle (Fig. 3.6) Double Bolted Shackle (Figs. 3.7 and 3.8) ‘These constructions kave been used where they must resist forces transverse to the spring (that is, in the direc- tion of the spring eye axis). Careful design with close fits ‘must be used to avoid loosening in service. Fig. 8.10—Spring eye bearing: Rubber bushings In shackleFig. 3.11—Spring end bearing: Rubber cushion (“Shock insula ph Riveted Bolted Shackle (Fig. 3.10) Well suited to rubber bushings, and therefore, in gen- ‘eral use on passenger cars and light trucks. 6. Center Bolt And Cup Center ‘The center bolt is required to hold the spring leaves together, and the center bolt head is used as a locating dowel during installation to the vehicle. For underslung springs, the head should be adjacent to the main leaf; and for overslung springs, the head should be adjacent to the short leaf. In most cases, center bolts are highly stressed in the ‘handling of the spring and in service. Therefore, it is nec- essary to use bolts and nuts of high mechanical properties. ‘The diameter of the center bolt hole in the spring leaves should be at Jeast equal to the thickness of the heaviest leaf in order to permit cold punching. If the diameter of the center bolt hole should be less than the thickness of the leaf, it may require heating the leaf in the area to be punched. However, itis not recommended to cold punch leaves which are thicker than 14 mm, (See Table 3.3 for sizes.) Generally, the spring leaf material at the center bolt area is inoperative when assembled to the vehicle. How- ‘ever, the diameter of the center bolt hole should not be too large in relation to the width of the leaf. When itis not desirable to use a center bolt and hole in the spring leaves, @ nib or cup, for nesting the adjacent leaves, is forged from the leaf material itself by forming a depression on one side and a corresponding projection on TABLE 3-RECOMMENDED CENTER BOLT ANO MUT DIMENSIONS (rm) _ tenate ‘ase * 15 | as | 100 | 700 | 1500 12 [138] 30 | treo | 00 | se0o | 207 | sor the other side, The leaves are then held together with clamps. ‘Cup centers are often used in heavy duty springs which may not safely depend on clamps and center bolts to prevent shifting of the spring on the axle seat due to driving and braking forces (See Fig. 3.14.) ‘When the main leaf is assembled adjacent to the axle seat as in underslung springs, the cup is hot forged in the ‘main leaf only, (away from the No. 2 leaf), When the shortest leaf is mounted above the axle seat as in overslung springs, all the leaves must be cupped toward the shortest leaf. ‘This method of cupping locks the main leaf to the axle seat. The horizontal forces which are applied to the main leaf will be resisted by the cup rather than the clamp and the center bolt. ‘ ‘There are many types of cup centers in general use, one ‘of which is shown in Fig. 3.12. The cup dimensions are listed according to center bolt diameter; however, the cup diameter should not exceed one-half the leaf width and the cup depth should not exceed one-half the leaf thickness. & 45086 Leo — Fig. 3.12—Dimensioning of typical cup center i Darerer® 7. Center Clamp The center clamp provides the permanent tie between the leaves, and between the spring and the spring seat. Figs. 3.13, 3.14, and 3.15 show some typical designs 1.20Fig. 3.19—Typical center clamping of oversiung spring Fig. 2.14~Center bolt assisted by cup centersFig. 9.15—Center clamping of undersiung spring with rubber pads The functions of the clamp are to attach the spring firmly to its seat to prevent leaf breakage through the center bolt section, and to prevent center bolt breakage due to horizontal forces. It therefore must remain tight in service. Excessive clamp length reduces the active length of the spring and wastes metal. The clamp length is usually be- tween 8-159 of the spring length. The ends of steel clamps should be well rounded to avoid sharp edges in contact with the spring leaves. Clamps with rubber pads (Fig. 3.15) are frequently used on passenger car rear axles because of their important influence on reduction of noise transmission. Similar to the rubber shackle bushings mentioned in Section 4 ofthis chapter, considerations of road holding, axle control, and steering control set a limit to the amount of softness which it is permissible to use at the center clamp. The center clamp has an effect on load rate and on loaded height of the spring. This is discussed in Chapter § Section 2 8. Alignment Clips Alignment clips are used to limit sidewise spread and vertical separation of the individual leaves in the spring. Bolt Clip (Fig. 3.164) ‘This lip is used for most heavy springs. The clearance between the bolt and the main leaf must be suficient to permit the main leaf to twist longitudinally so that this twist will not be concentrated in the free ends near the eyes. A spacer tube is recommended to prevent the sides 122 of the clip from binding the main leaf. For heavy duty applications, a double rivet construction may be used in springs 100 mm wide and over. Material is hot rolled steel strip of the following sizes: 4.5 x 20, 6.025, 6.0% 30, 8.0 35, 10x40. Clinch Clip (Fig. 3.168) This is used on springs where the clearances are limited. ‘The material is usually hot rolled steel strip 4.520 and 6.0x25 size. Single Piece Box Clip (Fig. 3.16C) ‘This clip is manufactured from hot rolled steel strip 2.5 X25. Its used with and without a rubber liner. Bolt sizes are M8X 1.25 and MIOX ‘Two Piece Box Clip (Fig. 3.16D) This clip is also made from 2.5%25 hot rolled steel strip, Note thatthe side ofthis clip, as in Fig 3.16C, are straight and provide a clearance forall leaves. Bolt sizes are MB 1.25 and MIOX 15. ‘Tab Lock Clip (Fig. 3.16E) ‘This clip is also made of 2.525 hot rolled stel strip. It is used with and without a rubber liner cn the main leaf only. ‘Tab Lock Clip With Locating Tang (Fig. 3.16F ) ‘This clip is similar to Fig. 3.16E except that it has a tang. for retaining it on the spring leaf.4119 G3LU3ANI “H Th fh ag ‘dy wwawuBny—1'e “Bis di19 dvuLs °D — £ st , ano wooT av 3 wr S : | ==. oa ef ve X09 393Id JIONIS: gees 5 Weer eee an SSS)Strap Clip (Fig, 3.16G) This clip has a rubber liner on all four sides of the spring, The strap is 0.5 X 16 stainless steel. The ends are ‘overlapped and secured by means of a fastener which is ‘crimped at assembly. Inverted Clip (Fig. 3.16H) This clip is used on heavy springs where there is a clearance problem. Itis usually made from 30 or 35 x 8.0 stock 9, Rebound Leaves Figure 2.1 shows an example of a spring with rebound leaf. There ae two principal objectives for the application ‘of one or several rebound leaves. One of these concerns the spring subject to very high and frequent windup loads which tend to distort the main leaf by separating it from the shorter leaves. The rebound leaf or leaves serving to prevent these excessive windup stresses in the main leaf act substantially like the other leaves and should be treated in the same manner as to length, free curvature, ete. They are loaded through the alignment clips, which must be properly placed and de- signed so as to maintain tip contact between rebound leaves and aiain leaf. “The other case is that of the truck spring which in extreme rebound will have to support the weight of the axle and of other unsprung components, thereby becom- ing subject to detrimental reverse bending stresses, partic- ularly in the main leaf, In this ease the rebound leaves are usually formed with less free curvature than the other leaves. When the leaves are bolted together in the spring assembly operation, assembly stresses are set up in the various leaves including rebound leaves, as described in Chapter 5, Section 4. The rebound leaf will protect the main leaf in the region between the eye and the nearest alignment clip from distortion in extreme rebound. ‘As long as rebound leaves are under load, they contrib- ute to the load rate in the same manner as other leaves, regardless of the free camber in the rebound leaves. 10, Variable Rate Springs Variable rate springs are used primarily on vehicles which operate with large variations in load, such as trucks and buses. Variable spring rates are generally required to provide desirable ride and handling characteristics under these conditions. There are several ways to obtain variable rates, some of which may be combined with others. The helper spring is one method of obtaining increased. rate with deflection. As shown in Fig, 2.8, the helper mounted above the main spring and has its own bearing pads. The helper spring does not support any load until contact is made with the bearing pads. The change in rate at contact is necessarily abrupt. Shackles may be used to obtain some variation in rate as described in Chapter 6, Section 1. Curved bearing pads or cams which shorten the effee- tive length of the spring as itis detected will provide a variable rate. Such a configuration is shown in Fig. 3.17A. ‘Another method to obtain variable rate is by means of A. VARIABLE EFFECTIVE LENGTH SPRING ne B. MULTI-STAGE SPRING Fig. 2.17—Variable rate springs 124the multi-stage spring, shown in Fig. 2.9, and Fig. 3.17B. This spring has one or more leaves called “second stage” leaves, mounted adjacent to the shortest leaf of the main ‘or “first stage” portion of the spring. This spring gradually increases in rate with deflection as the contact between the stages increases. Load and rate for each stage are usually 125 specified as shown in Fig. 2.9. They are generally checked in the same manner as single stage multi-leaf springs. Combinations, such as the use of curved bearing pads in conjunction with a multi-stage spring, are sometimes used to provide a greater change of spring rate.Chapter 4 Geometry 1, Deflection Theory ‘As a spring with leaves of constant cross section prop- erly stepped to approach the condition of uniform strength is deflected, it will assume the shape of a circular arc at all loads between zero and maximum, provided it has acircular are shape or is lat at no load or at any given load. Most springs approximate these conditions closely ‘enough so that the circular arc shape can be used to calcu- late their geometric properties. The following relations have been derived analytically and found to agree closely g | L arc HEIGHT with a number of actual springs checked. (However, see Chapter 10, Section 6 concerning the contour of single leaf springs in the free camber.) 2, Cantilever Spring For a spring of this type the center of the eye of the Berlin type moves in a path with a radius of 0.758 central to the main leaf, as shown in Fig. 4.1A. Ifthe eye center is offset the distance “e” from the center of the main leaf, the center of arc will be offset by O.Se in the opposite 3NITH3LN3O _ aN WLVa ainios~ + 206 ~1V3S ONIUaS “ y ANIOd aaxis 1.30‘TABLE 4.1—GEOMETRY FORMULAE FOR SEMI-ELLIPTIC SPRINGS |A.For conventional sping where W = Z/Y? equals one ten Bde, + 2-69) + 20(0, ~ 0) gexyety 2 oe wre oven 0573 + +b) SIV + HOLY 673 - oa) S73" +1) 57.30 - a) eas uy 7.31¥2 = 1) ot 8 For unconventional spring where W = Z/¥? does not equal one yw ze q = Same asin A 3 + Ya + NE + VP + A + 2 Fig. 4.4) 13, Depending upon th acuracy demande ofthe layout, a correction for the effect of the shackle may be necessary, particularly when the shackle angle is excep- tionally small (B less than 60 deg in the flat main leaf position) and the shackle is exceptionally long. The cor- rection may be made in the following manner: Locate point P at intersection of datum line and Ro, where R, is equal to distance OA. ‘After determining linkage layout for a given deflection, such as for rebound, point B, is foceted. Locate point P, on extension of line B.A. Locate point O, where chordal distance PO equals PO, Draw line O.M, to give corrected tilt to center link DE in rebound. In like manner, establish line O.M, to give corrected tlt to center link DE in compression. ‘These corrected positions of center link DE determine the corrected control in degrees per millimeter (equal to 6'/x) and can be used to establish the approximate radius Ry; for the corrected axle path 20H 2 eres, 7 w+ va @-wed (See Fig. 45) Zev we 5. Two-Point Deflection Method (Fig. 4.3) This method has the advantage that all of the layout work can be done within the overal length of the spring. In cases where the unsymmetry factor is small and the © point is far from the axle center, it is the only known procedure which permits construction within the confines of the standard layout board and straight edge. The principle of this method is based upon the use of the two cantilever deflections corresponding to a given deflection atthe center ofthe spring seat. These deflections ‘may be computed for two vertical positions of the spring. seat, for example maximum compression (metal-to-metal) ‘and’ maximum rebound. When they are applied to the three-link equivalent of the spring with the main leaf in the flat position, the path of the axle and the angles of the spring seat can be determined emtirely by construction. Procedure |, Start layout with main leaf in flat position with1H 90d Jo ured sutuso.0p 01 ssrowered Yo} we Yy id TT T T f oe a 8 on so 2 ¢, te * : 4 a 2 . 8 2 1% 31 2, me =e | ¥ (ae) ve7- v= Wy (OILVY aLWY) 94/29 = Z a = ul (O1LVa HLDNAT) 8/4 = A aa eA/Z=M [2(b + A) ZA + zIZA+Z) E1(L +A) a Zea +Z€lengths a,b, and L measured along the main leaf and axle center Hat distance h from center of main lea. 2. Establish lengths m and n, which represent the inac- tive material 3. Draw are R, and atthe intersection with 0.5e, locate point D. 4. Draw arc R and atthe intersection with 0.5¢, locate point E 5. Construct the three links AD, DE, and EB. 6. Locate point M at intersection of centerline of center bolt and link DE. 7. Draw reference lines AF and BG through the eye centers and perpendicular to the extension of the center link DE. 8. For any given deflections such asx, and x, compute f, and f. from the following formulae (see also Table 4.1) and draw arcs about point F. For a conventional spring: For an unconventional spring: yy +0 zZ+¥ 9. Similarly, for given defections x, and x., compute g, and & from the following formulae (see also Table 4.1) and draw arcs about point G. For a conventional spring: reeven(2) For an unconventional spring: (=) Y 10, Tangent line to arcs f, and g, establishes the position of center link DE in rebound, and tangent line to arcs { and g. establishes the position of center link DE in compression. 11. For each position of the center link DE the axle position can be located by the triangle DEH. When three ‘or more such positions have been located, the approximate radius Ry for the axle path can be established by geomet- ric construction, 12. The control in degrees per millimeter is equal to the angular change in the positon of the center link divided by the deflection x: @ = @/x (deg/mm). 13, In the symmetrical spring the control is zero, with the center link moving parallel to itself throughout the compression and rebound range. Actually, however, the center link undergoes a small angular change due to the vertical displacement of the shackled spring eye. 14, Depending upon the accuracy demanded of the lay- out, a correction for the effect of the shackle may be necessary, particularly when the shackle angle is excep- tionally small (B less than 60 deg in the flat main leaf 1.34 position) and the shackle is exceptionally long. The cor- rection may be made in the following manner: ‘Afier determining linkage layout for a given deflec- tion, such as for rebound, point B, is located. Draw arc d, where d, = (g ~ u). Draw tangent line to ares f; and d, to-establish cor- rected position of center link DE in rebound. (Note that the layout deflection x, is changed to x’, in applying this correction.) Similarly, after locating point B., draw are d, where @ +0). Draw tangent line to arcs f. and d, to establish cor- rected position of center link DE in compression, (Note that the layout deflection x. is changed to x’, in applying this correction.) These corrected positions of center link DE deter- ‘mine the corrected control in degrees per millimeter and ‘can be used to establish the approximate radius R'y for the correct axle path The difference between deflections x., % and X., X', respectively is so small in a full size spring layout that it usually can be neglected. However, when required, an overlay of triangle DEH positioned with point M lying on horizontal line through deflection points M, and M, and with points D and H lying on ares R, and R'y respectively, will determine exactly the angle of seat at the original deflections, x, and 6, Layouts and Nomenclature Figures 4.2 and 4.3 were constructed with exaggerated eye diameters in order to illustrate the mechanics of con- struction. Also, the relatively small shackle angle and long shackle used in these layouts results in a large vertical displacement to the shackled spring eye. This was done to better illustrate the effect of the shackle correction. Under these conditions the compression shackle (Fig. 4.2.) re- sults in a control increase during rebound and a control decrease during compression, while the reverse is true for the tension shackle (Fig. 4.3.) Nomenclature for Figs. 4.2 and 4.3 and for related for- mulae in Table 4.1 ad cantilever length (including inactive length m) ceallad “font” length b= Shackled cantilever length (including inactive length n) aed “rear” length bsp for compression (ounce) positon of krkage rangent arc rads at 8 where d = (g + u) ccentity = 0.5 ([email protected]. + 1) Yor Berlin eye, e equals 2000) Detection of point F Dottection of point G ‘Axle distance trom center of main leat ‘Spring rate, N/mm (clamped), See Chapter § Front cantlever rate, N/mm (clamped) ear cantlever rate, N/mm (clamped) = Total spring length—measured along flat main leatFront inactive length = Rear iacive length Distance of point O below datum tne Distance between pons © and M {see Figs. 4.2 and 445) the "equvalent torque arm” or verte loac—see Chapier 6 1 = Subsorpt or rebound positon of inkage Ry = 075 (am) Ry = 075 (on) Radius of ade path = path of point H (approx) jy = Radu of path of point M approx. Ry = A= {) Fo = Radius of pints P and 0 1° 2 Thickness of main lea 2. " 1.35 er eENncxgcc 0 = Distance between points 8 and G "= Front cantiever rate proportion (ky/k) abo 2/¥9 (see Fig. 4.4 ford and Fig. 45 for L/Q) Detection of point M = Lengt ratio (b/a) ‘= Gantlover rate rato (hy/ky) Shackle angie with main leat flat, deg ‘= Angular displacement of center lnk, dog = Parameter to determine Ry, (see Fig. 42) Control of center lok (87x, deg/mm). known a both ‘pring contol and “aslo con” Primes denote shackle effect covrectionsChapter 5 Design Calculations 1. Rate, Load and Stress ‘A leaf spring may be considered as a beam of uniform strength composed of leaves of equal thickness where the fiber stress i the same throughout the length ofthe beam ‘This approximation is justified for most springs within the accuracy necessary for layout work and—with certain correction factors—for estimates of the necessary length, thickness, width, and number of leaves. It also Serves as a base for more detailed calculations. Figure 5.1 shows a cantilever of six leaves and the same cantilever rearranged with the leaves split and laid side by side for comparison with the triangular beam of uniform strength which is shown in broken lines. It must be ap- preciated that the straight line contour shown in Fig. 5.1 applies only when all the leaves are ofthe same thickness. See the remarks under the “Stepping” paragraphs in Sec- tion 4 of this chapter. SIDE VIEW TSS PLAN VIEW CONTOUR OF UNIFORM STRENGTH BEAM LEAVES SPLIT AND LAID SIDE BY SIDE Fig. 5.1—-Description of multhieat spring 137 The formulae in Fig, 5.2 are given for leaf springs which approximate beams of uniform strength (except for the last column) and are derived from the following funda- ‘mental facts: 1, Stress is proportional to leaf thickness multiplied by change of curvature. 2. Change of curvature is proportional to change of bending moment divided by moment of inertia. 3. Stress is proportional to leaf thickness multiplied by bending moment divided by moment of inertia, For comparison, the formulae for the cantilever spring of one leaf or of several full length leaves (so-called “uni- form section beam”) are also given in Fig. 5.2. In this spring the stress is highest at the clamping area, and the stress formulae refer to this highest value. It will be noted that with the same load, length, thickness, and stress, the uniform section spring produces only two-thirds of the deflection and weighs twice as much as the uniform strength spring, or in other words, the uniform strength spring is three times as efficient as the uniform section spring AAs shown by the different formulae in Fig. 5.2, the stress can be calculated from strain, deflection, or load, depending on what information is known, tis evident from the “stress from strain" formula that for the same change in curvature (1/R—1/R,) the stress will vary directly with the leaf thickness ‘Again, it will be seen from the "stress from deflection” formula that the stress will vary directly with the leaf thickness and inversely with the square of the effective spring length. ‘The “stress from load” formula is the standard beam formula for stress where, for a given load, the stress will vary directly with effective length and inversely with the ‘square of the lea thickness. This may appear paradoxical ‘when comparing the “stress from load” formula with the “stress from deflection” formula, However, in the “stress from load” formula the deflection is not considered. If the expression for load (P) is replaced by the product of rate (&) and deflection (f, the “stress from load” formula will reduce to the “stress from deflection” formula, Although the foregoing substitution marks the two for- mulae as merely two different ways of expressing stress, the “stress from deflection” formula is particularly signifi- cant for this reason. It shows that for a given stress and deflection the leaf thickness varies as the square of the effective spring length, and since thin leaves will not sup- ply sufficient strength for the spring eyes, the formula ‘emphasizes the desirability of long springs. Another im- portant consideration for the use of long springs isthe factUnsymmetrcal Somtalliptic Ry AT CLAMP CLAMP — RADII, ANOR VARY ALONG THE LENGTH OF ‘THE SPRING. Cantilever Uniform Section Cantilever Se - = Type Symmatical (a= b) Sembatipnc Uneymmetieal Semele Canter [Urtorm Seeton Cantever Leon Y= brant iy = be Tie fey a @ Pre Ey r e | i Is. = op { ar \ zi i : Ee so IS ~ ar BI 7 aH ELE, ' =z @ f-t-Be [ts 3 ame So Swe om rs mean ae € cS Fig. 52—Design formulae for leat springs: Unloaded springs have a radius of curvature R which Ie eoneidered negative ‘when camber Is positive 1.38i i t windup stiffness also varies as the square of the ath. (See Chapter 6, Section 2.) in considering the load (P) and the rate (k), it must be atin mind that the test load and rate are not the same the desied installed load and rate. The effect of the tallation (shackles, etc.) is discussed in Chapter 6. Stitfening Factor ‘Actual leaf springs are not truly beams of uniform ength. How closely they approach such a beam depends jefly on the following factors: 1, Length of Leaves—Two or more full length leaves +sometimes used. The shorter leaves may be longer than ty would be for uniform strength. This is to reduce the tin leaf stresses in the area of the eyes. The spring will srefore be stiffer, and it will be intermediate between a iform strength spring and a uniform section spring. 2. Leaf Ends—Fig. 5.1 shows how the ends of the wves exceed the outline of the triangular beam. This tkes the spring stiffer. The various leaf ends and their sets are discussed in Chapter 3, Section 2. 3. Center Clamp—Standard procedure demands that liptic springs be tested without center clamp, and 2 formulae are given for this condition. Since the springs, + used with the clamp, the leaf lengths are designed for clamped springs. Theoretically, this requires subtract- ‘the length of the clamped past from the total length, jing two cantilevers for this reduced length, and add- 48 uniform section in the middle between the two can- vers. Actually, the effect of the clamp can be allowed c by using an “active length” in the formulae instead of 2 full length. ‘The amount of active length within the seat area de- ids on the design of the clamping parts and also on any er or insulator material which may be used between the imping plates and the spring itself. In semivelliptic and. ntilever springs, the active length excends into the seat ea and is therefore longer than the distance from the g¢ ofthe clamp to the point of applied load. For springs thout any liner material, the active length generally tends to the inside edge of the clamp bolt. Spring seat vers will increase the active length farther into the imped area, the amount depending on the compres + of the liner, when the clamping bolts have been tight- ed. ‘At one particular deflection the addition of the clamp vesnot change the load carried at that height. This is the ‘sition at which the curvature of the spring without amp fits the spring seat The influences of the lengths of leaves and the types of Af ends are taken into consideration by a “stiffening factor” which is designated as “SF” in Fig. 5.2 and Table 3.1. Note that the stress from deflection and the load rate formulae are multiplied by SF, except for the uniform section cantilever for which SF always equals 1.5. ‘The value of SF is exactly 1.00 when the leaf lengths and leaf thicknesses are selected to produce a uniform strength ‘beam (spring), and the resultant curvature ofthe spring in bending is a circular arc with radius The farther the design deviates from uniform strength, the farther will the elastic curve deviate from a circular arc. “R" will be only an approximate parameter for the clastic curve, and SF will have to be a higher value. SF will have its maximum value of 1.50 when all the leaves of a multilea spring are full length, or when there is a single leaf of constant thickness and width. A spring of this type is known as a uniform section spring (column 4, Fig. 5.2), where "Ris the smallest radius for the elasti ccurve, namely at the line of clamp or encasement, and “S' is the highest stress (at that line), as explained in Section 1 of this chapter. Selection of the correct SF value in the final spring design is predicated on factors gained from experience, since the value may vary from less than 1.10 to 1.50, depending on the design specifications. For the preliminary design calculation, the following SF values may be applied in formula I of Table 5.1 t0 provide a moment of inertia (1) for a selection of the approximate number of leaves and gages assenger car and light truck springs with tap- ered leat ends and more or less “nniform” stress design: SF = 1.10 For passenger car and light truck springs with tap cred leaf ends and extended leaf lengths: SF = 115 For truck springs with untapered leaf ends and more or less “uniform” stress desi SF = 115 For truck springs With untapered leaf ends and two full Tength leaves: SF = 120 For truck springs with untapered leaf ends and three. full length leaves: SF = 125 For the first stage of variable or progressive rate springs, before contact with the second stage: SF = 140 For the second stage of variable or progressive rate springs, when all leaves are operable: SF = 1.10 with tapered leaf ends, SF = 1:15 with untapered leaf ends For springs with all leaves full length: SF=1.50 139‘Table 5.1—FORMULAE FOR CALCULATIONS ON LEAF SPRINGS. Formula symmetrical Uneymmetical SemHENipie Semretipne ' 7 Sees emcee = ee “BE SEL BE SF Yat "2 SF - wer | as 4 ¥ EP m ‘Stress with F P standard gage. Eas (teat — MPa Tri mats om Table 53 1v | Approx Mass of] (The numberof eaves ofthe same gage) - UME ASS Hom TAB 5 ona na of sping length - SF pepe 1000 "9 “Where te grading of the spring has several gages, the several products are additive before multiplying by (one half of spring length) ‘and (SF), The denominator of 1000 reduces the unit mass values to kg per mm. NOMENCLATURE FOR FIG. 5.2 AND TABLE 5.1 Stifening Factor 3. Preliminary Calculations Table 5.1 lists the formulae (1, II, III) for the three ‘essential steps in calculating the spring design. Formula I is used to establish the total moment of inertia for the specified rate and length. Formula Il is used to establish the maximum permissi- ble leaf thickness within the specified maximum stress limit ata corresponding load. The grading ofthe leaves in the spring is established by selecting the number of leaves, the leaf width, and the combination of leaf gages to pro- Vide the calculated total moment of inertia obtained with formula I. The moments of inertia for the individual leaf sections are shown in Table 5.2 Length of semi-etiptc spring = am a length of cantiever spring mm a front length of somialiptis spring mm = rear length of somiatipte spring = mm y cantiever length rato in semi-elipic spring (=b/a) P toad 6 spring -N ' sptng detection = mm k ‘oad rate (change in oad / change in deflection) = Nim {otal moment of inertia (summation forall he leaves in the spring) = ext summation of the moments of inetia of those leaves comprising the section for which the stress is to be = emt calculated ‘maximum leat thickness for maximum stress < nm thickness of the lea selected rom standard gages (t < t.,) fr which the stess i 1o be calculated: ucually —— mm this is the mair lea and/or the leat ofthe greatest thickness in the sping maximum speciid strass (N/mm= = = MPa ‘Suess with selected standard gage MPa ‘modulus of elasticity (or stool: 200. 10°) = MPa Fadus of curvalue in the unloaded spring (shown in dash tines), considered negative mm Formula III is used to calculate the spring stress. For preliminary calculations of unsymmetrical springs in which the length ratio Y does not exceed 1.30, the symmetrical formulae in Table 5.1 may be used with as- surance that the results will be within 3% of those obtain- able by use of the more complicated unsymmetrical for- molae Formula IV is used to estimate an approximate mass of steel for comparing optional designs from preliminary cal- culations. ‘The exact mass of spring steel is calculated after the eye diameters and the leaf end constructions have been estab- lished. Generally, the exact mass is heavier than the ap- proximate mass from formula IV. 1.40TABLE 5.2—MOMENTS OF INERTIA (mm) FOR LEAF SPRING BAR CROSS SECTIONS OF MEAN DIMENSIONS PER TABLE 3.1 IN CHAPTER 3 (TABLE 1 IN SAE STANDARD J1123) Tike Moments OF ineria For Width OF ness | OO] S|] 80 | 70 | bod] oo |S To S00 | aera | aan | S003] “s6rz | x2 5% | aro | saa | ss60 | 6506 | 7599 360 | 3800 | cars | 7000 | 7000 | oss sco | esx | rea | asas | 701 | rose | soe | 1507 s20 | 7a | ssea | 1000 | 1123 | 1266 | sso | 1015 670 | 9560 | vo7e7 | ras | 1950 | sez | ter | 21a nao | s1373 | i200 | iea0 | t608 | ser | aise | 2600 | 2582 750 | tsz99 | ise | ieee | tase | atc | ose | 065 | s4n0 aco | tsz2o | weave | 201 | 220 | asec | aoe | 3720 | ata aso | “toe | "zie | 2ea5 | are | vor | a7or | ease | aooe | oaiz sco | 201 | 2600 | 2000 | 9259 | 67a | sa90 | 5205 | 5002 | 7376 sso | 2700 | cose | se0s | once | aazr | 5170 | 622s | ooze | 9077 soo | 8136 | asa | aos7 | eso | sooe | so06 | 7222 | 056 | 10080 | 12080 jos | sree | ats | e7oe | Soo | seve | 71ar | ase | 9572 | 11900 | 1s60 tz | aso | ser | ssa0 | oat | 7000 | sere | rover | sizer | t4190 | to050 tio | S11 | svar | bees | are | doa | sez | trees | tor95 | 16520 | 19860 taco | cose | seco | rose | aoa | orss | 11670 | 1407s | 1567s | 19600 | 20590 1320 voce | rors | sse60 | 1370 | 12580 | sess0 | 2aro0 | 27700 1400 torr | s2060 | izeso | 16250 | 19720 | 21990 | 27sa0 | soso 1500 rarz | t700 | s8r40 | 20070 | 24200 | 27010 | 33000 | aov70 1600 15aso | 1700 | 20250 | 24200 | 25060 | 32730 | 41120 | avo 1700 zraro_| 2az10 | 9080 | 95150 | s5200 | 4s270 | 59220 +800 asz00 | 200s0 | deeao | ares0 | eeuso | 58400 | 70300 1900 zo620 | 32500 | aoaro | aa600 | sase0 | 8650 | 62720 2000 seeio | 3000 | arom | seor0 | casio | e0000 | s64t0 2120 sseo | s7e10 | 7570 | asso | 114700 Zao ssi | 79210 | aeaso | 111000 | 194000 no ye190 | s2a10 | s0a220 | 120000 | 156600 2500 0230 | r09500 | 122400 | 154000 | 18000 2650 tes100 | 183400 | z2r900 ze00 71000 | 215000. | 260700 3000 200500 | 254000 | 20100 m0 306000 | 370000 x50 67000 | 44100 3550 527900 750 20100 For derivation of formulae for these tabulated values, see Appendix First Example Design a symmetrical semi-elliptic passenger car spring with tapered leaf ends to meet the following specifications: Rate, as tested without center clamp. k = 175 N/mm Design load P = 3500N Metal-to-metal clearance x, = 110 mm Length L = 1320 mm Maximum stress Sys = 1000 MPa Step One—Find the required total moment of inertia from formula I in Table 5.1 ee eeiioeious = NTS 1320 5117 mm 3E 200-108-110 = 5717 Step Two—Find the maximum permissible leaf thick- ness from formula IT in Table 5.1: 141 with Pray = 3500'4 17.5 - 110 = 5425°N _ 8: 5717-1000 $5717 1000 _ 6.39 mm 1320 - 5425 (nan Thinner leaves would give a lower stress, thicker leaves a higher stress. The nearest standard gage size is 6.30 mm, producing a 114% lower stress. Step Three—Find the Number of leaves and the leaf width (w) which will produce the required rate; using the moment of inertia values in Table 5.2: First option: assume w = 50 mm Number of leaves = 6 4 leaves gage 6.30 1 2 leaves gage 6.00 1S28 81756 N/mm k=175 3717 With this selection of leaf gages, the rate k is within less than 1% of the specified rate. The stress at Pinu, (5429N) from formula III 1520 630 . 5429-985 MPa be 8.5729 Approx, Mass = 4:2436.4 2. 2302 1000 B20 1.10 = 1045 ky Second option: assume w = 63 mm Number of leaves = 5 2 leaves gage 6.30 2532 3 leaves gage 6.00 I = 3280 31 = 5812 mm* 5812 = 175-32 - i779, k Sry 7 1779 N/mm With this selection of gages in the 63 mm width the rate k is within 2% of the specified rate. The stress at Py, (5457 N) Spa, = 1320: 6:30 8. 5812 2. 3.076 + 3-2.930 1000 5457 = 976 MPa Approx. Mass = - B20 110 = 10.85 kg Conclusions—Both options appear acceptable. The choice may depend on some of the following considera- tions: 1. Limited available space for the spring width will favor the first option (50 versus 63 mm width); 2. Restrictions on mass will favor the first option (10.45 versus 10.85 kg); 3. Mandate on low fabrication cost (6 versus 5 leaves) will favor the second option; 4, Demand for the greater lateral stiffness will favor the second option (63? versus 50° in the lateral rate for- mula—see Table 10.2) Unconventional Springs (See Table 4.1) Some semi-elliptic springs may require that one can- tilever end is made of more leaves than the other, or that its leaf lengths are extended substantially beyond the uni- form strength beam requirements. In such cases it is more ‘convenient to consider the complete semi-elliptic spring as ‘made of two cantilevers. The rates or deflections of both 1.42 cantilevers are calculated separately and combined by use of the following formulae: PL th + ah kykgl? Kea? + qb Load on spring Total spring length Fixed (“front”) cantilever length (including clamp length) Shackled (“rear”) cantilever length (including clamp length) = Deflection of front cantilever Deftection of rear cantilever Front cantilever rate Rear cantilever rate = b/a : Z = ky/ky Second Example Redesign the second option (w=63mm) of the firs example with the additional provision 10 provide ‘=0.030 deg/mm control. This will require an unsym- metrical spring desig. where W=Z/Y? does not equal one. (See Table 4.1B) Try front end length a = 580mm Rear end length b = 740 mm > _ 740 Length ratio Y = 2 = 29 = 1.276 ath ratio Y = 2 = Cantileyer rate ratio = ky Ly. 573: (V+) + OLY cane ky 373-4) ~ OL 2 = 1216. STMLZTO+1) + 0030.INO.276 57.3(1.276-+ 1) — 0.030-1320 za2s0 = e+) _ (Z-Y)-(Y¥+) (2: 1320(2.542.+ 1.2762) Q=1910 573 23 = 00: ‘mm a 30 deg/mm as 425.740 P, max = 3041 N 1320— 5425.580 Py max = Tp 7 AN So MZ+¥9) _ 17.50.54241.2762) Crag (127641 14.09N/mm 14.09 ky BS = 5.54N/mm % Soap 7 SANK Step One—From formula I for cantilever 6248mm* 5.54» 740? 2- 200-108 Bh 5102 mm* Step Two—Under a given load the rear cantilever will have a higher stress than the front cantilever, because of its smaller EI Therefore, tgmax will decide the selection of the standard gage to be used. From formula II for rear cantilever 25102 _ 1000 740” 2384 Q ‘Step Three—Due to eye strength requirements it is not advisable to use a gage thinner than 6.90 mm in the rain leaf. The maximum stress will then slightly exceed 1000 MPa in the rear (shackled) cantilever. This should ‘be acceptable a long as the stress in the more critical front (fixed) cantilever will remain below 1000 MPa. Establish the number of leaves and their gages in the specified width of 63 mm max 5.78 mm Front Rear Cantilever Cantilever Leat Gage 1 1 1 6.00 1099.4 1099.4 2 6.00 1093.4 1095.4 3 6.00 1093.4 1093.4 4 6.00 1099.47 1093.4 5 5.60 808.9 8889 6 560 888.9 none ST = e1stamn? 6262.5mm* If Leaf =5 were of gage 6.00 the front cantilever31 = 6355.9 and the rear cantileverSI = 5467.0 and since the required I for both front cantileverS1, = 6248 and rear cantileverS1, = $102 is between these two’ selections of ‘gages, actual tests for spring rates will establish the ‘Bifes 10 be used. For this analysis, it will be assumed that the checked rates with the 5.60 gage in No. 5 leaf will be satisfactorily close to the specified rates. 1.43 5, max = 580-60 304) = 860 MPa 2 61Sa S,max = 40-690 9384 = 1006 MPa F-52025 Estimate of approximate mass 42930 + 2- 2.736 Front Cantilever = 472950 42-2736 M . 1000 sa 1.10 = 5.484 kg 4.2930 + 12.736 Rear Cantilever = 4£+2930 +1 -2.736 ear Cantilever, i 0 110 = 5.884 kg. Total = 11.37 ke ‘A spring in which the specified control ® requires an ‘unconventional design will have more mass than a spring in which Z/Y? equals or nearly equals 1. This is verified by comparing the approximate mass of the Second Exam~ ple (11.37 kg) with the mass ofthe First Example second option (10.85 kg). Tt is always advisable to make sample springs in order to verify by test results the grading (gages) selected before issuing the spring design Third Example Design a symmetrical semi-elliptic truck spring with 3 full Tength leaves and no tapered leaf ends. Rate as tested with an ‘SAE loading block k = 110. N/mm Design load P = 14700N Stress at P (not to exceed)’ S, = 540 MPa Length L = 1460 mm width w= 75mm Step One—From formula I calculate the required mo- ‘ment of inertia' with the heavy truck SF value = 1.25 wa ve FF 14602 i= OE 1460 = 42792 mm* 32 - 200: 10? - 1.25 ‘Step Two—Find the maximum permissible gage thick ness from formula II S EMM 0 gg am Twa" Ta700Step Three—The nearest thinner standard gage is 8.50. Since 12 leaves gage 8.50 in the 75 mm width has aBI = 44484 which is too large, a combination of gages 8.50 and 8.00 will be selected. 9 leaves gage 8.50, 3 leaves gage 8.00, With this selection of gages the rate k should be very close to the specified rate. Stress at design load 14700N 1460 - 8.50 14700 = a 700 = 535 MPa Approx. Mass 94946 + 3. 4.657 1460 125 = 53.4 i 1000 2 oa Fourth Example tis interesting to note the difference in calculated data which would result in.the Third Example, if the specifica tions had not required the three full length leaves. \ The stiffening factor would te SF = 1.15 ‘Step One-— Ho. 1460) O16 45519 am! 2 32-200-10) 1.15 ‘Step Two— tne = 288819, $80 936 mm Tae” Ta00 ‘Step Three—The nearest thinner gage is 9.00 mm 8 leaves gage 9.00 I 35192 3 leaves gage 8.50 1 ua 1= 46313 mmé With this selection of gages, the rate k should be very close to the specified rate. Stress at design load 14700 N 1460 9.00 14799 8. 46313, S21 MPa 1.44 Approx. Mass = 825.235 + 34946 1460 5 1g = 1000 41.6 kg In comparing the Third and Fourth Examples, it is evident that due to the more uniform stress distribution throughout the length of the spring in the Fourth Exam- ple, the mass of the Fourth Example is less (47.6 kg versus 53.4 kg). However, where safety requires lower stresses at the bearing ends (eyes) with more than one full length leaf, the greater mass and increased cost of manufacture (12 versus 11 leaves) must be accepted. Summary ‘The four examples show that the formulae in Table 5.1 are sufficient for preliminary calculation, The total mo- ‘ment of inertia required for a specified rate (the most important characteristic of any suspension spring) at a specified spring length is determined in formula I. It may be satisfied by a variety of combinations of width, thick- ress, and number of leaves, provided the leaf thickness does not exceed the maximum calculated in formula II for specified maximum stress. This is checked in formula IIT in which the total moment of inertia for the chosen stan- dard leaf size is incorporated. 4, Stress Distribution When the types of leaf sections, ends, clamp, etc. (Chapter 3) have been chosen and the approximate spring dimensions have been determined as shown in preceding. sections of this chapter, further specifications are gener- ally developed in cooperation with the spring. manufac- turer. These specifications include thickness, length, free radius and peening of each leaf, and the amount of cold set to be given to the assembled spring, Spring manufacturers use various methods to arrive at these data, and their results may be different because they involve judgment in the compromise of contradictory re- quirements. In view of the compromises and assumptions involved in the detail design, service experience must remain the final test. Where sufficient background is available, service experience can be represented by the results of suitably arranged life tests (see Chapter 8), and in turn, life test results may lead to definite design rules for a particular type of service.The basic considerations which are presented in this se-sion will enable the user to follow the spring manufac K+ analysis, They will also help the interpretation of, service troubles in terms of spring design. Leaf Thickness Jn automotive practice, springs are usually “graded,” that is, composed of leaves of two or three different gage thicknesses. The main leaf, often together with adjacent leaves, is made one gage thicker, and several short leaves are made one gage thinner than the intermediate leaves. This is done for a number of reasons: to give the main leaf mote strength to resist eye forces: to allow more toler- ance of quench radius on the short leaves; to compensate for the difference in free leaf radii, and because desired rates can be approached more closely by combinations of standard gages than by using the same gage for all the leaves. Leaf Radii ‘The curvature is not the same for all the unassembled leaves of a spring. It becomes more and more negative or less and less positive from main leaf towards the shorter leaves. When the spring is assembled, the leaves are pulled up against each other and’a common curvature is estab- sd (which may, of eourse, vary along the spring). The eaves of an assembled spring in the free position are there~ fore under some stress. In the main leaf, this assembly stress is subtractive from load stress; in the short leaves, itis additive to load stress. This is done to reduce the main leaf stress and to insure that the leaf ends have bearing on adjacent leaves. ‘The quench radii or curvatures of the quenching forms are obtained from the desired curvatures of the individual leaves by allowing for springback, for the effect of shot pening, and for cold setting of the spring (See paragraphs on “Free Radi” in Section 5 of this Chapter). Stepping ‘The lengths of the leaves of a spring, together with the thicknesses and the individual leaf radii, determine the distribution of stresses along each leaf. They also control the shape of the spring under load and its rate. ‘The leaves of a spring bear on each other mainly on a relatively small area near the leaf tip. The center of pres- sure of this area (the “bedding centerline”) is some dis- tance behind the actual tip of the leaf. This distance ranges from about 10mm for blunt (full, thick) ends to about 50mm for tapered leaf ends. The distance from one center °* oressure to that of the next shorter leaf is the step or \shang. To determine the overhang of the shortest leaf from the edge of the clamp, itis necessary to consider the design of the clamping parts as explained under “Center 1.45 Clamp” in Section 2 of this chapter. The sum of all the steps equals the “active” length of the spring. To discuss the effect of stepping, a hypothetical imprac- tical spring in which all leaves are made withthe same fee radius will be considered first If ll leaves of this spring are of equal thickness, equal steps will give the closest approach to the beam of uniform strength. That is the condition shown in Fig. 5.1 If the spring has a mixed grading, that is, composed of more than one size of gage, the steps should be made proportional to. The streses wil be approximately uni- form along each leaf, but the various leaves will be stressed in proportion to their thickness, Which may not be the most desirable for fatigue life endurance. If several leaves are made full length, the spring will be stiffened (deflection reduced under a given load). For a given load, this will reduce the stress adjacent to the spring ends, while the stress near the seat area will remain the same, However ifthe applied load is increased to produce a given deflection, the stress near the seat area will be increased, Ifthe balance of the leaves are not extended in length with the full length leaves, the disturbance of efficiency will be confined to the full length leaves. However, if the balance of the leaves are also extended, th stresses will be less toward the leaf ends. The spring willbe stiffened still more, and it will deviate from the uniform strength beam in the direction of the uniform section spring. If the individual leaves have different curvatures, then assembly stresses are set up. These are desirable for vari- fous reasons. With assembly stresses it is impossible to have uniform steess along a leaf at all loads. By a suitable combination of assembly stresses and stepping, iti possi- ble to distribute the stress ina desirable manner among the leaves and to make it uniform along each leaf at one particular load. “Though the stress distribution can be improved by a proper combination of leaf thicknesses, free radii, and steps, the rate is the same as for a spring made up of the same leaves fitted “dead,” that is, without assembly stresses Fig. 5.3 shows the stress distribution along the main leaf ‘ona spring with and without assembly stresses, with three diferent leaf steppings, and at various loads. Iti intended to show only the principles discussed here. The spring itself is impractical. tis made up of two leaves of equal thickness, the mais leaf rectangular in plan view, the short leaf triangular to simplate the effect of a series of stepped leaves. ‘Without assembly stresses, the 250-250 stepping (equal steps) is obviously most efficient. Also, it has the lowest Tate and deflects with uniform change of curvature in the part made up of two leaves. The 400-100 stepping ap- proaches a rate and shape of the uniform section 5 With 300 MPa assembly stress atthe clamp, the rate is not changed, but the stress distribution and shape are changed.— MAIN LEAF fF ~ 250 ———-— 250 -—— —400 MPa STRESS IN MAIN LEAF WITHOUT ASSEMBLY STRESS (LOAD STRESS ONLY) P= 100 P=50N ps0 STRESS IN MAIN LEAF WITH 300 MPa ASSEMBLY STRESS AT CLAMP Fig. 5.3—Stress distribution along the main leat with and with- out assembly stresses The 250-250 stepping is now obviously inefficient, be- cause the main leaf has a peak stress at the tip of the second leaf. If the unassembled, leaves are circular, the shape under load will correspond to that shown in Fig. 5.4A. This condition corresponds to t stepping. With 325-175 stepping, the main lea will have uniform stress within the contact area with the second leaf under 100 N load and at that load will have a circular shape. At lower loads the shape will be intermediate between Fig. 5.4B and 5.4C. With 400-100 stepping, the main leaf will have uniform stress within the contact area with the second leaf under 50 N load and at that load will have a circular shape. At lower loads the shape will be more like Fig. 5.4B. In Fig. 5.4C the spring is shown flat under load in order to bring out the deviations from the desired shape more clearly when the backing leaves are too short as in Fig. 5.4A and when the backing leaves are too long as in Fig. 4B. In the following discussion the subscript “n” is used to denote stress (S), thickness (t), overhang (2), curvature @, etc, of “different” or “various” leaves. 1.46 To obtain uniform stress along the length of each leaf at a given load and to achieve circular or straight shape ofa spring made up of eaves which are circular when free, it is necessary that each leaf be subjected only to pure couples applied to its ends. This condition can be fulfilled only if the loads between the various leaves are all equal tothe load (P) on the end of the spring, In order to achieve this, the stresses (S,) in the different leaves are adjusted by proper choice of free leaf radii. At the same time the overhangs (,) of the various leaves are so chosen that at the particular load which is selected in advance as the load ‘at which the spring shall haye circular or straight shape, S, = 6 PLy/w th ‘Since P and w are constants, the overhangs 2., are made proportional to St, The stresses S, can be chosen, subject to the condition that the sum of the moments in the individual leaves ‘equals the total bending moment of the load. It has been found that lower stresses in the thicker main leaf and higher stresses in the thinner leaves give good results.A— BACKING LEAVES TOO SHORT | B — BACKING LEAVES TOO LONG C— LEAVES WITH St2 STEPPING (UNIFORM STRENGTH SPRING) Fig. £.4~Etlect of backing lea lengths on spring shape For “over-the-road” vehicles with a static deflection to design load of about half the total deflection, experience has shown that best service is obtained if the stresses are made uniform along the leaves under normal load. For “off-the-road” vehicles the stress at total deflection may be more important., Service experience and life tests must decide for which condition the spring should be designed. When the length of a leaf has to be different from the ‘value obtained from stress considerations (for example, to make room for a clip or to support the eye), the perform- ance and eficiency of the spring are affected least if all the ‘other leaves are made to the lengths originally obtained from stress considerations. ‘The change in spring rate due to changes in stepping can be estimated by comparing the overhangs as designed to overhangs proportional to t?. The spring designed with overhangs proportional tot? would have a rate calculated from the formulae in Table 5.1. A spring with all leaves full length would have a 50% stiffer rate. Te * Of Analysis, ‘Two different assumptions about the action of leaf springs are reasonable. They shall be called “point pres- sure” and “common curvature” 147 ‘Point pressure” means that the leaves touch each other only at the bearing points and at the center clamp. It is based on the observation of worn-in areas near the leaf tips. With this assumption, it is possible to calculate the tip loads of successive leaves from the condition that two contacting leaves must have a common load and a com- mon deflection. After the tip loads have been calculated, cach leaf can be considered as a simple beam. This method is used by some spring designers, but uniess the volume of work warrants the preparation of special curve sheets and leads to familiarity with the formulae, itis somewhat diff- cult, The assumption of “point pressure” is justified when the leaves are free to take the shape which corresponds to this load distribution. This is usually not possible for all leaves of a spring unless spacers are provided between them, “Common curvature” means that all the leaves of a spring touch their neighbors all along their length. This assumption leads to simpler calculations. It is in general justified at all points where a leaf is surrounded by other leaves, that is, everywhere except for the bottom and top leaves and except for the overhanging leaf ends. In some cases “point pressure” and “common curva ture” lead to the same result. For a really thorough analy- sis, it would be necessary to use each assumption only where it is compatible with its results. "Common curva- ture” alone is usually justified in the major part of the spring. Calculations on this basis will be very nearly cor- rect and are used here because of their relative simplicity. The following discussion, therefore, is based on “common curvature.” ‘A bending moment M will produce at any cross section ‘a change of curvature q=M/E 31,, where the summation includes the leaves at that section. The relation between curvature free and loaded is given by Qe +4 = oases ‘The correct signs for curvature must be introduced. Curvature is zero when the leaf is flat, and is positive under heavier loads (see the “Curvature” paragraph in Chapter 2, Section 1.) Therefore, on most springs the free curvature is negative. Numerically curvature i the inverse of radius, but the radiuS usually carries no plus or minus sign ‘Deftections can be calculated from q, just as for beams of variable cross section, by any of the methods explained in textbooks. Where only the deflection atthe point of load application is wanted, as in checking for rate the strain ‘energy method is the easiest of these. It is based on the equality between external work 0.5 Pf and internal energy in bending 0.5 f M q d2so that f=(5 M q aty/P.‘The load stress in any leaf at any cross section is Sp = @E y, where y is the distance from the neutral axis to the remotest fiber in stress. For SAE section leaves, y equals {t/2and the load stress then becomes S, = M1/2E1, where the summation includes the leaves at that section. The expression (231/1) is the section modulus. ‘The load stress §, would equal the total stress $ (where $ = S, + Sy) only in springs which are fitted “dead,” that js, where all leaves have the same unassembled curvature, ‘because in that case S, = zero. (S, = assembly stress in leaf) ‘in almost all leaf springs the unassembled curvatures qq are different in the individual leaves. In assembly a ‘common (unloaded) curvature q, is established which is Variable along the spring even if the leaves are made to circular arcs. The common curvature q, can be calculated, from the condition that the internal bending moments of all the leaves must cancel when the spring is assembled but not loaded. 35,1 = 0 do = Santh/Eth (weighted average) and the assembly stress S, can be calculated for each leaf at any section in the usual manner. Ey @y— 4) ‘The total stress $ isthe sum of the two stresses Sp and ,. The sum of the internal moments produced by the stress § equals, of course, the external moment Mat any section. $= 5,45, M = 58,2, where Z,is the section modulus flea. Ths last relation is implied in the previous ones, but may be used for re- checking “The design and analysis ofa leaf spring may therefore ‘be considered in four stages: Approximate estimate, aS shown in Section 3 of this Chapter. 2. Determination of leaf lengths and individual leaf radii from a desired stress distribution at one particular oad (see Section $ in this chapter). 3, Analysis of stresses and stress ranges at various points in the spring by common curvature or point pres- sure. “4. Check of the soundness of the assumptions in- volved in number 3 and analysis by @ combination of “common curvature” and “point pressure,” or by strain ‘gage measurements on the spring 5, Sample Calculation To design a spring for the following conditions: Rate, as tested without center clamp 17.5N/mm Design load 3500N Opening at design load (negative) =20mm Eye diameter, (front) upturned 40mm Eye diameter, (rear) upturned = 40mm Clearance (metal-to-metal) x, = 10mm Length center to center of eyes T= 1400mm Length front cantilever a = 570mm Length rear cantilever b= 830mm Width 6mm Maximum stress 1000 MPa Length of seat clamp 100mm Distance between inside edges of clamp bolts 75mm ‘Design the spring as a passenger car spring with tapered leaf ends except the main leaf, plus shot peening and cold setting, ‘Thickness of Leaves And 31 ‘Since the cantilever length ratio Y =1.456 exceeds 1.30, ‘unsymmetrical formulac of Table 5.1 must be used. 500 + (17.5 x 110) s425N Pras (at metal to metal position) Formule! wy 3 Yes 2 G+ SF - 14003 2 Formula IL pape +P "Peal Grading With Standard Gages ‘The nearest thinner standard gage is 6.70mm. Find required SI from Table 5.2 in gages of 6.70 or thinner in the specified 63.0 width. 1 leaf of 6.70 1 = 1522 thickness = 6.70 mm ‘3 leaves of 6.30 1 = 3798 thickness = 18.90mm ‘6413 thickness = 31.60mm 1.48LY BI = 6241 thickness Since 31 = 6413 is larger and 31 = 6241 is smaller than the calculated EI of 6360, only a test of sample springs made with each set of gradings will determine the selection for the spring, For this analysis we will assume that the grading of 1 leaf of 6.70, 3 leaves of 6.30 and 1 leaf of 6.00 has a test rate within the plus and minus $9% tolerance of the 17.5N/mm rate Stress Distribution Between The Leaves For calculation of stress, consideration is given to the seat clamp. The inactive length in the seat for a spring without liners is estimated as the distance between the inside edges of the clamp bolts, which is 75mm. The active cantilever length for both front and rear will be 37.5mm_ By using different free radii in each of the leaves, assem- bly stress can be added to or deducted from the load stresses of the assembled spring to obtain a more desirable stress pattern in the spring. The assembly stresses can be chosen more or less arbitrarily, except that they must be selected increasingly larger from main leaf to the short leaf, and the 3S,t? must be very nearly zero (S,=Assem- bly Stress) ‘A negative assembly stress in the main leaf is required to reduce the maximum stress (metal to metal position) to about $30 MPa. The longitudinal and lateral forces im- posed on the main leaf (with eyes) and its generally greater Stress range are the reasons for reducing its bending stress In the thinner gage leaves the stress can be somewhat higher because of their lower stress range. In view of the condition 38,1? = 0, several selections of deducted and added stresses for the individual leaves ‘may be analyzed before deciding on the best arrangement. les than the distance from center of eye to center of axle Leal cage = a seat. Te stress at 37.5mm from the center of axle seat can} GH ee = be calculated from 3 620 SoMa + 207 oe 2 620 yaaa + 988 Front Cantilever § = Patat 5 £00 a2hPa 4151 251 Sa? SBE whi Seopa Rear Cantilever $= Pofat ant The individual leaf stresses with these selected assembly \ esign Load == 3500N P, = 3500-230. = ag7sn stresses S, will then be: 1400 py = 3800. 57. ra2sw Front Cantiever Rear Cant i Lear [P=as00N [P= SH25R_ | PH3B00N | P= SHEEN + [S27 MPa | @4 WPa [520 MPa] 865 MPa a 2 | szoup | szoma | scoMPa | 348 MPo Cee ieee 3 | sssmpa | asimea | sesuea | s7oma + | sermpa | sssmra | syonea | eee ma Pymax = 3216N 5__| soma | eeampa | sromea | 901 we Pamax = 2209 &, = 510-375 = 532.5mm eae ‘The overhangs shall be arranged so that at design load hee eee the chosen stresses are uniform along each leaf. Ths im- io plies that loads between the leaf tips are all made equal at Front Cantilever $, = P+ 5325t _ goaispyt design load. The overhangs must then be made propor- aes tional to St : Py - 792.5 ‘Active cantilever length on front (hort) end is Rear Cantilever S, = Pe~7925t _ gootapy Gea oan . ‘The stress at design load (3S0ON) and at metal to metal position load (5425N) will be: eat | Gage | Front Cantilever: Rear Cantilover P=3500N | P=5425N [P=3500N | P=5425N | 870 | s77MPa | eaaMPa| S00MPa| —915MPa 2%.4| 630 | 54ampa| esimpa| ssmPa S| 600 | siz} sora | szavpa | stowPa 1.49 '532.5mm and on the rear (long) end is 792.5mm. St of the leaf vethang = Cantilever length - —St_of the leaf _ 7 : i tick ESt? of all the leavesront Cantilever Rear Cantilever teat | Gage | Front Total | Mass | Mase Pa. over | _WPa over ear | Bar | Bar | kg/m | of ‘eat_t_|ias00n) _st?_| hang | (95000) _st?_| hang Length | Length | Length Bar a ear Ber Tse] eo —PaeaT Tet ‘om_| mm _| mm kg 2 630 | s26 20056 | 1022] 540 avaaa} 522 —T—]-e70- “73070 1680] 270] Sat 3 630] 559 aro} 1070] ses esas} 1592 = 2 | aa0 | aze | 675 | t147 | 3078 | 3508 4 630 | 587 22804] 1097] 579 22981] 1692 3 | 630 | 57 | 50s | ez | 3076 | ese 5 600 | 550 _zorza| 981| 570 _zosgo|ras7_ «| 630 | 2az | 95 | 527 | 3076 | 1775 35%-=108100 ssf=1e00 5 | oo | 123 | 167 | 200 | 290 | 0679 The leaf lengths are obtained by adding the successive overhangs and rounding off. For the shortest leaf 37.Smm, per end must be added for the inactive length within the seat and at least 25mm for distance from bedding line to leaf end. If tip interliners are specified, the bedding line will be the &.of the tip interliner and the distance to leaf end will be measured from & of the interliner. In this example 25mm will be used. eat Theoretical Lengthe Selected Lengths No. Front Rear Front Rear 1 '70.0mm” ——830.0mnm" | B7Omm™ Boma 2 479.6mm ——6829mm | SO5mm 725mm 3 3774mm ——530.7mm | 390mm 555mm 4 274mm —a71.smm | 275mm 985mm 5 180.7mm__2083mm | 1@5mm_215mm, “The Bedding Line of the Main Leat isthe tof the Eye. Caleulate Mass of Steel in Spring \ Spring manufacturers have established (based on their processing equipment) amcunts in bar lengths required 10 be added for making eyes and to be deducted for tapering leaf ends. For this Example, Eye Requirement = 3 ID. of Eye + 3+ Gage Thk. or 3- 40 min + 3-6.70mm = 140 mm at each end of main leaf, Assume tapered leaf end thickness te=2mm and a length of taper TL=100mm at front end and TL= 150mm at rear end, extension of leaf end (Ext) due to taperinig operation may be calculated by formula Ext.= 049 TL—0.51 TL te/t. The Ext. is deducted from the finished (specified) leaf length to obtain the bar length required before tapering. For leaves of gage 6.30 Ext. =0.49(100)—0.51(100)2/6.30: Ext, =0.49(150)—0.51(150)2/6.30: For leaves of gage 6.00 Ext =0.49(100)—0.51(100)-2/6.00: Ext, =049(150)—0.51(150}2/6,00: Mass of each leaf will be calculated based on mass per ‘meter length from Table 5.3 (see page 52). Free Radi “Note: This section and the next section on, “Free Radii Including Shot Peening and Cold Setting,” are based on the premise that the free arcs of the individual leaves are 1.50 Total mass of steel bars = 14.328 kg formed to a circular shape. However, it should be pointed out that in practical spring manufacturing this is rarely true, because the overriding importance lies in the desire to secure an outline for the assembled spring which con- forms with the specified main leaf contour at “straight” main leaf (see “Seat Angle” paragraphs in Chapter 2, Section 1.) ‘The overhangs were developed so thatthe tip pressures at centerline of bedding are all equal at design load and the stresses in the leaves are the values calculated in the pres ‘ous section on “Stress Distribution Between The Leaves.”" ‘The free radii can now be determined from the condition that the change in curvature from free (unassembled) to design load must produce these stresses. Then the as- sumed conditions will be fulfilled. The camber at design load is calculated from the given ‘opening at design load, which is —20 mm., and the eye diameter, which is 40 mm). ‘Camber = Opening — % eye diameter -»- 2 — 40 mm Camber is converted to curvature according to the for- mule 8 X camber Curvature matt Curvature is called positive in the direction of inereas- ing load; camber is conventionally called positive in the opposite direction, therefore the minus sign. ‘The curvature at design load is (with a negative camber of —40 mm). +0.000163 ‘The change of curvature corresponding to a stress “S" is 28/Et and with E = 200 000 MPa gives the following free curvatures: Free curvature 4Front Cantilever ear Cantilever Let | ot Free Free Now| mm q__| mar | _asver 4, Radius | pat | _2svet a, Radius + Pero] amoares | 527] a 007a7 | —ara00EsE | T603 —]— $40} ~T.ddoRos | —TedEEAT | 1555 2 | 630 | oooo16a | 528 | 0.000838 | —0.000675 | 1481 | 540 | 0.000857 | —0.00060¢ | 1441 3 | 630 | oooo1e3 | $53 | 0.000878 | —oo0071s | 1399 | s85 | 0.000897 | —ooc07s« | 1362 2 | 630 | o.000163 | 567 | 0.000900 | —o0o0rs7 | 1357 | 579 | o.o0o19 | —o.000756 | 1320 5 | 600 | oooo16a | s59_| 0.000932 | -o.000769 | 1900 | 570 | 0.000950 | —oocove7 | 1271 *Siress at design load 3500 N ‘without any shot peening of cold seting LUnsymmettical springs require radi of diferent values forthe front and rear cantiovers (see above). however. for cantilever length ratios (of '958 than 1.30, a uniform radius may be used. Free Radii Including Shot Peening and Cold Setting ‘The radii of the quenched leaves are obtained by includ- ing the changes produced by shot peening and cold set- ting, Peening changes the curvature (in the direction of the change produced by load) equivalent 10 Cold setting also changes the curvature (in the direction of the change produced by load): = aay ¢ = average curvature at maximum set-down position (where “camber” is the arc height at this position) __ 8 camber assuming the spring is to be set down 230 mm beyond metal-to-metal position, _ 8( =230— 110-40) 00155 100 +0.001551 % 4, = (refer to the previous section on “Free Radii”) 014 t = curvature change due to 1400 MPa (stress at which cold setting begins) quench curvature (or the inverse of quench radius), when spring is to be shot peened and cold set =a -~%—% Note: Both shot pening and cold setting add positive curvature, therefore the quench curvature (q,) must be a larger negative value than the final free curvature (q,). Tength? 0.014 quench Leat a t a & a % radius Front Cantilever T OOOTEBT | —CUOROGO | ToMRE |“ O.BoGET—oDaHURS |= TUG TET—17T 2 001851 | 0.002222 | —o.000875 | 0.000078 } 0.00000 | -oc007ss | 1925 3 001551 | 0.002222 | -0.000715 | 0.000078 | 0.00004 | -o.0080s | 1198 4 o.001ss1 | 0.002222 | 0000737 | 0.000076 | 0.000066 | -o.00c879 | 113 5 001551 | 0.002353] —0.000769_| 0.000089 _| none =o0008s2 | 1174 Rear Cantiover 7 SOOTEST-] — OORT] —ODOOEIT | OwDOORT | —ToDaTOT | —ooo0eTa | Tee 2) c.0015s1 | 0.002222 | —o.00060 | 0.00007 | o.ooes | 0000793 | 1261 a) ‘.001ss1 | o.oozez2 | 0.000724 | 0.000076 | 0.000063 | —o.000873 | 1145, 4 co.oo1ss1 | 0.002222 | 0.000758 | 0.000078 | o.0008s | -o.000017 | 1091 5 o.oorss1 |_0.002333 | -0.000787__| 0.000083 | 0.000005 _| 0.000875 _| 1149, ‘Wen shot peering is not specified, replace q, with. = Quench cuvature whan spring isto be Cold sot only (that is, when shot pening is not specie) a “Tho actual spring caleulaton, of course, takes much lass time and paper than the sample on this page, which contains much text. This is pariculary tue it tho calevation is catied out according to @ definite pian 151‘TABLE 5.3—MASS OF LEAF SPRING BARS (kg per m length) OF MEAN DIMENSIONS PER TABLE 3.1 IN CHAPTER 3 (TABLE ‘YIN SAE STANDARD J1123) Thick Mass For Widih OF ness a a 500] 15ea | iras [tsar [etm | 24a 5.20 sae? | 2053 | 2301 | 2580 560 xgso | 2169 | 2430 | 2706 8.00, oes | 2322 | 2603 | 2990 | 3407 | 4198 6.30 zise | 248 | 2732 | 2076 | ser | aco7 670 2090 | 250 | 2904 | 3270 | 3e00 | aes7 710 2as7 | 2764 | 3076 | 3464 | 4196 | 4966 | 5520 7.50 2601 | 2007 | 3248 | 3658 | 4368 | 20s | 5030 8.00 277s | 3087 -| 2462 | 2009 | 4657 | 5500 | 6210 850 295 | 3277 | 3675 | s140 | aoe | soar | eos | 270 goo | 276 | ais | sas7 | aaao | a2 | s2as | 6289 | 6901 | arse 950 | 2919 | 3205 | 3656 | 4101 | aozr | 5523 | 6835 | 7377 | azar 1000 | 3061 | a4s: | a2 | arr | aas7 | 5001 | 6971 | 7750 | 0703 | 1166 soeo | 209 | 3654 | S063 | a565 | sia | ores | 7988 | azi2 | 1028 | 1256 1120 | 3417 | aa55 | 4203 | aaa | sasi | 649 | 7000 | 67s | 1086 | 1306 sveo | 3505 | 4056 | a5i7 | soo | 576 | sexe | za | o1ss | tras | 1375 x2so | 3e01 | 4260 | 4773 | 5361 | cose | 7222 | sexe | sere | i212 | 1457 1320 5038 | 5657 | 6980 | 7630 | 9176 | to21 | 1273 | 1538 1400 saat | soot | 6757 | ose | 9725 | toge | 1258 | 1631 1500 s702 | 6406 | 7228 | 8650 | 10400 | 1158 | 1459 | 17.87 16.00 6069 | 6820 | 7696 | 9215 | 11091 | tase | 1548 | 1863 17.00 za | area | o777 | t177 | 1310 | 1644 | 1979 18.00 zen | ser | 103e | i245 | 1396 | 1740 | 2094 19.00 sos | 9090 | 1090 | 1313 | 1461 | 1895 | 2209 20.00 eas | oss: | 114s | 1380 | 1598 | 1930 | 2324 2120 viz | 1461 | t627 | 2040 | 2a62 2240 w276 | 1539 | sta | 21a | 2592 23.60 1342 | 1619 | woe | 2266 | 2729 2500 vata | 1712 | 1908 | 2398 | 2589 26:50 2019 | 2599 | 1060 200 2130 | 2680, | 3290 30.00 zim | 2e6s | 3456 2150 200s | 3626 3350 sist | 3651 35.50 4075 37.50 4298 ‘Note:—When using this table, divide by 1000 to reduce the length in meters to millimeters. For derivation of formula for these tabulated values, see Appendix 8. 1526. Variable or Progressive Rate Springs << Appopular type of spring, particulary in the light truck Nodels, isthe variable or progressive rate spring design This is where a relatively low rate is desirable when the vehicle isin operation with only the driver and perhaps a light payload, bu a higher rate is quired when the vchi- cle is at design loud Variable rate springs have the potential of improved ride quality over a wide load range, provided an adequate total defection can he accommodated on the vehicle, and the design specifications assure manufacturing controfs to obtain the desired spring rates Tn the light load range, the second stage leaf or leaves are inoperative. As the payload increases. the frst stage, by rolling contact gradually engages the second stage When complete contact is made between the fist and second stages the spring is then operating inthe high rate range. Fora variable rate spring to produce the ride gualites desired, the specifications must clearly establish the type of load-deflection curve as shown in Fig. 2.9 with the inital contact between the two stages at a load greater than the curb load, and the complete contact at a load less than the design load. Where itis customary for constant rate springs to specify only a design load and height, itis necessary for variable rat springs to specify the curb load snd height forthe fist stage and also the design load and \ ight for the second stage. It is recommended that the @ OF sPG desired curve be drawn, showing a gradual transition from the first stage rate to the second stage rate, and that this, curve be made an integral part of the specification, Fig. 5.6 shows both the load-deflection diagram and the rate-deflection diagram for the Sample Calculation which follows. It should be understood that the rate-deflection diagram cannot be considered to be an exact depiction of the diagram obtained from an actual spring. It has been drawn without considering such details as the change in moment arm length when the spring is deflected from positive camber to flat 10 negative camber. Sample Calculation Design a variable or progressive rate spring for these conditions: 63.0 mm 1320 mm 2200 N 6000 N 25 N/mm Width Length between eye centers Load at curb height Load at design height Rate at curb or light loads Rate at design or heavy loads. 43. N/mm Metal to metal clearance = 100 mm Deflection between P, and P; loads = 120 mm Length of inactive spring seat = 80 mm [Note that the deflection between curb and design loads is the difference in heights between curb and design load positions. I= 13733 Fig. 5.5—Leat lengths determined from moment of inertia dia- ‘ram 153LOAD (NEWTONS) 10 000 8.000 6.000 4000 2000 8 838 8 RATE (N/mm) LOAD-DEFLECTION DIAGRAM LOAD AT M/Mfi0 300 INTERMEDIATE | LIGHT LOAD RANGE | LOAD RANGE| HEAVY LOAD RANGE fe— kt = 25. N/mm 43 N/mm —>} WY CURB LOAD, 2% 200 (P1) as SPRING DEFLECTION (mm) ZERO AT DESIGN LOAD 1 1 1 \ 1 ' RATE-DEFLECTION aioe: 1 1 1 INTERMEDIATE Je LIGHT LOAD RANGE- LOAD HEAVY LOAD RANGE RANGE Fig. 5. —Variable rate spring: Load-deflection and rate-detiec- tion dlagrams 1.54All leaves except the main leaf are to have tapered leaf ends. his spring will be designed assuming that the front and [Raf length ratio Y=b/a is 1.30 or less, With this condition the symmetrical formulae from Table 5.1 can be used. Step One—For the first stage, consideration will be given to a three or a four leaf design. Because of the extended lengths of these leaves when operating without ‘contact with the second stage, an SF value of 1.40 will be used in the design calculations, as explained in Section 2 of this chapter. Calculate the moment of inertia requirement for the first stage leaves. From Table $.1 formula I 1320)? 32-200-10°1.40 Selection Of Gages For The First Stage—Three leaves, ‘of 750 gage have 31 = 6402 mm* or, considering a four leaf design, one leaf of 7.10 gage and three leaves of 6.70 gage have 31 = 6377 mmt. Either the three or the four leaf design should produce a first stage rate close to 25 N/mm. The three leaf design of course would be more economical. Step Two—Calculate the moment of inertia require: ont for the second stage leaves. Consideration will be given to two leaves and when selecting the four leaf design for the first stage there wi bea total of six leaves in the spring. Since all six leaves will bein active operation after the first stage makes complete contact with the second stage, the SF value will be 1.10 (Gee Section 2 in this chapter.) = 6417 mm* From Table $.1 formula I kL? __ 4313207 “ Jy= LE 43.0320) _ _ 14048 mm‘ Th= ESF ~3a20010%110 ~ ‘The moment of inertia requirement for the two second stage leaves will be the difference between 31 and 31; or 14048 — 6377 = 7671 mm, Selection Of Gages For The Second Stage—Two leaves of 9.00 gage have a moment of inertia of 7356 mm‘, Add- ing this moment of inertia to the four leaf first stage as selected under the previous section, “Selection Of Gages For The First Stage,” the total moment of inertia will bbe 13733 mmt. This is within 2% of the calculated 14048 mm*. It is recommended that sample springs are made and tested to verify the selected gages. Length Of Leaves With the gage of all the leaves established, the moment of inertia diagram can be drawn for the spring, see Fig. 5.5. The length of leaf No. 6 will be determined by its mo- ment of inertia relative to the total EI plus the inactive spring seat length (14 - 80 mm) plus the length at the ends for bedding (25 mm). 2 (2 ary 02040425 Length of No. 6 leaf = 2(166 + 65) = 2-231 mm ‘The length of leaf No. 5 will be determined in the same ‘manner, where the moment of inertia I will equal 2(3678) =7356 mm 7356 13733, Length of No, 5 leaf= 620465 2(3324.65)=2:397 mm For eye support the length of the No. 2 leaf will extend tothe center line of load application with the bedding line between leaf No. 1 and No. 2 at 25 mm from the end of the leaf The length between No. 2 leaf and No. 5 leaf is 660 — 397 = 263mm. ‘The stepping between leaves No. 2, No. 3, and No. 4 will be established proportionate to their individual mo- ment of inertia values. Since the moment of inertia of all three leaves is the same, the stepping will be equal or 263 + 3 = 87.67 mm. The leaf lengths are as follows: 1_| Front Ena Total eT Tae mi 3333 | 660 1320 mm ess | 572 572 | 144mm ea77 | 408 42a | 968 mm ross | 397» | 397 |}. 794mm saraa | 201, zar__| 462mm Stresses Stress calculdtions for variable rate springs are more involved than they are for uniform rate springs where all the leaves are operative under all loadings. Stresses in the main leaf will be calculated at the follow ing sections from which a stress diagram can be drawn, See Fig. 5.8, 1.55‘Section Moment Arm | 1 Stress (MPa) in Main Leaf at Load of mm F008 3000 N 6000 N TOROS 4 ‘CURB CONTACT DESIGN MoM BO=BEF ETT TT Ta 787 38 320 8 2855, 162 220 a 757 c ear? 7 20 81 826 o 677 380 518 5 = 2 19703, = s «7 982 For moment arm lengths see Fig. 58 “At loads of 2200 N and 2000 WN only leaves 1, 2.3, «are operative (I = 6977) STRESS—LOAD DIAGRAM FOR MAIN LEAF AT LINE OF ENCASEMENT (SECTION D} 1.000 982 9 ATM/M 900 800 700 3 3 37 ATDESIGNLOAD £ 600 3 al ‘590/AT COMPLETE CONTACT * . z g 500 518 AT INITIAL CONTACT a 6 7 3 400 4 v BOATCURBLOAD 2 3 i co '293 AT 1700 N 3 = 4S 200 2 8 CURB LOAD 1.000 eff: 7.000 8.000 LOAD (NEWTONS) Fig. 5:7—Varlable ate spring: Stress-load diagram for main lest at line of encasement 1.56 9000 10000(ness mea) 900-—:--- AT 10300 .N |_}-— Line oF ENCASEMENT ¢ f+ 332mm ——— |_| 620mm = * INDICATES MOMENT ARM Fig, £2 Variable rate spring: Stress detrbution long main leat under diferent lade 187Sample stress calculations Sa(at 2200 N) ‘Sp(at 3000 N) = Scat 6000 N) = S00 288 3000 - 620 7.10 3008 518 MPa 4 6377 Ps Spfat 3000 N) 1 The stresses at line of encasement (section D) for loads between design position and metal-o-metal position can be calculated by establishing the amount of load on the spring which is not supported by the second stage leaves after complete contact is made between the first and second stages. The load on the spring at metal-to-metal position equals 6000 N (design load) + 43 N/mm - 100 mm (x,) = 10300 N In Fig. 5.6, the dash line extension of the second stage te line intersects the zero load at a position of minus 140 mm. At this position the frst stage leaves have supported 1700 N (68 mm-25 N/mm); this load is not transmitted to the second stage leaves, The stress for that 1700 N load unsupported by the second stage leaves is: 1700 - 620 - 7.10 = 293 MP: 4-637 " At metal-to-metal position the stress for the load sup- ported by all the leaves is (10300 — 1700) - 620 - 7.10, 4.13733 = 689 MPa ‘Thus the total stress at metal-to-metal position equals: 293 + 689 = 982 MPa Similarly, the total stress at design position equals: (6000 — 1700) - 620 - 7.10 13 + {6000 — 1700) - 620 «7.10 ae 4 13733 = 293 + 344 = 637 MPa Figure 5.7 presents a detailed diagram of the stresses in the main leaf at line of encasement (section D), ‘The stresses in leaves No. 2, 3, 4 are 6% lower than the ‘\~ stresses in the main leaf since their gage thickness is 6.70 mm compared to 7.10 mm in the main leaf. The ratio of these gage thicknesses is 6.70/7,10 == 94%. 158 Since 1700 N is not supported by the second stage leaves, the stress in the second stage (leaves 5 and 6 with 9.00 mm gage thickness) can be calculated by using the transmitted load of 10300 — 1700 = 8600 N and a mo- ‘ment arm length from line of encasement to bearing line between leaves No, 4 and 5 (section C). This equals the length of No. 5 leaf (397 mm) less the inactive seat length (40 mia) and less the amount of bedding length (25 mm), thus equaling 332 mm. ‘The moment of inertia (I) for the two second stage leaves is 7356 mm’. Then at meta:to-metal position, here the spring load is 10300 N, the stress in leaves No. 5 and 6 is: (10300 — 1709) - 332 - 9.00 4. 7356 873 MPa, Note that these calculated stresses do not take into account any nip between the leaves which for the main leaf ‘would slightly reduce the stress from the calculated value. Optional Second Stage The second stage can be made of a single leaf, provided the stress levels are acceptable and equipment is available for making the required taper contour of the leaf ends. 7. Strength of Spring Eyes The bending stress in the main leaf adjacent to the spring eye or in the eye itself, due to longitudinal forces acting on the spring, can be calculated by the following formula: 3FD +9 ew The stress calculated by the above formula applies to upturned, downturned, and Berlin eyes, except that in the case of the Berlin eye the stress is zero under a compres- sion type longitudinal force. Thus, the Berlin eye has a strength advantage over the other two types. Tests have shown that when the stress, a calculated by this formula, reaches the yield point of the material, the eve will begin to open. If leaf springs are used in Hotchkiss drive suspensions where they carry the longitudinal force due to braking and driving, a large factor of safety should be allowed. The longitudinal forces in such applications are often much larger than those calculated from static forces, because shock loads may be applied either by the driver, by hop- ping the wheels on the road, or by longitudinal shake of the axles. The maximum calculated stress should there- fore not be allowed to exceed 350 MPa. Where vehicles such as buses are subject to frequent start and stop opera- stions, the required eye strength should be determined by life tests which reflect these operating conditions, bending stress resulting from the press fit of metal ‘buttrigs in spring eyes may be calculated by the following formula: 4 AEt 7 D+P ‘The maximum axial force and torque obtained without slippage between the bushing and spring eye will depend upon the finish of the engaging surfaces, their hardness, and the degree of lubrication present when the press fit is, made. 159 [Note that the stresses produced by bushing press fit and. by horizontal forces on the main leaf eye may be algebrai- cally additive. Nomenclature A = Difference between OD of bushing and ID of eye Stress ID of spring eye Thickness of leaf at eye Width of leaf at eye Modulus of elasticity = 200 000 MPa Longitudinal force s D t E FChapter 6 Installation Effects 1. Characteristics of Shackles ‘The rate of deflection of a spring is defined as dP/4f, the slope of a tangent to the load-deflection curve. A leaf spring tested on the rollers of a universal load scale has a rate which varies only slightly as the chord between the ends of the spring changes in length. However, due to the way the spring is installed, the rate on the vehicle may be different from that obtained on the load scale. Knowledge of installation effects will reduce the amount of experimental testing required to obtain the desired installed rates, and it can also be useful in obtain- ing variable rates where desirable, through the use of ordi nary springs with particular arrangement of shackles ‘The installation may involve contact pads, bushed eyes with one spring end fixed and the other end shackled, oF bushed eyes with both spring ends shackled. If contact pads are used, the active length of the spring may decrease as it deflects under load, and the rate may accordingly be increased (see Fig. 3.17A). ‘When one eye is fixed and the other eye is shackled, 1wo effects will result. As the spring deftects, the length of the chord changes, and the shackle will swing and change its angle. In swinging, the shackle may lift or lower the eye of the spring and with it the point of load application. This is the fist shackle effect. When the shackle is not perpen: dicular to the datum line of the spring, the shackle load will have a longitudinal component either compressing oF stretching the spring between the eyes. Compressing will decrease the rate of the spring, while stretching will in- ‘erease the rate. This is the second shackle effect. In the first shackle effect, the raising or lowering of the shackle eye changes direction when the shackle passes through the perpendicular position and also when the spring passes through the position where the distance be- tween the eyes is a maximum. This is represented by “Rat linkage,” where the three-link equivalent (see Chapter 4, Sections 4 and 5), is stretched out in one straight line, ‘except that in springs with eyes of unequal size or unequal offset, the “flat linkage” is only approximately fat. The terms “compression shackle” and “tension shackle” refer to the fact that load on the spring induces a compressive force in a compression shackle and a tensile force in a tension shackle. In the second shackle effect, the compressing or stretch- ing of the spring changes when the shackle passes through the perpendicular position. The amount of shackle effect ‘depends on the load which the spring carries rather than (on the rate of the spring. The rate of an installed spring with shackle may easily be 50% higher or lower than the 161 nominal rate. To simplify calculations, the charts in Figs. 6.2-6,5 may be used for springs with one shackled eye and fone fixed eye ‘These charts give the installed rate in percent of the nominal rate. The nominal rate is the rate calculated as shown in Chapter 5,Section 1, or the rate obtained from tests with spring eyes on rollers and with the spring in the fully extended position represented by “flat linkage.” For calculating these charts, the bending stiffness of the spring. was replaced by torsional stiffness in the hinges of the equivalent linkage. Experimental checks confirmed the calculated results within the limits of accuracy of the measurements. ‘The rate of spring with shackle depends on the nomi- nal rate of the spring, the position of the shackle, the length of the shackle, the “camber” of the spring, and the load on the spring. ‘i For the purpose of these charts, it is convenient to ‘express the “camber” of the spring by a term which is characteristic of the linkage directly and includes the effect of eye offset and shackle. The geometric deflection is used for this purpose. It is measured along the line of load application. It isthe distance by which the load appli- cation point is displaced from the position at which the ‘equivalent linkage is flat (maximem distance between eyes). (See Fig. 6.1.) The geometric deflection is zero when the linkage is lat or stretched, positive for increased loads (with increasingly negative spring camber), and negative for decreased loads (with increasingly positive spring cam- ber). It includes the effect of stiackle displacement. In Figs. 6.2-6.5, the geometric deflection is given as a per~ ‘centage of the (stretched) spring length (L), so that the ‘curves can be used for all sizes of springs. ‘The “stretched” length is different from cither the straight or the active length, but itis sufficiently accurate to use the straight length in this instance. To express the effect of load on the rate, it is most cofivenient to define the load by the geometric free camber of the spring, When the linkage is flat, the geometric de- flection is zero, and in that position the load does not depend on the shackle arrangement. This load is called P,, The geometric free camber is defined as the quotient of the load P, divided by the nominal rate of the spring (P,/k). In Figs. 6.2-6.5, the geometric free camber is ex= pressed as a fraction of the spring length, so that the ‘curves can be used for all sizes of springs. Curves have been drawn for various values of geometric free camber, ranging from 0.05L-0.20L. ‘The curves for symmetrical springs are shown in Fig. 6.2 (with compression shackle) and in Fig. 6.3 (with ten-‘A— SPRING WITH UPTURNED EYES AND COMPRESSION SHACKLE GEOMETRIC DEFLECTION (NEGATIVE) OPENING (Positive) ‘SHACKLE POSITION WHEN LINKAGE IS FLAT — SHACKLE wie ah DROP OF SHACKLE END 2 Pa GEOMETRIC DEFLECTION = — (OPENING + te +104 2 1 sHackLe DROP) 22 GEOMETRIC DEFLECTION = — (OPENING + phe + ge [SHACKLE DROP NEGLECTED] 8 — SPRING WITH DOWNTURNED EYES AND TENSION SHACKLE GEOMETRIC DEFLECTION OPENING; (NEGATIVE) (POSITIVE) i SHACKLE POSITION WHEN LINKAGE IS FLAT —— GEOMETRIC DEFLECTION = — (OPENING + i . fe + ; ‘SHACKLE DROP) GEOMETRIC DEFLECTION = — (OPENING - fro. + qo [SHACKLE DROP NEGLECTED] Fig. 6.1~Explanation of geometric detection 1.62sion shackle), Refer to Fig. 6.1 for sketches of springs with ‘compression shackle or with tension shackle. There are also curves for unsymmetrical springs in which the length ratio b/a or Y equals 2. These are shown in Fig. 6.4 (with compression shackle) and in Fig. 6.5 (with tension shackle). It is recommended that the curves in Figs. 6.2 and 6.3 be used for springs with length ratio up to 1.40, the curves in Figs. 6.4 and 6.5 for springs with length ratio exceeding 1.80. Where the length ratio is closer to 1,60, an interpolation between the curves may be used if they differ sufficiently to demand such a step. In the interest of readability the number of curves on each chart has been kept to a minimum, Near zero geo- metric detection, the curves for a given geometric free ‘camber follow similar patterns regardiess of shackle length. Curves for different geometric free camber values ‘can be interpolated or extrapolated as follows: In a given cchart (given shackle angle), with a given shackle tength st a given abscissa (geometric deflection), equal differences in ‘geometric free camber call for equal differences in ordi- nates (shackled rates). Example: In the symmetrical spring with tension shackle of 0.100L length and with 110 deg shackle angle, the shackled rate at —0.087L geometric deflection is 0.80k for P,/kL = 0.100 and 0.90k for P,/kL = 0.150; there- fore, it will be 0.95k for Py/kL = 0.175, Another example: In the unsymmetrical spring with compression shackle of 0.075L and with 90 deg shackle angle, the shackled rate at 0.0791 geometric deflection is 0.70k for P,/kL. = 0.200 and 0.90k for P,/KL = 0.150; therefore, it will be 1.00k for P,/kL = 0.125. The charts show that with compression shackles the load rate drops off to zero at high geometric deflections (or loads). To obtain the increase of rate with load which is generally desirable, springs with compression shackles should be used with negative geometric deflection (posi- tive camber or upturned eyes) at light static load. In gen- eral, this type requires the use of stronger bumpers or helper springs. With tension shackles the rate curves rise sharply with higher geometric deflections (or loads), which is desirable. ‘The curves indicate that at low shackle angles the springs ly high static deflection should stand near zer0 geometric deflection under light static loads, while springs with relatively low static deflection should stand with positive geometric deflection (negative camber or downturned eyes) under light static loads. The shackle effects are more pronounced with shorter shackles. Curves are given for tree lengths of shackles. The relation of shackle length to spring length is what is important, and the shackle length is therefore expressed in Percent of the spring length. Solid line curves show what happens ifthe shackle length equals 10% of spring length Dot-and-dash line curves apply when the shackle length equals 7.5% of spring length, and broken line curves when it equals 59% of spring length. ‘The shackle setting is the remaining factor which must bbe considered. The different shackle settings are identified by the minimum shackle angle, that is, the shackle angle when the distance between spring eyes is longest (or the spring has zero geometric deflection, or the linkage is fat) It is measured as shown in Fig. 6.1. Curves are given on ‘each of the charts for six different minimum shackle an- gles ranging from 60-110 deg. Shackle angles increase when the spring is deflected from the fat position; therefore, the shackle may fall in line with the adjacent equivalent link. If this happens in rebound, the shackle may “toggle” or swing over from a compression position into a tension position. This danger exists often on compression shackles with large minimum shackle angles, and it is then necessary to provide either 4 rebound stop for the spring iself or for the shackle. ‘As an example of the application of these charts, the installed rate is determined for a spring withthe following specifications: 1270 mm 560 mm 710 mm Length center to center of eyes L Length fixed (front) cantilever a Length shackled (rear) cantilever b Nominal rate; clamped between rubber pads : k= 21 N/mm Design load Po = 3340N Opening (positive) at design load 25 mm Upturned eyes. inside diameter ID. = 38 mm Thickness of main leaf t= 670mm Clearance (metalto-metal) © x, = 115 mm Rebound x = 135mm 100 mm 75 deg. Compression shackle, length Minimum shackle angle ‘Since the cantilever length ratio Y = b/a = 710/360 = 1.27 is appreciably less than 1.40, the rate curves for symmetrical springs will give adequately accurate infor- mation. (See earlier discussion in this Section.) ‘The opening at zero geometric deflection is (see Figs. 4.1 and 4.2, and Figs. 6.1 and 6.6) ett _ osID+H+t + 2 2 0.5 (38 + 6.70) ~ 6.70 2 = 14.53 mm The geometric deflection at design load is, therefore: = 14.53 = ~ 39,53 mm This equals —3.1% of the spring length. Since it was obtained by neglecting shackle effects, a correction may have to be introduced later. The load at zero geometric deflection (P.,) will be the same as the load on rollers at 39.53 mm beyond static deflection, 1.63a é é Zo00 SUI SAACREE ANGEE) . SHACKLE ANGLE AT eet | = 1 ZERO GEOMETRIC 8 L 21754 t 5 H CAMBER INDICATED ht 5 isolr tt a 1 2 1 & 4 2 GEOMETRIC FREE Ee 100 (CAMBER AS FRACTION 5 OF SPRING LENGTH a 76 g INDICATED z 50 : & -15 -1 -6 0 +5 +10 +15 GEOMETRIC DEFLECTION (IN PERCENT OF SPRING LENGTH) 80° SHACKLE ANGLE 90° SHACKLE ANGLE 8 a. i. Te 15D: |.050 i -15 -10 -§ 9 + +10 +5 -15 -10 -5 0 +5 +10 GEOMETRIC DEFLECTION (IN PERCENT OF SPRING LENGTH) SHACKLED RATE (IN PERCENT OF NOMINAL RATE) Fig. 6.2-Rale variations of springs with one compression shackle and length ratio 1.00-1.40 1.64z & @ 2 70° SHACKLE ANGLE — $ 200 1 SHACKLE LENGTH = | 3 ; | IN PERCENT OF 5 754 ela $ | e SPRING LENGTH: Evo —_|—] : | 2 A | =10 Percent # Pe 2% zie 1 —-— = 75 percent = He 8 = | —— ss percent gz WIS TL | si | 200-7 | 2 os Se 0 -| gilt 100 2 i oso} AN oe ers ee een onetime 15 GEOMETRIC DEFLECTION (IN PERCENT OF SPRING LENGTH) 100° SHACKLE ANGLE 110" SHACKLE ANGLE & 200- wy 7 5 FF 1 ~ § tee | ' _ 5175 tt — t ; 5 M ie 2 t s L % 160 1 ' 5 beau eae § 12s: P10 | a res 123] g { 2 = 10 a PoE # Sn é 15-200 = 7s s al Saal nee so! 8 | A173 | 00 Hate 139 = AP P80 |os0 -h to], 030 g “0-5 0 +5 0-15-10 SOO 5 GEOMETRIC DEFLECTION (IN PERCENT OF SPRING LENGTH) Fig. 62 continued 165SHACKLED RATE (IN PERCENT OF NOMINAL RATE) SHACKLED RATE (IN PERCENT OF NOMINAL RATE) ” SHACKLE ANGLE on T Tae SHACKLE ANGLE AT i HERES LEE Ea ce ZERO GEOMETRIC CAMBER INDICATED 150 125 GEOMETRIC FREE CAMBER AS FRACTION 100 OF SPRING LENGTH 2 1200] | INDICATED 507-15 10-5 05 HO HS GEOMETRIC DEFLECTION (IN PERCENT OF SPRING LENGTH) 80° SHACKLE ANGLE 90° SHACKLE ANGLE Ft ie it q sons 10-5 =O 5 1B 10-5 0 HHO GEOMETRIC DEFLECTION (IN PERCENT OF SPRING LENGTH) Fig. 63Rate variations of springs with one tension shackle land fength ratio 1.00-1.40 1.66SHACKLED RATE (IN PERCENT OF NOMINAL RATE) 70" SHACKLE ANGLE ‘a BBlEta IAL RATE) SHACKLE LENGTH 3 IN PERCENT OF g SPRING LENGTH: s 5 = 10 percent 5 +75 veneer E as PERCENT & é a 5 . ar a @ GEOMETRIC DEFLECTION (IN PERCENT OF SPRING LENGTH) 100° SHACKLE ANGLE 110° SHACKLE ANGLE é : : Va + |t60 so 0-8 O18 105 GEOMETRIC DEFLECTION (IN PERCENT OF SPRING LENGTH) Fig. 6.3 continued 1.67SHACKLED RATE (IN PERCENT OF NOMINAL RATE) SHACKLED RATE (IN PERCENT OF NOMINAL RATE) 60° SHACKLE ANGLE SHACKLE ANGLE AT ZERO GEOMETRIC CAMBER INDICATED GEOMETRIC FREE CAMBER AS FRACTION \ Cy OF SPRING LENGTH \ ut a 1 F INDICATED At Awl 0-5 0 45) HO +15 GEOMETRIC DEFLECTION (IN PERCENT OF SPRING LENGTH) 90° SHACKLE ANGLE 80° SHACKLE ANGLE \75 175° [100 150 150 |o50 =15 -10-§ 0 +8 +10 +16 -18 -10 -§ 0 +5 +10 GEOMETRIC DEFLECTION (IN PERCENT OF SPRING LENGTH) Fig. 64—Rate variations of unsymmetrical springs with one 1.68‘SHACKLED RATE (IN PERCENT OF NOMINAL RATE) 70° SHACKLE ANGLE 200 pq ‘SHACKLE LENGTH IN PERCENT OF ‘SPRING LENGTH: = 10 PERCENT SHACKLED RATE {IN PERCENT OF NOMINAL RATE) —-—. = 7.5 PERCENT a = 5 PERCENT ‘a 10 #15 GEOMETRIC DEFLECTION (IN PERCENT OF SPRING LENGTH) 100° SHACKLE ANGLE 110° SHACKLE ANGLE te p=: T i E i cea las 100 “AA ae chet” > <4 = | 00 pera 1S ane iC WN 1 “16-10-50 45 «#10 -15 -10 -5 0 +5 GEOMETRIC DEFLECTION (IN PERCENT OF SPRING LENGTH) Fig. 64 continued 1.69SHACKLED RATE (IN PERCENT OF NOMINAL RATE) ‘SHACKLED RATE (IN PERCENT OF NOMINAL RATE SHACKLE ANGLE og HE to |} “15-10-50 Oa GEOMETRIC DEFLECTION (IN PERCENT OF SPRING LENGTH) 80° SHACKLE ANGLE SHACKLE ANGLE AT ZERO GEOMETRIC CAMBER INDICATED GEOMETRIC FREE CAMBER AS FRACTION OF SPRING LENGTH INDICATED. 90° SHACKLE ANGLE —15 10-50 45H HS =10 5045 +10 GEOMETRIC DEFLECTION (IN PERCENT OF SPRING LENGTH) Fig. 6.5—Rate variations of unsymmetrical springs with one ten- sion shackle and length ratio exceeding 1.80 1.70Cy SHACKLED RATE (IN PERCENT OF NOMINAL RATE) ) 70" SHACKLE ANGLE 200 a1 ‘SHACKLE LENGTH IN PERCENT OF ‘SPRING LENGTH: = 10 PERCENT = 7.5 PERCENT ————— + 5° PERCENT 50° = 10800 GEOMETRIC DEFLECTION (IN PERCENT OF SPRING LENGTH) SHACKLED RATE (IN PERCENT OF NOMINAL RATE] 100° SHACKLE ANGLE 175 160 rst 100-4 50 7 4 ~15 -10 6 ~O OS 15-10 GEOMETRIC DEFLECTION (IN PERCENT OF SPRING LENGTH) Fig. 65 continued 1O 1270 — —————___- OPENING AT ZERO GEOMETRIC DEFLECTION = 14.53) GEOMETRIC DEFLECTION! AT DESIGN LOAD = —39.53 SOLID LINES: OPENING AS SPECIFIED FOR DESIGN LOAD (ON ROLLERS) BROKEN LINES: AT ZERO GEOMETRIC DEFLECTION 3: 25 OPENING AT DESIGN LOAD NEGLECTED Fig. 6 6—Spring for sample use of charts P, = 3340 + 39.53(21) = 3340 + 830 = 4170 N The geometric free camber is: Po/k = 4170/21 = 198,58 mm This equals 15.6% of spring length, so that an interpola- tion about one-quarter of the way from curves for Pe/kL 0.150 to curves for P,/kL. = 0.175 will be in order. ‘The shackle length is 100 mm which equals 7.9% of spring length. This makes the “dot-and-dash” curves nearly applicable, or an interpolation nearly halfway from broken line curves to solid line curves. The geometric deflection at the metal-to-metal clear- ance position will be —39.53 + 115 = 75.47 mm, This equals 5.9% of spring length. ‘The geometric deflection atthe rebound position willbe = 39.53 ~ 135 = — 174.53 mm. This equals ~13.7% of spring length. ‘The specified minimum shackle angle is 75 deg, midway between the curves for 70 and 80 deg. ‘The following values for shackled rate (in percent of nominal rate) may now be established: 1. At design load position (where geometric deflec- tion equals —3.1% of spring length): Shackle Ange: 70409 seg TS eg" ‘Sackl Lang: 086 GTGBL COBOL” CoB GION OO COBH PKL = 0.180 88100 8 ~ 0200, 96 103 % Ofse 8 100 SG 172 2. At metal-to-metal clearance position (Where geo- equals + 5.9% of spring length): rde0 W009 Theat sac eng DSL BT Bill” GowH SL BR” SEE Pu = 0150 Bee ies Sm oe tee & te Sige fe oe 8s 3. At rebound position (where geometric deflection equals — 13.7% of spring length): sacle gle roses ode 18s" snectie Lng, 2075 OTOL BOT GOTEL C100. GowA- Cosme Pum = 0150 117108 16128 ous, es: ‘er ee Cr a a CC) tnerplated vais. To summarize: wot nominal rate New At design toa 8 206 ‘At metalo-metal clearance 105 220 ‘vecage in jounce range 102 ata ‘At rebound 120 252 ‘Average in rebound range 109 229 ‘The data (and the curves) show that @ reduction in shackle angle would produce a higher rate in the jounce» range and simultaneously a lower rate in the rebound range. To obtain values for load at different deflections, the geometric deflection must be multiplied by the average rate between zero and that deflection, The resulting prod- uct is the load difference between that deftection and zero geometric deflection. The load at zero geometric deflec- tion is not affected by the shackle and can therefore be caleulated as shown above (4170 N). For this example, the load'at —39.53 mm geometric deftection would be: 4170 — 39.53(21.4) = 4170 — 846 = 3324, and the geometric deflection under 3340 N design load would be: —39.53 + 16/21 = —39.53 + 0.76 = 38.77 mm, This is the correction mentioned above. It is small in this case, but may be large in others. Similar effects on rate may be produced by lengthwise tension or compression of the spring and can be causeid by driving thrust, braking, or nonhorizontal position of the springs. 2, Windup of Springs {In many applications, leaf springs are loaded not only by vertical forces but also by horizontal forces and torques in the longitudinal vertical and transverse vertical planes. Torque in the longitudinal vertical plane (windup) is usually produced by a longitudinal force applied above or ‘below the spring seat. Where the spring does not carry the longitudinal force but only the vertical force and the windup torque, stresses and deflections can be calculated fairly simply with reasonable accuracy as shown in the section, “Springs Carrying Vertical Load and Windup Torque.” However, when the longitudinal force is also carried by the spring, asin the Hotchkiss drive suspension, generally valid formulae become very complicated, and the problem must be treated as described in the section “Springs Carrying Vertical Load, Windup Torque, and Longitudinal Load.” In all formulae given in this section, the total length should be used as the flat length L. This will give correct values for deflections and sufficiently accurate ones for stresses. vs TaNv=L= T BL .25kL? kL [ WINDUP STIFFNESS w = 2 = KL? v4 Fig. 6.7—Windup of a symmetrical spring under brake application 1.73‘The stress formulae apply only to springs built as uni- form strength beams. The deflection formulae, however, are not so restricted; they apply to all springs having certain rate and length ratios, but they are true only for small displacements because changes of ratio and length are neglected. Springs Carrying Vertical Load and Windup Torque ‘The center or seat of a leaf spring is elastically re- strained against deflection produced by a vertical force due to its rate k(N/mm); it is restrained against windup oF rotation produced by a torque in the longitudinal ver- tical plane due to its windup stiffness or windup rate ‘w(N : mm/radian). This windup stiffness is often used to resist the driving and braking torque, as in the Hotchkiss drive suspension. In “symmetrical” semi-elliptic springs, a vertical force produces'no rotation of the seat, and's torque produces no vertical deflection of the seat. Such a spring with a torque applied to it is shown in Fig. 6.7. ‘The windup stiffness is «» = kL*/4 (N-mm/radian). The stress is, expressed by the windup angle V s. = 2Et .v.sF or expressed by the torque T SE pgp 27S ww ot where SF is the stifening factor (See Chapter 5, Section 2), S is the stress.caused by a deflection f, and V is the wvadup angle in radians. The formulae show the great importance of spring length for resistance to windup. Windup stfu incteaues s the square of the spring tength or springs of equal load rate. Thestress produced bya given Corgue decreases inversely as the length for springs of equal loed rate and equal stress rate S/t Semi-lipie springs ae fequently made unsymmetri cal to obtain desirable geometry. This can be done by neque! division of length, or by adding leaves (0 one of the arms, or by eombining both methods. Fig. 68 shows a spring with arms of unequal length under vertical and tongue loading ating together. These loadings are pro- died under forward brake application in a Hotchkiss drive suspension. | duce beside vertical deflection, a tilting ofthe seat and a tora wil produce, besides tilting, a vertical defection e s Fig. 6.8-Windup of an unsymmetrical spring under brake application 1.74‘When the torque is applied with the spring seat prevented from deflecting vertically, the windup stiffness and the windup stress as expressed by the windup angle are in- creased in the ratio ‘The relationship between the vertical deflection f (mm) and the tilt angle V (radian), due to a vertical load, may be expressed as the equivalent torque arm Q (mm) which is shown in Fig. 4.2. There the deflection is called “x” and the tilt angle “8.” Q="v ‘The length of the equivalent torque arm Q can be cal- culated ifthe lengths of the spring cantilevers and the ratio of their stiffnesses are known: With length of fixed cantilever length of shackled cantilever rate of fixed cantilever, rate of shackled cantilever and with Consideration of cantilever deletions will show that wk + by Le +9 ak = by ZW + D For springs designed to deflect in circular ares (without reinforcing leaves), Z = Y? and the formula becomes Q This has been shown in Table 4.1, and the relationship between L, Q, Y, and Z is graphically demonstrated in Fig. 4.5. The zero ordinate for L/Q (where Q becomes infinitely large and the seat angle change or tilt angle becomes infinitely small) is crossed by each Z line at the point of Y = Z. Thus, whenever Z equals Y, a vertical force will produce no tilting of the spring seat. When Z becomes smaller than Y, the equivalent torque arm Q will {20 in the direction of the longer spring arm, as indicated by the negative ordinate values. The relationship between the vertical deflection f (mm) and the tilt angle V (radian), due to a torque while the spring is under a constant vertical load P, may be ex: pressed as the effective swing radius B (mm), which is, shown in Fig. 6.8. Om cD ey The virtual center of rotation for this torque loading is a point at a distance B from the spring seat. B equals zero. if Y equals Z, and the center of rotation is then at the spring seat. For larger or smaller values of Z, the center moves closer to the end of the spring arm which is stiffer than the value corresponding to Z = Y. The windup stiffness of unsymmetrical springs can be calculated from the cantilever rates and lengths Bat k © = (ak, + bby) , Eth koko? (+ DRL? tke ZED For convenience, itis expressed as a fraction of the stift ness of a symmetrical spring of the same length L and of the same rate k. The stiffness ratio is oo Marky + biky) KL (KF RL kk + ky) 40" +2) 7 G+ + ‘Thisis plovted on Fig. 6.9 over a“Y" abscissa, It shows, that the value «»/(kL‘/4) can exceed the value | only in 1 few cases which have little practical application: 1. When Z equals 1 while Y is either substantially ‘more or substantially less than 1 2. When Z exceeds 1 while Y is less than 1 3. When Z is less than 1 while Y exceeds 1 Lines have been drawn both in Figs. 4.5 and 6.9 for W = Z/Y* to indicate the extent of the distortion (com- pared with a conventional spring layout) which is required for a given relationship between Y and Z. (see Chapter 4.) ‘The stress due to windup in unsymmetrical springs is normally higher in the longer (usually the shackled) can- tilever b; the windup stresses in the two cantilevers have the relationship Z Sy ¥? SF, ‘The stress in the longer cantilever expressed by the windup angle V s, 2Et 2 + Sy = 2B. 5 F, ee 2Ve(Z Ee) _ by the torque T SE 204 Dg, aD ayaa TSF 175where SE, and SF, are stiffness factors that may be deter- ‘mined as stated in Chapter 5. When Q, k, and w are known, they can be used to express the relations between vertical load P, torque T, Springs Carrying Vertical Load, Windup Torque, and Longitudinal Load ‘The three-link equivalent shown in Figs. 4.2 and 4.3 baal vertical deflection f, and tilt angle V: very closely reproduces the action of the spring under vertical force and windup torques, ifthe bending stiffness reba t ee Of the leaves is replaced by torsional stiffness in the two ee x Te hinges ofthe thee inks It als gives a reasonably accu For vertical load only, the defiection f = ©, rate equivalent under longitudinal forces. It may therefore P 7 be used to calculate the effect of these forces. the tilt angle V = 2, and Under longitudinal forces such a linkage evidently has Qk a marked toggle effect which increases if the spring is . © = Qas previously defined. made shorter. Under braking and accelerating forces, the results depend upon whether the spring cantilever which For torque only, the defection ¢ = 1, transmits these forces from spring seat to fixed eye is in tension or compression. Ifthe fxed cantilever isin tension. ae eee there isa stifening effect, and both windup and vertical @ deflection are decreased. Ifthe fixed cantilever isin com: t ag previ pression, the toggle action produces'a deerease of rate, and eee cee eee ees both windup and vertial defection are increased. Incom- pression, this effect can be strong enough to make the yee eee | ate, WS [| %s%o [% ‘THE ORDINATE FURNISHES THE WINDUP * Fig ST est | STIFFNESS OF THE UNSYHMETRICAL 10) 1 = | zor SPRING (01 AS A FRACTION 0 [4 pe OF THE WINUP sriFFNESS me eee | os eves OF ASYMMETRICAL SPRING OF EQUAL LENGTH (L) AND OF EQUAL RATE (kt | | os | “svmmernicat * KL © SUNSYMMETRICAL _4(2+2) KL ewe woz? FF Yb Wenate nario} 2 = kay (RATE RATIO) LT | 8 10 12) 16 18 20° 22 y LY Fig. 6 9—Windup stittne 188 of unsymmetrical springs 176mechanism completely unstable and buckle the spring if it is relatively short or if the main leaf is too thin. The buckling tendency also increases in proportion to the static deflection ‘An unequally divided (unsymmetrical) spring offers _ereater resistance to buckling when the shorter cantilever is between the seat and the fixed eye, as in the usual Hotchkiss drive suspension, due to its greater bending resistance and the additional vertical load produced by a horizontal accelerating force. In this case, the horizontal force not only produces the windup torque T, but also the additional vertical load P that opposes the windup lifting force T/Q. The vertical load is obtained from the follow- ing formula FHL Tire friction force Vehicle center of gravity height above ground Vehicle wheelbase The net spring defection from the static postion may be obtained by substituting the value for vertical load P in the defection formula, which shows that vertical load and torque are related in the following manner to produce vertical deflection Ree ke In acceleration, deflection (P/k) is always positive (pring in compression) and deftection (T/Qk) is always negative (spring in rebound). In braking, the signs ofthese deflections are always opposite because the horizontal (ire friction) force, vertical load, and windup torque are reversed. When the deflection f is positive, the vehicle drops, and when fis negative, the vehicle rises Rates of Unsymmetrical S} Roll ings Under Bounce and It should be noted that the vertical rate k equals the sum of the cantilever rates (k, + ky) only if Z = Y, which happens in symmetrical springs where Z = Y = 1. In un- symmetrical springs, k is usually smaller than (k, + kj): However, k may become equal to {k, + ky) when the longer cantilever is made stiffer by additional leaves or longer leaves to balance the deflections. Fig. 6.10 shows the relationship of k,, ky, and k. Fig. 6.11 shows the ratio (ky + k,)/k for various values of Y and Z. The fact that in unsymmetrical springs (k, + ks) is larger than k can be used to provide additional roll stiff ness without using a roll stabilizer. The lower diagram of Fig. 6.10 shows the normal position and the deflected Positions of two unsymmetrical springs which are rigidly ee mounted on a sti axle, one end of which is moved up and the other end down. If both ends were moved down together, the spring seats ‘would tilt one way, as shown in the upper diagram. If they were moved up together, the seats would tilt the other way. If one end is moved up and the other end down, tilting of the spring seats is prevented because they are forced by the axle to remain almost parallel. The axle imposes a windup torque on the springs and is itsel stressed by a twisting torque. The rol stifness of a pair of springs rigidly mounted on an axle is rot = (ky + ky) C'/2 (N+ mm / radian) where C is the distance between spring seat centers. Any torsional deflection of the axle will reduce this roll stiff: ness. For symmetrical springs or for springs whose seat is not forced to tilt with the axle, the roll stiffness is kay = KO? Additional roll stiffness is produced by twisting of the spring (sre Section 3 in this chapter). This raises the total roll stiffness by 20-40% above the values given in the last two equations. Summary Springs are made unsymmetrical either for reasons of geometry, 1 obtain a desirable equivalent torque arm or center of rotation, orto secure increased rol sifiness with- ‘out increasing the ride rate by the ratio (ky + y)/k. ‘Under torque loading only, windup stifness is reduced if springs are unequally divided as shown in Fig. 6.8, However, in a Hotchkiss drive suspension, an unsymmet- rical spring has less buckling tendency because of the stabilizing effect of additional (dynamic) vertical load and the increased bending resistance of the fixed cantilever. The most economical springs are designed to bend in a (nearly) circular arc when loaded at the seat. In such springs, the rate ratio ky/ky equals the cube of the length ratio b/a For these useful springs, most of the formulae can be simplified by substituting Y? for Z. In Figs. 6.9 and 6.11, these springs are indicated by the heavy line for waz 3, Twist of Springs In the usual suspension applications, leaf springs may be subjected to twisting, for example, by an obstacle under ‘one wheel of an axle, 17sear anave | pe IN PITCH OR BOUNCE k 1S LESS THAN (ka + kb) | | IW ROLL k IS EQUAL TO (ka + kp) kaP/E ka=Palfa kb = Pb / fb IN ROLL THE AXLE PREVENTS THE CHANGE OF SEAT ANGLE. THIS ADDITIONAL RESTRAINT PRODUCES INCREASE OF RATE. ‘THE TWISTING OF THE SPRINGS RESULTS IN A FURTHER RATE INCREASE. Fig. 6.10—Unsymmetrical leat springs with rigidly connected axle 178kathe _ (Z+1)(Z+Y?) kath ee Sa . Kat kb 1 k Zv+ 12 [ FOR CONVENTIONAL SPRING: “2™“t = y—1+ 7 Y = b/a (LENGTH RATIO) 10 Le fee ' L Z* ka/kp (RATE RATIO) a, SYP git vt Z fy 2 w=z/v3 . 8 4 we | T 7 I 6 4 3 x atl oe + + ib aol eral tees ae 0.8 1.0 1.2 14 16 Solid Lines — Conventional Springs Broken Lines — Ur Fig. 6.11—Relationship between spring rat ‘Twisting a spring leaf (having the substantially rectangu: lar cross section of the SAE leaf spring steel with width w and thickness t) through a degrees in length & (for example, between eye and seat) will produce a shear stress $ = 1400 t a/2 MPa (Approx) and a torque Inconventional Springs te (K) and sum of cantilever rat (hy + Ke) When military wrappers or reinforced eyes are used, both first and second leaf willbe twisted and their torques will add, The flexibility of brackets and shackles will reduce twisting of the spring. ‘The resistance of the spring against twisting increases the roll stiffness in the usual application Sometimes an axle is supported by two springs which are not parallel. Convenience of attachment is the usual reason for this practice. Ifthe angle included between the springs is kept small (up to 10 deg), the action ofthe spring is changed only a little. Parallel motion of the axle (ride) vill produce a small twist ofthe spring, and in unsymmet- rical springs, a small windup torque, and the spring rate will be increased in this way. = 420 wt a/2.N - mm (Approx.) To keep the twisting stress low, it is necessary to distrib- tute the total twist angle over as long a length of the spring. 8s possible, which means that the clips should not restrain the main leaf from twisting. 179Chapter 7 Interleaf Friction 1. Characteristics Interleaf friction can be defined as the force which op poses the relative motion of adjacent leaves. The friction force provides damping in the suspension system. It also resists the initial deffection of the spring, making the sus- pension system less responsive to dynamic forces. Many heavy trucks using multiple leaf springs rely solely on interleaf friction to provide damping and, there- fore, shock absorbers are not required. In suspensions that use single leaf springs or springs where tip inserts have been installed, shock absorbers must be used to damp out the vibrations. ‘The magnitude of the friction force depends on the condition of the leaf surfaces (coefficient of friction), on the load carried by the leaf surfaces, and on the speed of sliding between leaves. Fig. 7.1 shows a load-deftection diagram as the result of| atest conducted at a slow rate of deflection without rap- ping the spring, It shows that at any given deftection, the Toad may be any value between an upper and lower limit, depending on the direction of motion and on the distance \ rom the last reversal of motion. Tests conducted at faster rates of loading and unloading will show the same basic characteristics, but the width and the shape ofthe diagram. will be different. ‘The procedure for testing springs to obtain the friction loop is given in Chapter 2, Section 1 Springs with high frictional forces will have a higher effective (dynamic) rate when operated through small am- plitudes. At larger amplitudes, the effective rate will ap- proach the rate determined by a slow test. 2, Measurement It has been found convenient to compare one spring with another by means of a friction factor: Friction Force Fri ‘Average Load jon Factor = Though not strictly correct, it is permissible to assume that the friction force equals one half the difference be- tween the compression load and the release load. The average of the compression load and the release load at a sven defiection is termed the average load. (See Chapter <4, Section 1). 1.81 3. Control ‘The friction factor, as measured by a slow test, may be as high as 0.10. To obtain lower friction factors, the con- tacting surfaces should be smooth, leaf ends should be flexible (tapered) and lubricated. Under these conditions, the friction factors may be between 0.02-0.05 on a new spring. However, when the spring becomes dry, dirty, and the contact surfaces are scored, the friction factor will In some applications, various types of interleaf liners or tip inserts are used to eliminate squeaks and obtain low friction factors. A typical tip insert installation is shown in Fig. 3.16E, Section A-A. LOAD-N DEFLECTION — mm Fig. 7.1—Typical diagram of a leat spring tested at low speed and without rapping to show Interteat frictionCHAPTER 8 Operating Stress and Fatigue Life 1, Operating Stress In order to hold the mass of a spring to a minimum, it is necessary to use the highest stress that wll give satisfac- tory operation. This stress is limited by three factors: set- tling under load, fatigue life, and quality of spring material and processing ‘The settling of a spring of given hardness is a function of the maximum stress. Any spring which is repeatedly loaded to a high stress will settle somewhat during the frst few load cycles. This settling can be reduced toa negligible amount by the proper presetting as outlined in Chapter 5, near the end of Section 6, and Chapter 9, Section 4, For automotive suspensions, the design load stress is usually in the range of 600-750 MPa for passenger cars ‘and 350-550 MPa for trucks. There is some possible over- lap of design load stress between passenger cars and trucks for special vehicles (‘ight trucks, station wagons, taxicabs, cetc.). The maximum stress should not exceed the mini- um yield stress of the spring material. For properly heat \efeated alioy steel, the minimum yield stress is generally accepted as 1200 MPa. ‘When a spring is subjected to windup under engine or brake torque or any other forced external means, stresses in addition to those due to vertical load may be present and should be considered in computing the stress. 2. Fatigue Life Fatigue life is expressed by the number of deflection cycles a spring will withstand without failure and can be estimated by the use of Fig. 8.1. This has been constructed With the help of data obtained from laboratory fatigue tests which were conducted on various spring designs be- tween 1950 and 1970. Improvements in the processing of the steel which may be expected in the futuce both at the steel mills and at the spring plants can be taken into account by shifting of the ‘cycles to failure” lines in Fig. 3.1. Such an adjustment would be based on the results ‘obtained from extensive controlled fatigue tests. In order to establish the fatigue life cycles which are acceptable in any spring design, it is desirable to have road durability tests run over a prescribed course so that fatigue life test data and actual road durability results may be correlated. 1 must be understood that the number of estimated life \ajcles is a statistical average and that fatigue test results will show scatter (or dispersion) even under closely con- trolled test conditions. The extent of the scatter will de- pend on the consistency of surface condition, fabrication, and the general quality of the springs which are tested The average life “cycles to failure” lines, as shown in Fig. 8.1, apply to present springs but do not include the effect of surface cold working as produced by shot peening (sce Chapter 9, Section 4). Tests have proved that the fatigue life of a spring can be greatly improved by shot peening the tension surface. The increase in fatigue life ‘will depend on a number of factors, such as peening inten- sity and coverage and the condition of the surface prior to shot peening. Fatigue testing is an accelerated method of examining. springs for design adequacy and for quality controt pur- poses Reexamination of the design will be in order if the fatigue tests result in failures which are confined to one section of the spring. It is recommended that the fatigue setup produce atleast an average of 50 000 and preferably ‘an average of 100 000 cycles. At higher stresses (shorter lives) the scatter is theoreti- cally reduced, so that fewer test samples will produce a sven degree of precision in the estimated life ofthe entire population. However, lower stresses (longer lives) produce more realistic results, since they duplicate more nearly the actual service conditions and the spring is less likely to settle during the test. Also, comparisons between different ‘groups of springs will be more distinct at lower stresses, since different S-N curves tend to diverge the more they approach the fatigue limit (or limiting value of the stress at which 50% of the population would survivea very large number of cycles, usually 10 000 000.) ‘leaf spring used in a suspension will undergo a large number of cycles of small amplitude near the design load position without failure. Under greater amplitudes the number of cycles without failure will be reduced, since the maximum stress as well as the stress range are increased, and both are determining factors in the fatigue life of a spring. See examples D versus C (both with 700 MPa stress at design load) in Fig, 8.1 The metal-to-metal position (vertical load limit) is fre- quently used as the maximum deflection position of the spring in a fatigue test; but in heavy truck springs this, deflection is often considered excessive for the test setup, as it is rarely reached in actual service. The length of the test stroke is selected from experi- ence. A frequently used method of establishing. the 1.830 400 100 1300) 1 1 p——_*# ; U i INITIAL STRESS—MP2 YU =r 7 i + 1.200) i | Ht 1 it a i £ i) Z [ wo HAG POINT OF I 5 wy INTERSECTION ie = | (HORIZONTAL LINE). 70 - Sy INDICATES LIFE EXPECTANCY : | “ ‘AS DETERMINED BY LOCATION OF POINT OF INTERSECTION —| WITH REGARD TO THE DIAGONAL CYCLES-TO-FAILURE LINES ; + 600) i | 7 1 INITIAL | STRESS | MAXIMUM | EXPECTED | Exampces | STRESS | RANGE | STRESS | LIFE CYCLES “0 | a ° 700 700 75.000 j 8 25 550 825 80.000 c 500 | 400 900, 120 000 ° 550 300 850 1.000 000+ Y Fig. €.1—Diagram for estimating fatigue Iie cycles of steel leat ‘springs (preset but not shot peened) 1.84length of test stroke is to add to the compression stroke (from design load to maximum test load) one-half of this length for the release stroke (from design load to initial test load). This practice may require modification in those cases where it would produce less than 50 000 or much more than 100 000 cycles, according to Fig. 8.1. Example: To estimate the expected fatigue life of the first example shown in Chapter 5, Section 3, compute the stresses at maximum and initial test loads as follows: Deflection to design load ='3500/17.5 = 200 mm Metal-to-metal clearance (compression stroke) 110 mm Total deflection to maximum load 310 mm Stress at metal-to-metal position (6.30 gage) 985 MPa Stress rate = 985/310 = 3.18 MPa/mm. Release stroke = 44 + 110 = 55 mm Fatigue test stroke = 110 + 55= * 165 mm Initial stress = 985 ~ (165-318) = 460 MPa From Fig. 8.1, the intersection of the horizontal tine representing 984 MPa and the vertical line representing 460 MPa gives the expected average fatigue life cycles as, 65 000. A fatigue life of 100 000 cycles under amplivudes repre- senting these stress conditions will generally assure a road life of more than 160 000 km of “on highway” operation. In this example, shot peening of the leaves of the spring is recommended if 160 000 km of highway operation is expected. Any “offthe-road” operation will reduce the service life of the spring. Springs for “off-the-road” duty should, therefore, have a lower operating stress than those designed for highway operation. ‘Another example: To estimate the expected fatigue life ‘of the third example shown in Chapter 5, Section 3, (clear- ance of 90 mm to one-half bumper compression specified Additionally), compute stresses at maximum and initial test loads as follows: Deflection to design load = 14 700/110 134 mm Clearance to one-half ‘bumper compression 30 mm Stress at design load 535 MPa Stress rate = 535/134 3.99 MPa Release stroke = % - 90 45 mm Fatigue test stroke = (90+45) 135 mm Initial stress = 535 — (45 - 3.99) 355 MPa Maximum stress = 535 + (90 - 3.99) 894 MPa From Fig. 8.1, the intersection of the horizontal tine Fepresenting 894 MPa and the vertical line representing 395 MPa ges the expected average fatigue lie cls 75 000. ™ : : * * 1.85 3. Evaluation of Fatigue Test Results It has long been recognized that considerable variation is present in the fatigue life of springs loaded with the same stress cycle, and that the average life ofthe tested springs isnot sufficient by itself to establish a judgment either on the design, or on the material, or on the production method which they represent. The relationship between the number of applied cycles and the percentage of springs which failed at these cycles can best be analyzed with the help of statistical techniques which will systematically de- scribe the “dispersion” or “spread” or “scatter” of the recorded test results Sampling One of the main purposes of statistical analysis is to draw inferences about the properties of a large group (the population”) from the results of tests on a small group (the sample”). If the entire population were tested, one would not have to infer anything; one would know how the population reacted to the test. This would be called 100% confidence. If 99% of the population were tested, ‘one would be 100% confident about that 99%. Also, ifthe sample of 99% were considered representative of the re- maining 1%, one would be close to 100% confident of predicting the results if that 19 were tested. If only 5% are tested, one would know about that 5%, but how much could one infer about the remaining 95%? Actually, if certain conditions of sampling are met, one can infer a ‘great deal about the entire population from tests on small samples. What is required is a good, honest sample. ‘The primary condition for a good sample is that it be taken at random under conditions which ensure that all springs of the population have an equal chance of being chosen. This is obviously impossible in the case where the sample consists of a few handmade springs of a design which has not yet gone into production. Only experience can tell the engineer whether the various properties ofthe sample which can affect the test result (in regard to mate- rial as well as to production methods and controls) will also be present in the production springs (the “popula- ion”). This determination is outside the realm of stat tics. However, statistical mathematics are based on the inherent assumption that the sample isa true representa- tive of the population. Distribution If the entire population were tested under identical test conditions, the results could be shown in graphical form by arranging them in ascending numerical order and plot- ting the cumulative fraction (or percent) of failures over an abscissa of “life cycles.” A sample selected at random. from this population can be expected to exhibit a similar distribution of fatigue life; the larger the number of springs in the sample, the closer will be the similarityIt is possible to calculate the likelihood of similarity for samples of any given size. Tables are available based on such likelihoods; Table 8.1 is an example. It presents the “median rank” of each test result for a sample size be- tween I and 30, A rank is assigned to each individual test result corresponding to that portion of the population which it is most likely to represent. The median rank is used as an estimate of the true rank because it is just as, likely to be high as low. Table 8.1 lists percent figures. ‘A good approximation formula for the median rank (which may be used for larger sample sizes than those in Table 8.1) is: 100 2= 93, N+ 04 where: J = Position (in ascending order) for each test result in, the sample N = Total quantity of springs in the sample ‘The median rank line constructed from such data pre- dicts that certain percentages of the population will sur- vive specific cycles-to-failure. But any such estimate may “TABLE 81 —MEDIAN RANKS (PERCENT) FOR SAMPLE SIZES 1 TO 20 err substantially on either the high or low side. The ques- tion is: How confident can the engineer be of such an estimate? When he has accumulated a great deal of experi- ence in comparing the results of small samples which represent springs of different materials or different designs oF different production methods, he may judge, after con- templating two such median rank lines, that he should sive preference to one set of springs over the other because of its apparent superiority in fatigue life. However, when {judgment based on experience is not considered adequate for a final decision, then it will be necessary to construct lines of higher confidence. In many cases a confidence level of 90 or 95 or even 99% will be required, so that there will remain only a 10 of 5 of 1% tisk of the estimate being either too high or too low. For a chosen confidence level, the life cycles of a given percentage of the population will be found within ‘a certain “tolerance interval”. On the median rank graph this may be represented by a “tolerance band’ to either. side of the median rank line, Wider bands indicate in- creased doubt about the line truly representing the popula- tion. With a given sample size, the bands will be wider for higher confidence levels. With a given confidence level, the bands will be wider for a smaller sample size. Sample Size ani Rank Order 1 2 a “ 5 ‘ 7 . ° 10 Order 1 ooo =e masa rz04ssomt = aaas goo rate? 1 2 yor) Sao RST. Shae) Baas zee 2011S tTaee 62282 4 yas — etazr Sama et Seal? Seana aoe, aa87 ‘ biov agin S7aso 50000 uts gages asta 5 tross —masss uses S804 50000 aste9 ‘ sooo Tras) gaan 6 > sos = Tomy 71376 seas ° sim = aos ate . esr are 0 mm 0 Sample sie Ranke ae Ont ot 2 2 “ 1s 6 w 1° Orr . sor sera ao ase a0 5888 ase . 2 wares tases 1257147020028 7 2 3 asm bee boos tga7 aso aes a 13827 3 <— Seam Bars 275m mse ante aeare 21.178 tase + 3 thts Gras Sst 275 ase ae ao aaNet 5 © $000 45051 azo asaya 705 e084 maz ‘ > Saat Same Som asts Asano 02st Seas 7 3 ereco cata? Srane Soass 5.000 aeraazte meso a oYeaat—(Yozee = casee | oust Sty S089 000 re) 5 10 esau raat Taare g7a2s, 630095077 58.708, prs) 0 11 gaan eave 7amss ass gous gs205 sat ss nase 2 bese? erat erase host That. oa gram 3 seaoe —sezee ace 77825 73.080 foo ame “ 95163 se0e0 —aa6as = 7mat joa | orms Mt 15 feuee 60790 ATO yee aie 8 5700 a2 so ro 8 pees fess 7 sae sess 18 wea oe 8 woe | 1.86If a 90% confidence level has been chosen, the lower and upper limits to the estimated fatigue life distribution of the population can be shown by constructing “5% rank” and "95% rank” lines (see “Theory and Technique of Variation Research” by Leonard G. Johnson. Elsevier Publishing Co., 1964). The numbers 5% and 95% repre- sent the chance of being either t00 high or too low in assigning the given ranks to the individual springs in the sample, and they are used to establish the limits within which the true population is expected to lie. For a concise summary of distribution mathematics see “Engineering Considerations of Stress, Strain, and Strength,” by Robert C, Juvinall, McGraw-Hill Book Co,, 1967. Weibull Plot Several systems of mathematically organizing the test result data have been established. In the past, the normal (or Gaussian) distribution has been most widely used. It is graphically represented by the familiar symmetrical bell-shaped distribution curve which is completely defined by two statistical parameters: 1, The mean life—In the test sample it is the sum of | all the recorded test result values, divided by the sample size, It then becomes an estimate of the population mean life. A population with normal distribution has the mean coinciding with the median (which is the middle result when all individual results are arranged in order of magni- tude), and also coinciding with the mode (which is the ee value at which the greatest number of failures oc- 2. The standard deviation—It describes the scatter on cither side of the mean. For the test sample it is math- ‘ematically defined as the square root of the sample vari- ance (which in turn isthe sum of the squares of the differ cence between each recorded test result and the mean, divided by the sample size minus 1). It then becomes an estimate of the population standard deviation, ‘When the test results are arranged in ascending numeri- cal order, and the cumulative percent of failures is plotted (using median ranks) over an abscissa of life cycles on normal probability graph paper, it will be found that a straight line can be fitted to the results as long as the distribution is normal. This becomes an estimate of the population distribution. ‘While the normal distribution has a number of attrac- tive attributes and has been the subject of many publica- tions, it must be recognized that in spring fatigue testing. the results are usually not normal in that they cannot be ‘Sample Size Ranke ack, Ont ot 2 a ™ * ™ ” ” ” 2 Oraee 1 aze7 310) a0 zee 2704 2s tues tee aa ‘ er i i rt éia Sacer ae 3 ast 1397 nyse soge7 0860 om aerate | 2 8 g 6 5 & x 8 z & 2 60 1. Connect total degrees of freedom with Weibull stope and locate intersection point on auxiliary line. 2. Connect life ratio with intersection point and continue to intercept on confidence number. 3. For unequal Weibull slopes perform operation for each slope and average the confidence numbers so obtained. Fig. 84—Contidence nomograph at mean tite level 1.92 MEAN LIFE RATIO Het 3 Pra By a eet DEGREES OF FREEDOM ya PateWEIBULL SLOPE, b 115 10 12 99.95 13: 99.5 14 90 15: 16 6 18 2 2 ° g 348 S780 w £ 2 Z ats a s 2 a 5 = = z 61a 3 3 ‘i 2 8 s+70 Z z g 10 8 2 2 gt 2 60 55 104, DEGREES OF FREEDOM 1. Connect total degrees of freedom with Weibull slope and locate intersection point on auxiliary line. ratio with intersection point and continue to intercept on confidenice number. 3. For unequal Weibull slopes perform operation for each slope and average the confidence numbers so obtained. Fig. 85—Contidence nomograph at Bro ie level 1.933ON3GI4NOO LN3ONad LIFE RATIO Fig. 8.8—Contldence interpolation[7 weiput, SLOPE 30 . — — ss 4 20 — + | LIFE Minus “a” LiFe (1000 cycLes) (1000 cycLes) 10 . 1 2 30 40 60 60 70 8090100 200 °«9002=«:SCC CT Fig. 87—Welbull plot for springs with minimum lite greater than zero 1.95Chapter 9 Material and Processing 1, Steel ‘The basic requirement of a leaf spring steel is that the selected grade of steel must have sufficient hardenability for the size involved to insure a fully martensitic structure throughout the entire section. Nonmartensitic transfor. mation products detract from the fatigue properties. ‘Automotive chassis leaf springs have been made from various fine grained alloy steels such as: (G92600 (SAE 9260) 40680 (SAE 4068) G41610 (SAE 4161) (G61500 (SAE 6150) G86600(SAE 8660) GS1600(SAE 5160) G51601(SAE 51860) HS1600(SAE 5160H) G50601(SAE 50B60) In the United States almost all leaf springs are currently made of chromium steels such as G51600, G51601 or their H equivalents. For example, with G51600, the chemistry is specified as an independent variable (while the hardena- bility of the stee! is a dependent variable which will vary with the chemistry). H51600 is essentially the same steel ‘except that the hardenability is specified as an independent Variable (while the chemistry is a dependent variable which may be adjusted to mest the hardenability band requirement) In general terms, higher alloy content is necessary 10 insure adequate hardenability when the thicker leaf sec~ tions are used, When considering the grade of steel to be used, it is recommended either that the hardenability be calculated from the chemistry (for example, G51600), or that the hardenability band charts (for example, H51600) for the various H steels, as published in the SAE Hand book, be consulted. ‘The following “rule of thumb” may be useful for cor- relating section size and steel grade: Thickness-mm Steel 80 max 51600 16.0 max 151600 31.5 max 51601 2, Mechanical Properties Steels of the same hardness in the tempered martensitic condition have approximately the same yield and tensile strengths. The ductility, as measured by elongation and reduction of area, is inversely proportional to the hard- ness. Based upon experience, the optimum mechanical properties for leaf spring applications are obtained within the range of Brinell hardness numbers 388-461. This range contains the six standard Brinell hardness numbers 388, 401, 415, 429, 444, and 461 (corresponding to the ball indentation diameters 3.10, 3.05, 2.00, 2.95, 2.90, 2.85 mm obtained with an applied mass of 3000 kg). A specification for leaf springs usually consists of range covered by four of these Bhn’s, such as 415-461 for thin section sizes. Measurements of typical mechanical properties of leaf spring steel are given below: Hardness: Bhn 388-461 (3000 kg mass) Brinell indentation diameter 3.10~ 2.85 mm Rockwell C 42-49 Tensile strength: 1300-1700 MPa Yield strength (0.2% offset): 1170-1550 MPa Reduction of area: 25% min Elongation: 1% min 3, Surface Decarburization Surface decarburization may reduce the fatigue durability of the springs; therefore, itis important that surface decar- burization be at a minimum. Hot rolled steel bars, as received from the mills, have some decarb, at least of the minimum Type 3 (see SAE 3419 and SAE J1123 in the SAE Handbook), where more than 50% of the base carbon content remains atthe surface (i.e., some partial but not more than 30% loss of carbon). If decarb is of Type 2, where 50% or less of the base ccarbon content remains at the surface (i.e., appreciable partial but not total loss of carbon), the decarb normally does not exceed a depth of 0.25 mm for steels of thick: nesses 5.00 through 12.50 mm, nor a depth of 0.50 mm for steels of thicknesses over 12.50 through 37.50 mm, With sections over 25.00 mm in thickness, some of the hot rolled steel bars may have decarb of Type 1, in which virtually carbon free ferrite (.c., total loss of carbon) exists for a measurable distance below the surface ‘The depth of decarb varies from mill to mill, from roll- ing to rolling, and from bar to bar. The extent to which the depth and type of the decarb can be acceptable will be slibject to agreement between the steel producer and the spring manufacture. 1.97‘The edges of the bars are somewhat higher in decarb than the flat surfaces: decarb on both the edges and the flat surfaces usually has greater depth with increased bar thickness, After forging and non-atmospheric controlled heat treat- ing, the spring leaves will have greater decarb. Scaling of the steel in this processing reduces the thickness of the leaf, While some of the surface decarb is removed. with the scale, the final depth of decarb is usually greater than it was in the steel bats as received from the mills, 4, Mechanical Prestressing Presetting, shot peening, and/or stress peening at ambi- ent temperatures produces large increases in fatigue dura- bility without increasing the size of the spring. These pre- stressing methods are more effective in increasing the fa- tigue properties ofa spring than are changes in material ‘When a load is applied to a leaf spring, the surface layers are subject to the maximum bending stress. One surface of each leaf will be in tension and the opposite surface will be in compression. The surfaces which are concave in the free position will generally be in tension under load, while the convex surfaces will generally be in compression. Fatigue failures of the eaves usually start at Or near the surface cn the tension side. Since residual TENSION 0.25 0, PRESET, COMPRESSION RESIDUAL STRESS. SHOT PEENED stresses are algebraically additive to load stresses, the in- troduction of residual compressive stresses in the tension surface by prestressing reduces the operating stress level, thereby increasing the fatigue life Presetting (synonymous terms are: cold setting, bull- dozing, setting-down, scragging) produces residual com- Pressive stresses in the tension surface and residual tensile stresses in the compression surface by forcing the leaves to yield or take a permanent set in the direction of subse- quent service loading. While this operation is beneficial to fatigue life, its primary effect isthe reduction of “settling” (Goad loss) in service. Presetting is usually done on the spring assembly. Shot Peening introduces compressive residual stresses by subjecting the tension side of the individual leaves to ‘stream of high-velocity shot. The SAE, “Manual on Shot, Peening.” HS-84, deals with the control of process varia- bles, while techniques for control of pening effectiveness and quality are explained in, “Procedures for Using Stan- {dard Shot Peening Test Strip,” SAE 5443. Cut wire shot, size CW-23 to CW-41, and cast steel shot, size $-230 to S-390, are generally used for this purpose. Shot pening intensity is expressed as a dimensionless Almen number. The intensity of shot peening applied to light and medium type springs is usually in the range of 10A-20A as read on the Almen gage. For heavy springs, the intensity is usually 6C-14C as read om the Almen gage. Coverage in both cases should be at least 90%. DEPTH BELOW TENSION SURFACE, mm 150 075 1.00 Fig. 9.1—Beneticial stress patterns induced by presetting and ening in the absence ofStress Peening (strain peening) is a means of introduc- ing higher residual compressive stresses than is possible on 1 ow 1 H+ Belt-H 1 o ° © <1 tn Yi Hee em det Hlt—el ves Vi see Fig, 10.3 2 + AVAL + de + Hts) (000 Fig, 10.11 (son Fig, 10.2 with H = 0) (aoe Fig. 10.2 with H-0.1,02,03,04,05) (soe Fig, 10.1 ' 2—e? 00 Fig. 10.5 190 Fig. 105 — 0 Fig. 10.7 with 2-4 3Ut~Jellde8) ~ 6 nde H0.1,02,03,04,05 (swe Fig. 105) (soe Fig. 10.7 with = 0}, 21-49)" 10.8 with 10 Fig, 10.6 00 Fig, 10.6 ) ~ Ble? Andy (ase Fig, 10°8 with H = 0) 10:1,02,03,04,05‘TABLE 102-VOLUME, RATES, EFFICIENCY OF VARIOUS TYPES OF SINGLE LEAF SPRINGS (CONTINUED) Fie Type a Fa 3 ZI ma Peak true Pa) Spears | 0 8 50 S0 8 Undee Verte! Unitorm Over | rom Encasement | Unitorm Over tod P(N Entire Length | to Distance | Entire Length So Stes at {Yor locaton of Speak emo) Enearement tee Chapter 10, eee Sxction 4] wate? Elutic Energy (mJ= New): Energy Factor= | 1 1 1 1 1 neray Fecior wotoe Wwe & Volum Etficieney (mi/mn | Shoat Shak 1 Sheak Shenk 3-H on. | — We SEI) Gy se we tw ues for length, width, and rate are specified the thickness then remains to be established. wth The vertical rate for the cantilever spring is Eustl cy pan aaibe Ew, t k= E&ete oc, and ae where C, is the vertical rate factor for whatever single leaf eek ed inder c ° EwL Cy type is under consideration (see Section 3 in this chapter). ‘The subscript "o” refers to the values at the line of encase- ment for the cantilever spring or at the start of the active length for each half of the semi-elliptic spring (located within the axle seat area as discussed in Chapter 5, Section 2), In the conventional semielliptic spring, the thickness t, can be established by combining the rate formulae either for the fixed end or for the shackled (or free) end. The rate formulae at the fixed end are combined to y= - Bet gy In the same manner, the lateral rate for the semi-ellipti spring can be determined from kb? _ Et) ky = HE = Eat oc, aL ae then EtawiL ky = Eid eb where: C, = lateral rate factor. In the unconventional spring, the value of “Z” will be established by means of the formula showin in Table 4.1 1.106t Peale Troe Towed Tee Fe 7a ma 7 1 % % % For 4205: 59 For 4205: 89 From Encooent| Uniform Over | From Encarement ‘o Dinone "| Ene Cogn Drone % ene eno For ec For tyc05:——— wat) alto 1 1 1 Foc 208: 1 For a>osit For WSO: ie Fortec: Bde"? shea 2-iP| shat Set For 208 For 4208: ce 20 fw | we wee eax 1 Shea 1 + 4 14 es = | Bette) Ge Seine Ie ve Paco me Faor vo ee For Je<0.5: For Je 108 oe cc + IP ey COC + 1 & 23460, € > 7.1 (by tial solution) acP(c +P 4.8, 7D? 8X 13000C(C + 1) 8x OWE + DL om X 125? ao CIC + 1) < 406; C < 6.7 (by trial solution) Since no value of C will satisfy the requirements for both L, and S,, two springs must be considered. B—Two Springs: by trial select C = 8 125 a+ 1d = 19 8x 13911 8x 108 X 19 300 X 125 x 65 ox 52x BX i & 8.3010. Ly = Nu dy = 830 x 14.7 = 122(0K) 79 300 x 125 LR, = BMX 5 _ ay, : Tx 830 8 9 ~ 4 ett 12. Na = 830244 = 107 s-17 br, = n4[ET i] = 196 13000 x 324 4p, = POON 14 Py 8 8100 8x 8 x 8100 x 92 ss, = SX Sx Sox f ie 855 (= 8.) (OK) 16, vy = EXE @ 1.59 x 10° + 17. V, =, 1.50 108 x 0498 = 747 x 10% mm Helical springs of rectangular wire are used in special applications where a greater volume of material is required in a limited space, and they may provide a solu- tion to a problem of high loads in a small space. A helical spring of cylindrical or conical form, made from a round bar of tapered cross section, is feasible and finds an application as a vehicle suspension spring. The variable bar cross section can impart a variable rate to the spring as the smaller diameter portion of the bar closes solidly during the early stages of spring compression. This feature can provide improved ride ‘‘characteristics” to loaded vehicle, The use of a papered bar, combined with variable coil pitch, reduces the linear space and spring mass requirements. Manufacturing methods have been developed to form the tapered round bar prior to coiling. ‘The tapered bar helical springs can be designed with the tapered bar section at one or both ends of the spring, depending upon the application and the characteristics desired, Due to the added variables involved, the design Of these springs becomes complicated but can be handled by developing modifying factors on the standard design formulae and then experimentally verifying the spring characteristics. Another approach is to develop new for- mulae to encompass the added variables and constrai imposed on the variable rate and the stress distribution along the bar length. 2. Static Versus Dynamic Loading ‘The term “‘static loading" implies a force application of sufficiently low velocity that the spring is not deflected beyond an equilibrium position, and deflections and stresses are stabilized throughout the spring. In practice, 26 a force is usually applied until the deflection is limited by some external means. If this position is reached at a ow velocity, no surge within the spring will occur and it may be considered as statically loaded. When a mass W (kg) is gradually lowered on a vertical helical spring of rate R (N/mm), assuming that the standard accelera- tion of gravity g = 9.81 m/s prevails, and since force ‘equals mass times acceleration, the mass W (kg) will exert a vertical force P = Wg = Wx9.81 (N)on the spring, when at rest, and a deflection F = P/R (mm) will result. ‘The term “dynamic loading" implies an application in which the velocity of the moving spring end is apprecia- ble. This is true for all impact loadings as well as for sinusoidal loadings under high frequencies. Because of the inertia of the spring coils, the whole spring deflec- tion is not equally distributed among the coils so that some of them must undergo a stress range greater than the corresponding stress range with static loading. For example, let the mass W be suddenly released from a point where it just touches the top of the free vertical hel- ical spring of rate R. In this case, the spring will undergo a maximum deflection F, before it comes to rest at the ‘equilibrium position under deflection F. ‘When the mass descends a distance F, (mm), its poten- tial energy decreases by W g F,/10? (J); assuming no damping, the strain energy in the spring increases by RFY/Q x 10) (). Therefore: w 2 ah _ RF We, 2248 op 1 2x 10" R R It then follows that, for this instance of dynamic load- i, the maximum deflection and, consequently, the max- imum stress, is twice as large as would result from the gradual (static) application of the same mass. When the ‘mass is released from some point higher than the top of the free spring, then the maximum stress is more than twice the static stress. ‘Two basic instances of dynamic loading can be distin- guished. In one instance the stroke is practically constant. This applies to the driving spring of a machine gun, since the striking part is finally stopped by a stiff buffer act- ing in parallel to the driving spring. The other instance is that of a buffer spring, which absorbs almost the full kinetic energy of the striking part and therefore has a vari- able stroke. In the first instance, the spring itself may take a greater deflection than that prescribed by the striking. part owing to the kinetic energy of its coils. In addition, at the minimum compressed length, the whole deflection of the spring is not equally distributed among the coils, so that some of them must necessarily take a higher than average stress. Therefore, the conventional formulae give too low a maximum stress, and special formulae are required for the maximum dynamic stress.‘An example of dynamic loading will serve to illustrate the relationships. A precompressed spring is suddenly compressed by a heavy mass moving with constant veloc- ity v over a given stroke. The motion of the front end of the spring causes a surge wave which is propagated along the spring wire with surge velocity c, while the coils themselves move axially with velocity v. This wave may be called “‘single wave," as long as no reflection occurs at the rear end of the spring. The amplitude of this sin- gle wave can be expressed by one of the following 1. Change of load, AP (N) 2. Change of stress, AS (MPa) 3. Change of deflection per coil, AF (mm) thas been found that all these changes of load, stress, and deflection per coil are proportional to the velocity of loading v (m/s). They may be expressed as follows: ap ne N Fert x 10 = 72. vi76 (S) as as_t MP2 v ap iY? 7S (aa) os 1s rDe ey @ oR Te wn) where: T, = travel time of the wave over the whole wise length, or “surge time" s = 7850 kg/m" for stee! G = 79 300 MPa \ Ga) 88 <3 cu) ar _ wo va™ ‘mn wo (rn) “The additional stress AS is independent of the spring dimensions; for any compression spring made of steel, cach m/s of impact velocity increases the stress by 35.3 MPa. ‘After reflection of the single wave from the fixed end of the spring, the incident and the reflected waves are superimposed by. ordinary addition of the stress ampli- tudes. Therefore, the maximum stress built up in the a7 spring depends on the stress amplitude of the single wave as well as on the number of superimpositions of the sin- gle wave over the duration of the compression stroke. While the stress amplitude depends upon the impact velocity only, the number of superimpositions is a func- tion of the surge time T,, that is, the spring dimensions. With slow dynamic loading during compression time, there are many superimpositions of the surge wave of small amplitude. Therefore, the dynamic stress will not noticeably exceed the static stress. However, with rapid dynamic loading, the compres- sion time can be of the order of the surge time. There are few superimpositions, but the amplitude is high and, consequently, the dynamic stress can exceed the static stress substantially. In such instances, it may be advisa- ble to tune the surge time T, with regard to the compres- sion time by proper selection of the spring dimensions. In most instances of dynamic loading, the analysis of stress is not as simple as in the example given. However, in general, two groups of factors appear to be decisive for the maximum stress built up and the spring life expected: 1. The “loading function” v(t), that is, the time- velocity function of the moving end of the spring, with stroke, compression time, maximum velocity, and maxi- ‘mum acceleration as important parameters. 2. The spring, characterized by its material, maximum, static stress, surge time T,, and inherent damping. The influence of the loading function v(t) on the life of springs has been strikingly proved by fatigue tests: The average life of springs under sinusoidal loading may be reduced as much as 50:1 under impact loading, with the same stroke and frequency. However, the importance of ‘each individual life factor remains to be explored by more fundamental testing. Gun springs are among the springs which are subjected. to severe dynamic loadings. The following methods of design have been successfully tried: 1. Single wire springs—Proper adjustment ofthe surge time T, to the given compression time reduces the max- imum dynamic stress. 2. Stranded wire springs, coiled from 3-wire or 7-wire strand. Their greater inherent damping improves the spring life. 3, Two-section springs that i, a front spring in series with a somewhat stiffer rear spring. In comparison with a single spring, @ two-section spring gives more possiil- ity of adjusting the spring assembly o the given loading function since a more equalized distribution of stresses over the wire length can be obtained at the moment of, ‘maximum compression. This results in lower peak stresses and better life than with a single spring.3. Natural Frequency With dynamic loading of compression springs, the iner- tia of the spring coils, in addition to their elasticity, becomes effective. When moving one end of a precompressed free spring in an axial direction, the dis- turbance of the equilibrium at this end is propagated along the spring wire in a transient wave with a definite surge velocity. where: mass of active part of spring = W = wo #dN,Dy 4g @ < spring rate.= R = @ Length of active part of spring = L = DN, (4) Substituting W(2), R(3), L(4) in (1) 6 of meter, as follows: ©) a 8) 9,806 650 m/s! 79300 x. 10 N/m! (for Steel) 850 x 9,806 650 N/m? (for Steel) @ G aa Here all length units must be used uniformly in terms | Substitute g(6), G(7) and (8) in (8) c - 1247 4 (mm/s) (9) D sav ‘The travel time of the wave over the whole wire length or ‘surge time” is (L, D, d specified in mm here expressed as m) 2DN, = DN) 10C ) with mean value for spring steel DN, Baa) 10% 6, an 28 Active solid length of the spring = L, = Nyd (12) race G) 6-8 Gb T, increases with the active solid length and with the index of the spring. A spring of index D/d = 7 has a surge time of 69 x 10° s/mm of active solid length. In some instances, another formula can be used to good advantage. (4) where W(kg) is the mass of the spring and R (N/mm) is the rate. When both ends of the spring are supported, as is, usual, one cycle of vibration comprises just twice the surge time T,, therefore, the natural frequency of the spring, measured in Hertz, is: 2.247 x 10d rr, 2D°N, = 358. 10% Gy = 4 Hy as orf = 388 x 10% ayy a9) LD —1_ ote 2V Re -issVR a 0D ‘The natural frequency decreases when the spring index and the active sold length are increased. A spring of index D/d = Tand an active solid length of 250 mm has a fre- quency of 29 Hz. The natural frequency of valve springs, as used in automotive engines, tends to be somewhat lower than that calculated from the above formula. Actual measurements reveal that the frequency may be as much as 10% less ‘than that calculated, due, evidently, to the end effects introduced by more active coils during a spring surge. NOTE: It is nevessary to distinguish clearly between the natural frequency of the spring itself, as described above, and the natural frequency of a mass W, (ke) attached to spring of much smaller mass and rate R (N/mm).‘The natural frequency of such a mass W, is: ‘The mass W,, when attached to the spring, will cause a static deflection F = We R ‘The natural frequency of the mass W,, relative to the spring deflection F (measured in cycles/minute) 4s 945 (cycles/minute) VF f= 159 x 60 Residual Stresses Residual stresses (sometimes called “trapped stress”) may be set up in springs in a variety of ways including heat ‘treatment, shot peening, and hot and cold presetting. As a rule, the presence of residual stresses of proper sign ‘makes possible the utilization of higher working stresses; in springs, this is of considerable advantage since the energy storage capacity increases as the square of the working stress. For coiled springs, the shot peening and presetting methods of inducing residual stresses are most practical. Shot Peening The shot peening process consists essentially of sub- jecting the spring to a stream of shot moving at high velocity. The peening action of the shot sets up benefi- cial compressive stresses in a thin surface layer and also results in cold working this layer. The depth of this cold-worked layer depends, among other things, upon the size and impingement velocity of the shot used, the direction of impingement relative to the surface, and the spring material. Shot peening, when properly con- ducted, results in relatively large increases of the stress range through which the spring can safely operate under fatigue or repeated loading conditions. In commercial use, two types of methods are most adaptable. In the earlier method, air is used for the propelling agent as in the common sand blast equipment. ‘The second method is mechanical, the shot being fed on a rapidly rotating wheel having radial blades. Thus, the energy of the wheel is transferred to the shot by the blades and by centrifugal force, the shot in turn impinging on the parts. Tests have shown that either method can pro- 29 duce the same results in increasing fatigue life, but the mechanical method is more economical for high production. ‘The increase in fatigue life is accomplished in two ways, One consists in stressing the surface layers in compres- sion. Fatigue failures are due to tensile stresses and, if ‘a compressive stress.can be set up in advance where sub- sequently the load application will produce a tensile tress, the part will carry considerable load before itis actually stressed in tension. The second effect is the cold work- ing and strengthening of the surface layers which raises the physical properties where the highest stresses occur. ‘This second effect is of minor importance compared with the first, Because of the twofold beneficial results imparted by shot peening, this surface treatment has done ‘more to increase the life of springs than any alloy steel or other process ever employed. Shot pening can be successfully carried out on wire sizesrin excess of 0.75 mm. But it will be effective only if itis thoroughly and evenly applied over the entire sur- face, with particular emphasis on those parts of the sur- face which will undergo the highest load stresses. Thus, ‘on compression springs the inside of the coil is most important, whereas on torsion springs, it is the out that is important. Time of exposure to the shot is one of the prime factors to obtain effective pening, as in¢ cated by the curve in Fig. 1.2. For each type of spring, a knowledge of this time factor, with a definite type and size of shot, is needed ~—wwrensiry oF Penne GAGE READING Fig. 1.2~Intensty determination curve The effects of shot size and peening time can be checked by means of Almen strips and other observ: tions. The height of the arc in the Almen strip is not itself a measure of fatigue lite, but it furnishes a check on uniform operation of the shot pening machine. The ‘SAE Manual on Shot Peening (HS 84) details the methods for controlling the machine with the help of Almen strips and for measuring coverage. The importance of proper Peening coverage cannot be overemphasized, ‘Since shot peening is a mechanical working process, it can be affected by heat. If springs are heated suffi-ciently after shot peening (Fig. 1.3), the benefits will be reduced and the endurance limit will drop, eventually to that of the wire as originally received. The use of shot peened springs in applications involving high tempera- ‘ures necessitates careful planning. If the spring is heated above 430°C during manufacture, the maximum resis- tance to settling at temperatures usually encountered in. automotive practice is obtained. But in this event, the beneficial shot peening stresses are completely removed, and no increase will occur in fatigue life due to the shot peening operation. Therefore, compromises must be made if an endurance isto be realized which exceeds that Of the untreated wire. Fig. 1.3—Reduetion in effectiveness of shot pening by subsequent heating or stress relioving Presetting Residual stresses are most commonly set up in helical compression springs by means ofa presetting operation, known also as scragging or cold setting when performed at ambient temperatures. When the presetting operation is performed at an elevated temperature it may be called hot pressing, heat seting, or warm setting. The need for presetting depends upon the design stresses, the application and its conditions and require- ments. The use of presetting is most beneficial when design stresses are at or near the yield point, and settling or sag prevents the spring from performing as required. Presetting is an operation that is performed during the manufacturing of helical compression springs in which the spring is compressed heyond the yield point of the niaterial. The yielding of the surface layers of the wire which occurs during the presetting, produces beneficial residual stresses, thus increasing the elastic limit of the spring and thereby reducing the amount of setling or sag in subsequent service. The spring is initially coiled to a free length greater than the final desired free length and is then compressed to a point where the spring may be at or near the solid length (Fig. 1.4, point B). On releas- ing the load, a permanent set OC is produced while the spring behaves nearly elastically along line BC. In spe- cial instances several additional compressions to point B may be carried out, DEFLECTION ° e ¢ . Fig. 1.4—Typical load-deflection diagram of helical spring during preseiting ‘When presetting on helical compression springs at some elevated temperature, the operation known as hot press- ing, heat setting, or warm setting may be employed to produce greater resistance to relaxation than cold setting. Heat setting consists of compressing the spring on a fixture, subjecting the compressed spring to a tempera ture higher than the desired operating temperature for a time suitable to insure complete penetration of the heat, and then cooling to room temperature before releasing For a description of warm setting, see Chapter 4, Sec- ton 5. Hot pressing consists of heating the spring in its free or relaxed position to some temperature for sufficient {ime to insure complete penetration; then while the spring is at the temperature, it is compressed to some height below the installed or operating position and released. Fig. 1.5A represents the stress distribution over the cross section at a load which gives a stress § somewhat below the yield point. For purposes of illustration, itis assumed that no strain hardening occurs, even under the higher presetting load Gg. 1.4, point B) with which Fig 1 5B deals. Here, yield 18 occurs at stress S, and the distribution of stress wi be as indicated by the shaded area f-h-i-k. The nominal stress calculated from the prsetting load by the usual for- mula is equal to S, and without yielding the stress tribution resulting from this presetting load would be as indicated by the triangular area f--k. Actually, yielding has started at the surface (wire diameter d) and has progressed inward to a diameter: 4, = Y= SA, In Fig. 1.5B, the ratio S,/S, has been chosen as 31/24, In that instance, d, becomes d/2. ‘When the presetting load is reduced to zero (Fig. 1.4, point C), residual stresses are set up as indicated by the 210Ye Fig. 1.5£tfect of presetting on stress distribution over cross section of helical spring shaded area in Fig. 1.5C. These are obtained by subtract- ing the ordinates of the dashed line f-1 from those of the broken line f-h-i in Fig. 1.5B. The reason for this is that ‘on removal of the load, the spring will behave approxi- mately elastically. Therefore, removal of the load is ‘equivalent to subtracting the clastic stress distribution f-1 from the plastic distribution f If the spring is loaded subsequent to presetting, the resultant stress distribution is found by adding the load. stress distribution to that of Fig. 1.SC. For example, a load may be applied to the spring which gives a stress S, the same as that shown in Fig. 1.5A. The resultant stress distribution under that load is indicated by the shaded area f-0-p-a in Fig. 1.5D. It will be seen that, due to the presetting operation in this instance, the peak stress has been reduced from a value S (Fig. 1.5A) to a lower value S, (Fig. 1.5D). Still dealing with a round wire helical compression spring of large index and still assuming that no strain hardening occurs, it may be considered that the material Ids a an essentially constant stress equal to the sheat ing yield point and that complete yielding, over the entire ‘ross section, will occur theoretically at a load equal to 1.33 times the load at which yielding first starts. This ‘would indicate that, for these theoretical conditions, the a ant ‘maximum presetting load would be about 33M% above the load corresponding to the torsional yield stress and that, after presetting, the spring would behave elastically up toa load corresponding to a stress 1.33 times the torsional yield point. Because of curvature and strain hardening effects, the ratio of the maximum presetting load to the load at which yielding first starts, tends to be higher than 1.33 in actual springs, and, in some instances, for the smaller spring indexes, it may approach 2. Since the spring will behave elastically nearly up to the presetting load, this means that the elastic limit of a spring may be increased considerably by presetting; much of this increase is attributable to the residual stresses set up in the presetting operation. In actual practice, presetting deflections equal to 2, and in some instances even 3, times the elastic limit deflection may be used (OD = 2t0 3 times ‘OE in Fig. 1.4). tn other instances, however, lower values of presetting deflection may be required to avoid exces- sive distortion of the spring. In cold wound helical torsion springs, residual stresses are also set up by the coiling operation; tension residual stresses are set up on the inside of the coil and compres- sion stresses on the outside. If the spring is loaded by a torque so as to wind up the coil, it is clear that the residual stresses, in this instance, will also subtract from the max-imum load stresses near the outer fibers. To take advan- tage of the residual stress pattern, torsion springs should therefore be loaded in the windup direction. Residual stresses set up in cold wound helical compres- sion springs as a consequence of the winding process are usually removed or reduced by a low temperature, stress- relieving (bluing) treatment. These stresses tend to aug- ment the effects of the torsional working stresses in such springs, and hence, their removal is beneficial. Temperature Effects on Springs Spring design becomes complicated for extreme tem- perature applications. Generally, the usual spring ‘materials will operate satisfactorily at very low tempera- tures under static loading conditions. The combination fof extreme cold temperature and impact loading represents a more severe situatioi; however, it may not be catastrophic because of the resiliency of the spring, material. ‘Under high temperature conditions, springs undergo load losses which in most instances are due to a decrease in the torsional elastic limit of the material. Springs may bbe designed to operate satisfactorily at normal tempera- tures, but under elevated temperatures these same springs undergo a slow piastic deformation (commonly known as “relaxation” when the spring is at a fixed length, or a “creep” when spring is under a fixed load) land take a permanent set. The amount of set depends upon the working stress level, heat intensity, and the dura- tion at the elevated temperature. See Fig. 5.7 in Chapter 5. ‘Thus, in any spring problem involving elevated tem- perature, the designer should be fully aware of the impending difficulties and make suitable allowances for all the factors involved. In general, the spring stresses should be reduced to the lower practical values. How- ever, there are applications where such modifications are not possible, and in these instances, some form of preset~ ting may be used. See Chapter 1 - Section 4. ‘The temperature, time, and stress level of any of the presetting methods may be varied to minimize the loss in load when subjected to the stresses and elevated tem- peratures subsequently imposed in operation. This relax- ation or sag may be reduced as much as 50% when compared to a spring that has not been preset. For recommended guidelines on selecting materials to resist the effects of elevated temperatures, see Table 2.1 in Chapter 2. If springs are subject to elevated temperatures during. operation, residual stresses may be greatly reduced or ‘even lost. For example, shot peening begins to lose its effectiveness at temperatures around 260°C, with complete loss at temperatures around 430°C on steel springs. This indicates that, at these temperatures, residual stresses are 212 considerably reduced, and the benefits of presetting will be reduced correspondingly. 6. Fatigue Durability ‘Some springs are required to function cyclically with accompanying repetitive stresses. Failure of these springs is commonly known as fatigue. Generally, these springs perform an essential function and should be designed to have high fatigue durability. This applies particularly to valve springs for internal combustion engines and to injec- tor springs for diesel engines. Designing a spring of this type requires a consideration of the stress and stress range, the waveform and harmonic content of the cyclic stress, the environmental conditions, and the specialized surface treatments such as presetting and shot peening. From fatigue test data on cyclic loaded springs an S-n ‘curve (endurance factor S¢/Sy versus life in number of cycles) can be plotted as shown in Fig. 1.6. This is a typi cal log-log plot of the endurance factor Ke which is nor- malized to represent the ratio of the peak alternating stress as a fraction of the minimum ultimate tensi strength of the material. The log-log coordinates permit the test data to be reasonably well represented by straight lines between 10° and 10* cycles of life. Spring steels develop a “knee” in the curves at about 2 x 10* cycles beyond which infinite life cau be expected. oF so 08 108 10 LIFE, CYCLES (leg) Fig. 1.6-—Typical Sn curves for helical springs Springs under a fluctuating load normally are alter- nately stressed about a mean level with the actual stress value not going through zero. The S-n curve of Fig. 1.6 refers to a completely reversed alternating stress with a ‘mean value of zero. Therefore, another form of diagram, known as the “Goodman-type Diagram’” as illustrated in Fig. 1.7, is employed to correlate the actual stress range with the conventional S-n diagrams by determing an equivalent value of Ke. As an example, illustrated in Fig. 1.7, a typical stress range between the initial stressof Kz, reference line and the maximum stress Ks: on the “Goodman” line establishes Ke, which is one-half of the overall peak-to-peak cyclic stress range and represents a spring having equivalent life. These two diagrams are ‘the basis for the three fatigue strength diagrams presented. in Chapter $ as Figs. 5.1, 5.2, and 5.3 for use in actual spring design work. They are intended to establish safe stress limits consistent with expected life endurance. YIELO CRITERION: 5 AT STATIC SERVICE ENDURANCE FACTOR Ke INITIAL STRESS FACTOR Kst Fig. 1.7—Goodman-type fatigue strength diagram for helical spings It should be understood that the S-n curves and “Goodman-type Diagrams" are limit lines established from accumulated test data by many investigators. They hhave no theoretical basis, but are conveniently presented in mathematical and graphical form for design purposes. ‘The lines given tend to be conservative and are intended only to serve as a guide for the designer. In the final appli- cation, the designer should consuli with the spring ‘manufacturer to establish realistic values for stress limits in cyclic stressed springs. The following formulae are the basis for developing the fatigue strength diagrams. They may be used in con- junction with computers for spring design calculations instead of using the fatigue strength diagrams directly. Ke = Cent Ko(Ks: — Ks) _ Se 2Ku — Ke + Ks) Se 2KcKe + Ks(Ke — Ke) ror =A+BKs Pe Min. TS oe ale 8] Max. Kez = sn! A= Keats, = 0 B= (Ko ~ AV/Ky Where: Ce, Cs M and ¥ are constants Ke = Endurance factor Ks, = Initial stress factor Ke = Maximum stress factor Ky = Ultimate strength factor Life cycles Spring load Endurance stress Initial stress TS = Ultimate stress (min. tensile strength) Refer to Tables 5.3, 5.4, and 5.5. In selecting a spring material for prolonged fatigue durability, first consideration should be given to carbon steel valve spring quality wire. In the normal sizes, it usually performs as well as alloy steel and is generally more dependable. In most instances, the use of pretem- pered steel is advisable to avoid hardening cracks that can ‘occur in material hardened and tempered after coiling. Alloy steels are usually more subject to seams and have a greater tendency toward quench cracks than are car- bon steels. Therefore, it is advisable to subject such springs to magnetic particle inspection. While this proc- ess is not ordinarily used for springs in the automotive industry, itis seldom omitted on aircraft engine springs. Another technique, more recently developed and com- monly used for quality spring manufacturing, is magn inspection of spring wire at the coiling machine. . Common Causes of Spring Failure A spring designed and produced with due regard to all the foregoing details should not fail. In spite of this, a rare failure may occur on a properly designed spring. It can usually be traced to the following causes: 1, Surface Imperfections—They occur as hardening cracks, seams, pipe, laps, pits, die marks, and tool marks. 243‘Of these, magnetic particle inspection will locate only the first three. The other imperfections must be found by careful inspection of the raw materials, usually by etch- ing both ends of every bundle of wire and by visual inspection of the finished part. The difficult part of this procedure is that the small sharp pits or marks may be overlooked while the broad, round bottom marks, which are harmless, are easily observed. For critical springs such as automotive valve springs it is common practice to inspect the wire continuously by an electromagnetic device which rejects springs with defects. 2. Corrosion—This is a more common cause of spring breakage than is usually understood by the spring user. In valve springs, it may be caused.by condensation of the products of combustion on the springs. A small pit is formed or hydrogen embrittlement has occurred from the oxidation of the steel. It is not generally recognized how fast this process can lead to breakage. Springs which were perfect have failed in’3-6 h in an engine which was started and stopped before it could run long enough to scavenge the crankcase. In applications of this type, it will be neces- sary to provide corrosion protection for the springs. (Refer to Section 8 on surface protection.) 3. Improper Heat Treatment—It is a threat which must not be overlooked, even though it is not frequently encountered. If springs are overheated, the grain struc ture will be coarse and the fatigue life poor. Failure to hheat to a proper temperature, of tc heat long enough to ‘get all the carbides into solution, will result in ferrite patches and low endurance limit. Automatic time- ‘temperature control in modern heat treating departments has almost completely eliminated this once common cause of breakage. Decarburization—This is the least frequent offender in the entire list of imperfections. Partial decarburization is usually present in spring wire, at least toa slight extent. The degree of permissible decarburization will depend it is used. When the examination by a competent metal- Jographer reveals a full ferrite ring around the circum- ference, the wire is subject (o rejection because the surface condition will ultimately lead to premature failure. ‘The foregoing data cover the general considerations which deal with fatigue durability springs, and some of the special methods which have been developed to meet the conditions where the consequences of breakage may be disproportionally serious. 8. Surface Protection ‘The most widely used metals in the spring industry are the carbon and low alloy steels. These are also the least cortosion resistant. The wide acceptance of these steels as spring material directs consideration toward choice of the proper type of finish for a particular spring applica- tion, The most important factors to be considered are: 1, Corrosive severity of the spring environment. 2. The degree of corrosion protection required to give spring life expectancy. 3. Effect of coating and its method of application on the mechanical properties of the spring. (See guidelines in following paragraphs to minimize hydrogen embrittlement.) 4, Comparative corrosion resistance and the cost of applying the coating. 5. Effect of the protective coating on associated members. 6. Availability of the protective materials and the equipment necessary for the application of this coating. Protection From Mildly Corrosive Environment For many applications, springs are coated with one of several slushing oils or rustproofing compounds. A more durable coating, if required, is a phosphate coating with rustproofing compound or organic finish. The organic finish can be compounded with low, medium, or high oil content in order to give the degrees of protection required, though baking may be required to dry the springs coated with compounds containing the higher oil content. An example of an organic coating with a low oil concentra- n is the asphaltic-based material known as Gilsonite paint or Japan Black, while the medium and high oil con- centration paints are alkyd enamels. The addition of zine ‘chromate pigment to these synthetic enamels may be used to further enhance their corrosive protection. ‘When springs are protected by applying a zinc or iron phosphate coating, the thickness ofthe coating depends, ‘among other factors, on the precleaning process. When ‘an emulsion cleaner is used, a very fine coating is formed; an alkaline cleaner produces a thicker, hard, ti ing; if an acid is used, the coating is coarse and very thick. ‘A.uniform coating depends on good cleaning and chem- ical control. This phosphate coating isa bass forthe sub- sequent coating with a rust preventive oil, wax, grease, ‘or organic finish providing good protection from corro- sion. The mechanism of these coatings is the ability of the minute pits, surface irregularities, and the erystalline structure of the zinc and iron phosphate crystals to retain the subsequent coating, thereby providing maximum cor- rosion protection. The phosphating process must be tai- lored to protect the spfing from acid pickling, both in the precleaning sequence and during the phosphate proc- ess by controlling the free acid ratio. When the end use of the spring is in a mildly corrosive atmosphere, a black oxide finish provides some degree 214of corrosion protection. A biack oxide coating provides some of the same mechanics to hold rust preventive materials as the zinc and iron phosphate coatings, but with an overall lesser degree of protection. When black oxide coatings are applied they have the advantage of causing very slight, if any, dimensional changes and no hydrogen embrittlement. Frequently, itis advantageous to use precoated spring. wire where mildly corrosive conditions are encountered. ‘These wires are coated by dipping into molten metal or by electroplating. They may be coated and then drawn to size, or they may be coated at the finished size. Cad- mium, tin, and zine are the metals employed for the wire coating process. The commercial importance of the proc- ss is considerable because the parts formed from such ‘wires may be used immediately without having to undergo additional operations normally associated with the plat- ing of finished parts. In the case of extension springs, the use of precoated wire is preferred because the tightly wound coils prevent applying a satisfactory coating after coiling. Electroplating Many springs are being plated with a variety of metals of which cadmium: and zinc have generally proved to be the most satistactory ‘Cadmium and zinc plate are bluish white metals hav- ing good ductility and providing a protective coating on springs for most applications. When properly applied, cadmium or zine plating does not flake or peel. They pro- tect the base metal by sacrificial galvanic action. The pro- tective life of both zinc and cadmium plate may be extended by adding a chromate coating. The degree of protection will still be dependent upon the thickness of the plate which may vary from 4 to 13 um. Neither of these metals is resistant to attack from strong acids or alkalis. Both zine and cadmium are poisonous when vaporized. Neither should be used on springs which come in contact with food or drink. ‘The end use of the spring and the environment deter- ‘ines which coating is to be applied. When either zinc or cadmium is indicated, zinc may be chosen because it is less expensive; the choice may be cadmium because of the longer time before producing white corrosion products. Cadmium can be baked at a higher tempera- ture for relief of hydrogen embrittlement without the haz~ ard of blistering. Copper plate on springs should be used only for identification purposes ait provides only slight, corrosion protection for steels. The use of a copper, nickel, or chromium plate may be indicated for decora tive reasons but is not generally applicable. Springs requiring nickel plate may be plated using the Sulfamate, Watts, or electroless process. When the nickel corrodes, 215 it forms green nickel corrosion products, and the’ pro: tection depends on the thickness and continuity of the plate. Plated nickel is dense and prevents the escape of hydrogen which enters the steel during the precleaning, pickling, and plating processes. Hard chromium plating of springs is generally unsatisfactory because of the heavy thickness (50 um or more) required to give corrosion pro- tection. Chromium plate is hard with low ductility; there- fore, it is subject to peeling and cracking while causing reduction in spring fatigue life. ‘When springs are exposed to severe environmental con- ditions such as salt water, acid, and caustic atmosphere or solutions, the most resistant coating is lead alloy. Lead coatings should be 25 ym in thickness to afford maxi- ‘mum protection. Heavy tin coatings, exceeding 25 ym in thickness, should be used for springs coming in contact, with foodstuff because of the resistance of tin to weak organic acids and its low toxicity. Lead and tin coatings are electropositive to steel (more noble in the electromo- tive series); therefore, these coatings will only protect the steel spring when the coating is continuous and of a suffi- ccient thickness, Lead and tin plates are soft materials; therefore, abrading of the coating must be avoided in order to provide the necessary continuous thick protec- tive coating. ‘Techniques for Minimizing Hydrogen Embrittlement Hydrogen may be introduced either during the clean- ing steps in preparation for plating or it may be introduced during the plating process. If springs can be fabricated without producing heavy scale or film, the first hazard will be greatly reduced. Residual stresses should be reduced as much as possi- ble before plating. If scale removal is required, a mechanical process is preferred. Unless the parts would distort, they may be blasted with shot, grit, glass, aluminum oxide, ete., or they may be tumbled or vibrated with various abrasive media. When either a sulfuric or hydrochloric acid solution is used to remove scale'from springs, the use of a suita- ble organic pickle inhibitor will climinate excessive pick- ling and tend to reduce hydrogen embrittlement. Springs with a hardness over Re 40 should not be descaled in acid. ‘The resistance of the passage of hydrogen from the steel increases as the thickness of the plate increases; there- fore, excessive plate thickness should be avoided as well as the conditions producing hydrogen embrittlement. The following procedure should be followed to achieve ‘The anodic or cathodic cleaning and pickling cycle should be as short as possible to reduce possible cracking of the spring in highly stressed areas. The zinc or cadmium plate should be initially deposited at a low amperage until the springs are completely coated. The amperage should then‘TABLE 1.1—CORROSION PROTECTION FOR SPRINGS. Ctra fo Recommended ‘Dimensional ‘Special Determicing Protection in ‘Buu, Methods of Aavantages, Protection’ Order of Prefer. Minimum/Surtaco, ‘Applying Precautons, and tobe Used ence am Protection Limitations Directy ex Phosphate coat 8 ip. spray ‘Goce tts wail pee! posed to ‘an pair, OF fr brush paint weather, paint alone not making close fis Deeaty ex Cacmum or 3 Electoplate Corrosion protection ‘po8ed 10 "ine pale ‘Qood. Peli o! Py ‘weather, ‘rogen embrite: making ‘ment essertal a. ose fis ter electroplaing® Not girecty Cagmium o 4 Blectroplate Same as above ‘exposed to ‘ane pale weather, ‘making tose tis Not arecty Phosphate coat 8 oO ‘exposed to ‘Dw of weather. Pate coat plus making paint lose tis Feiaively Precoates Thickness is vara- Hot ip or Fair at purpose coat mis corr ‘sping wire ie and is depen ‘lectroplate ing applied to, Sve eondh ent on size spring wire sut tone ‘able lor migly cor > tesive conatons ‘mere coro- Oxide black Lees than § ‘Alxaline ox: ‘sve cond. izing 60: tions are tutor very ml Midest cor Shshing ait or Dependent on Good tor protection rosive con: rust proofing ‘mattiod of app tor to instalation ions compounds Caton and/orvis: fnd for short pe cosily of materia! fed ater instal *See text for alerate methods of plating and recommendation for ree of hydrogen embritlement be raised, as the plating operation is continued ut specified thickness is achieved In spite of these precautions, springs which have been cadmium or zine plated have been exposed to a severe hydrogen embrittlement environment during the preplate and plating steps. To produce a satisfactory spring, con- ditions must be set up to force the hydrogen to pass from the steel spring through the cadmium or zine plate, The time and temperature required to relieve hydrogen embrit- Hlement may vary from Ih at 180°C, to as much as 24 h or longer at 220°C, depending on the tensile strength, hardness, residual stress, size of the spring, and thick- ness of the zinc or cadmium plate. This operation should bbe performed as soon as possible after plating. The following processes for applying metallic coatings may be used with no exposure to hydrogen or substan- tially less exposure than occurs with conventional elec- troplating: mechanical plating, peen plating, vacuum plating, and low-embrittling electroplating. 216 the Effect of on Environment ‘The choice of finish depends also on the conditions encountered in use. Each individual application presents a different problem since the type of protective finish suitable for one may be wholly inadequate for another. For example, consider the question, "Should zinc- or cadmium-coated springs be used in hydraulic systems using petroleum hydraulic fluids?” While the zinc- or cadmium-coated spring will function perfectly for years, in this medium, conditions can develop causing the petroleum hydraulic fluid to break down producing acid, sludge, and varnish by-products. These breakdown products are caused by temperature and time. When a breakdown occurs, it will be accelerated by the presence of zine or cadmium. Zine or cadmium metal will act as a catalyst speeding up the formation of acids, sludges, and varnishes. These conditions may cause little or no etching of the steel spring but will prevent normal oper-ation of the hydraulic system due tothe sludge and var- before final decision, so as to coordinate the finish with nish deposits, the manufacturing procedure I is essential (0 consider all of the factors involved in Table 1.1 can be used as a general guideline for select- order to select the finish best suited for a particular pur- ing finishes. Ke pose. if possible, the spring maker should be consulted 247Chapter 2 Spring Materials 1. Material Selection ‘Among the factors governing the selection of a mate- rial for a spring are the desired load on the spring, the stress range through which the spring operates, mass and space limitations, expected fatigue life of the part, the environment in which the spring will operate with respect to temperature or corrosive conditions, and the severity of deformation encountered in its fabrication. As loads are increased and space is limited, springs must be made {from higher tensile strength materials. If the load cycles, also increase as the stress levels rise, the spring material must have correspondingly better fatigue properties. Music spring wire has the highest tensile strength of all the common engineering materials, and valve spring wire has the best fatigue properties. ‘Short descriptions of the types of wire and strip most generally considered for use in the manufacture of springs. ‘will be found in the following sections. Complete details and tolerances of the materials are covered in the refer- ence specifications. The SAE specifications also give the Fequirements of the finished springs made from these materials. Cold Wound Spring Material Spring wire is produced from hot rolled rods by cold drawing through carbide dies to obtain the required size, surface finish, dimensional accuracy, and mechanical. properties. By varying the chemistry, the amount of cold. reduction and other mill practices including heat treat- ment, a wide diversity of mechanical properties and fin- ishes is made available. ‘The several types of wire used in spring manufacture are made from a variety of chemical compositions and are either annealed, hard drawn, or pretempered to defi- nite mechanical properties. The performance of springs is dependent upon the mechanical properties of the mate- \ rial used or the properties which may be developed as a result of subsequent treatments. Some of these proper- ties are given in Table 2.1. Tensile strength values are shown in Tables 2.2-2.20. and 2. 3. Ferrous Wire Hard Drawn Carbon Spring Wire—SAE J113, UNS ‘KO6501 and Special Quality High Tensile Carbon Spring Wire—SAE 3271 ‘The mechanical properties are developed by cold draw- ing patented hot rolled material. Patenting may be defined as heating to above the critical range followed by rapid cooling to transform at an elevated temperature in the 455-465°C range. This operation produces a tough uniform structure which is suitable for severe cold reduc- tion without actual or incipient breakage. To replace patenting, a new process has been developed wherein the hot rolled rods are given a controlled cooling treatment on the hot bed prior to coiling into bundles. The result- ing structure and physical properties are equivalent to those obtained by patenting. This treatment is limited to a maximum of 13 mm rod diameter. The subsequent cold. drawing imparts a fibrous structure and increases the ten- le strength of the wire without appreciable loss in duc- ity. Hard drawn wire is intended to withstand more severe deformations than oil tempered grades. For the ‘most part, hard drawn wire is intended for use in springs subject to static loads, low stress, or infrequent stress repetitions. Tensile strength values of SAE J113, Class Land Class 2 hard drawn wires are shown in Tables 2.2 SAE 4271 is similar to SAE J113 except the former tricted wire tolerances and is used for appl: cations requiring higher stresses. See Table 2.4 for ty cal tensile strength values. ‘Music Spring Wire—SAE 3178, UNS K08500 Music spring wire represents the highest quality of hard drawn steel spring wire. Its manufacture involves care- ful selection of heats of steel or portions thereof, special hot rolling practice, and very closely controlled heat treat- ‘ment and cold drawing operations. The attainment of the high mechanical properties of this grade is dependent upon the chemistry of the steel, the high degree of crafts- employed in processing, and the amount of cold reduction (greater than that used for SAE J113)after the last of several patenting operations. This greater degree of cold work involves an increased number of drafts, a lower percentage of reduction per draft and slower wire 2.19TABLE 2.1—PROPERTIES OF COMMON SPRING MATERIALS. Speciicaton_ Modus ot Eiaststy | pvaiatie | Maximum Name ws ToT fae ace | wire Ses | Tamperatie> VY umber | cae | Comparable | arm | Tension (€) | Shear (a) | #2 S28 | Tempera asT™M Meas | MPa High Carbon Stee Herd orn xossar | visa | ager] 7850 | 205 000 | 79200 | oso | 110 Hard draw vave xosror J u72} = | 7as0 | 205 000 | 79 300 | 22-65 0 Nard drawn special quaity | 08200 | sar | aso | 205 000 | 79 300 | O5—1 no Music oasoo | usa} zee | 7250 | 205.000 | 79300 | o10-6s | 40 Oi temperes oven | sae | az2a | 7860 | 205 000 | 79300 | oso—16 | 160, Oi tempered vob xoorar | sss1 | aza0 | 7560 | 203 000 | 79 300 | ees 160 annealed sro650 |p raso | 205.000 | 73300 | twws 110 sise80 | = 7e80 | 205 000 | 79 300 | Overs 0 Aly Steels Chromium Vanadium scisoe | — | azar | 7080 | 205 000 | 79 300 | os—12 230 Grromum Vanadium vate | 15047 | size | x23 | 7850 | 205 000 | 79300 | 05—12 230 Chromium Sion sso | i157] aaor | 7850 | 205 000 | 79 300 | 08-11 250 CGhromum Sicon valve = asrr | 7es0 | 205 000 | 79300 | 08-10 250 Staines Sie! ‘Austere (18-6) 20200 | 1220] sis | 7900 | 190.000 | 68900 | oze—10 | 290 177 PH sivreo Juiz} vero | 200.000 | 75600 | o5—i0 540 Copper toys Phosphor Bronze cse100 | sees] 109 | 8800 | 108 000 | 43 400 ‘ : C1000 | 469 | 8159 | e860 | 105 000 | 43 400 | 10-12 : Berylum Copper cr7eq0 | vss | arse | 260 | 190 000 | 48 300 ‘ : civz00 | ssa | 197 | e260 | 130 000 | 28 300 : Silicon Bronze Gossoo | 4469] B99 | 8500 | tos 000 | 43 100 : Brass 26000 | v463] 325 | 8800 | 105 000 | 37 900 : cere0n | 463] Bis¢| 8470 | 105 000 | 37 900 . Nickel Atoys ‘incon 600 ossoo | ue70} —— | e800 | ars 000 | 75.800 | o1o-r2 | 420 Inconel X750 nav7so | us70} = | 300 | 218 000 | 75 800 | d10—1« | S00 Mone! 400, nossoo | sera] = az | 180 000 | 65 500 | o1o—12 | 230 _ Mare 500 wossoo [ro] | e470 | 10.000 | 6s 500 | ovomi2 | 200 correcied sess for 190 > ‘Dela fortis material are ot avaiable, “Sheet and stip. ‘TABLE 2.2—HARD DRAWN CARBON STEEL SPRING WIRE, SAE J119, CLI Tensile Strength Tense Strengine Diameter MPa Diameter MPa. Min | Max Min | Max oso | 1950 | 2250 | 280 | s4a0 | 080 08s | 1920 | 2200 3.00 | 1450 | 1650 060 | 1900 | 2180 350 | 120 | 1620 oes | tea0 | 2160 | 400 | 1400 | 1600 070 | aso | 2130 J aso | 1380 | 1500 080 | 1800 | 200 | 500 | 1250 | 1550 80 | 1780 | 200 | 550 | 1320 | 1500 100 | 1750 | 2020 800 | 1300 | 1480 110 | 1720 | 1980 700 | 1250 | 1420 120 | 1700 | 1950 | 300 | 200 | 1380 140} 1680 | 1920 | 9.00 | 1180 | 1350 reo | 1650 | 1900 J 1000 | 1150 | 1320 180 | 1600 | 150 f 1100 | 1120 | 1900 200 | 1580 | 1800 | 1200 | 1100 | 1280 220 | 1580 | 1720 | 1400 | 1050 | 1220, 250 | +500_| 1700 Pretered sizes, ‘Tensile stength values for intermediate sizes may be interpolated ‘United Numbering System for Mets and lays. SAE HS 1086 (ASTM 0.56) ‘Temperatures represent ne appoxmate point ef which a compression spring wil experience 10% load loss when loaded to 700 MPa drawing speeds in order to control heat buildup, Music wire is used in the more severe types of spring applica tions which require high tensile strength (see Table 2.5) and exacting surface control to obtain the necessary fatigue properties. . a Oil Tempered Carbon Spring Wire—SAE J316, UNS KO7001 ‘The mechanical properties for this grade are developed bby heat treating the wire at its finished size. The general processing to this point is the same as for hard drawn wire. The heat treatment consists of heating the wire in the strand form to a temperature above the critical range, quenching in oil and then tempering by passing it through ‘a molten lead bath maintained at the proper temperature to produce the desired mechanical properties. The result- ing tempered martensitic structure provides greater free ‘dom from relaxation under repeated and continued stress 2.20‘TABLE 2.3—-HARD DRAWN CAR- TABLE 2.5—MUSIC WIRE, SAE J178, UNS KO8500 BON STEEL SPRING WIRE, SAE sna, cuz Tensile Song Frensie Sram ee Diameter ia ee A ‘Tensile Strength mm [}———+ _ ameter ee win_|_ Max Win [Max tm [ Ma o1@ | 3000 | 000 eo | 2450 aie | 2000 | 3200 iso | 2100 oso | zee | 2500 are | 2600 | 100 20 | 2080 eee ora |p ce ors | aro | oso roo | 2350 oes | a0 | 2ea0 ozo | 270 | 00 aso | 2300 om | bo | 240 oz | aes0 | 2500 ooo | 2250 Rege e| eirote | asa ozs | 2650 | 2950 180 | 2200 pane flail |e cos ozs | 2000 | 2320 {280 | 2200 110 2000 | 2250 030 | 2600 | 2900 1900 | 2150 120 190 | 2220 03s | 2550 | 2820 reso | 2100 ta | ts0 | ben te | tao | 2120 ou | 2500 | 2750 1220 | 2050 ts | tas0 | oe0 4s | 2480 | 2700 {00 | 2000 oe 1820 260 0.50 2400, 2650 1750, 1950, 3% | veo | 2000 ! 20 | veo | 2000 oss | aae0 | 2020 700 | 1900 3a | v0 | ta00 060 | 2290 | 2000 130 | 180 3S | eso | aso oss | 2320 | 2580 teso | 1380 is veo | ta00 070 | 2300 | 2550 1620 | 120 iS | isso | 980 450 | 1550 | 1750 820 | zea | 2500 -tso0 | 1800, 5 | iso0 | 100 1500 | 1750 800 vaso | 1080 0 ‘20 | tee “Freer sae. a0 taso | t580 “Tonle svengh values for trmediae sizes may be nerpoaed soo | i350 | 550 1000 | taeo «| tsa0 Moo | tao | 00 to | tas | tas0 "1400 1200 1360, ‘Preterred sizes. application as compared to hard drawn wire. Oil tem- STonstesengh values forte pered wire is also more suitable for precision forming and acer eee coiling operations than hard drawn because of the close control of tensile strength and its superior straight- poe ness. Tensile strength and hardness values of Class 1 and HARD ORAwN STEEL WE, SAE ‘Class 2 oil tempered wires are shown in Tables 2.6 and art 21. ameter | TorsieStoraie ba Typical Oil Tempered Carbon Valve Spring Wire—SAE 3351, => UNS K06701 050 2600 0.80 2500 080 2400 ‘Valve spring wire represents the highest quality of oil $90 7200 tempered wire. It is used for springs subject to dynamic iz in stress applications with maximum life expectancy. To 200 2000 meet this requirement, valve spring wire has the highest 250 1980 degree of uniformity with respect to surface, structural fea aos soundness, and mechanical properties. The production 5.00 1780 of this quality requires rigid inspection, selection, and 20 1700 continuous control of the product from the steel melting aoa be furnace to the finished wire. The surface of this wire will 200 1450 have no areas of detrimental carbon depletion. Its free 00 1400 of seams, scratches, die marks, pits, or other surface rotoned sizes defects that would impair the fatigue life of the spring. \\ ene stength values for in Wire tensile strength and hardness values are shown in termeciate sizes may be ier Table 2.8. palates. 224‘TABLE 2.6—Oll. TEMPERED CARBON STEEL ‘SPRING WIRE, SAE 4915, CLI ‘TABLE 2.7—OlL_ TEMPERED CARBON STEEL SPRING WIRE, SAE J316, CL2 Tensie Sven Tele engi ‘Diametert MPa Hardness Diameters MPa . Hardness ae iain tnt Jee Roca 15 Rockwa 15 oso | zoo | ze | eo" soo os | zo | 2a | soo So os | dee | 20 | ao | Soo oss | dem | demo | S00 | seo sc | do | 200 | 0 | Soe deo | doo | oo | soo | sro ces | taco | Zim | sso | soo des | fim | 200 | 0 | St om | te | iso | oe | os oro | feo | Zam | 30 | sto oa | dso | seo | aes | sos geo | taco | ao | ars | sos too | fio | Zao | so | S00 oo | tes | fom | aro | fo 100, we00 | 2000 | 265 | eas Hine a liecrne || een l| area || tos io | S000 | Zao | ars | aos ti | iso | ie | S60 | a0 120 iso | 1950 | s60 | a0 120 jeanne | reales er De | eel Ho | jos | die | So | a0 Poche 45 ‘ : Poche 41 4 | 199 | sooo | seo "y goo seo | 1900 | 2100 | soo" eto a | eso | tes | So | foo 180 1620 | 1820 | 530 | seo 380 ee aoe eset cra zoo | ts | ‘oso | ae | fro 200 | re | 0 | ses | srs oe | taco | foo | ss | cos im | two | imo | $0 | dro 35 | to | ‘iso | sso | ao $8 | tao | imo | 83 | Se 20 | to | an | $0 | oo 2s | tsa | tr | sos | ass 300 | too | two | 3 | ss Rockwell C Rockwell C goo | oo | zoo | crs aso | sso | eso | arse 38 | a | wo] | 8 to | tem | iso | | 3 to | tim | i] | 8 a3 | iso | io | a | 8 s3 | tw | iso | i | soo | te | woo | a | so |i | bao | oo | o ss | tw | ion | a | 8s sso | tsa | soo | a | ow fo | ies | eo | & | 8 5 | iso | ro | |g aco | io | eo | | a a ee eae 9.00 vaz0 | 1620 | a4 | a9 ea oe 10.00 1aoo | 1600 | 43 | a8 so | tam | two | a | a to | too | tw | 3 | roo | 200 | reo | ar | as vm | toso | ts | a2 | Noo | to | i390] a | tio | 0 | eo] & | woe | tio | ie | s&s | ison | tao | em | on | te tio | ite | ta | a | tt veo | tee | 0 | a ff Protaea cee “Teeter vas rite ze maybe nar ‘reer sizes. "Tense svengih values lor intermediate sizes may be nter. polaied Hard Drawn Valve Spring K06701 A hard drawn grade of valve spring wire is produced with all the restrictions as to processing and controls of uality equal to the oil tempered grade. The final mechan- al properties are the result of the amount of cold work which follows the final patenting treatment. The fabri cation of this material is similar to the hard drawn car- bon spring wire described in the first paragraph of Section 3. See Table 2.9 for tensile strength values of the wire. Annealed Wires (SAE 1065, UNS G10680 Thru 5.0 mm; SAE 1566, UNS G15660 Over 5.0 mm) Annealed wire is produced for certain applications 222 polated where there is a high degree of severity of deformation in the fabrication of springs and formed parts. This wire is not commonly produced to definite tensile properties, as the parts formed are subsequently heat treated. This grade may be furnished lightly drawn after annealing in order to aid in dimensional control during forming. Allloy Stee! Spring Wires—SAE 3132, UNS'K15047; SAE, 3157, UNS K15590 Alloy steel wire is used in the manufacture of springs which operate at moderately elevated temperatures up to 230°C: Because of their resistance to relaxation under these conditions, their high tensile strengths, and their torsional elastic limits, the most commonly used alloys are SAE J132, Oil Tempered Chromium Vanadium Valve | |TABLE 28-OlL TEMPERED STEEL VALVE SPAING TABLE 210 CHROMIUM VANADIUM STEEL SPRING WIRE, SAE \ VaRE: SAE Jos UNS ROS7OY Fis, UNS wasoer \ Dareter | Tensia Svongm Tense Svengh Reason \ Pa eS Harness 3 er | ove [tw | vin [wae | mn | Moe vin | wx | wa [| we | smn Rockwell 45N ‘Rockwell 15N | Bees | eo | eo | caps — | os | 2050 | 2200 | ss")'sr | — ie | 38 | tem | te y |S os | oe | dom | 200 | aes | os | = a8 | os | S00 | ats | seo | 993 | ocho © Gs | $a | isso | Zico | ers | see | = h as | ao | 1590 | a0 | ees: 28 | 12 | ts00 | dom | ero | oss | = 1 38 | te | tee | to | ss | & I fo | SS | ise | wwe | fe | fs Pockwal sh | $3 | °2 | too | ves | af | 12 | 18 | 00 | isso | sro"y"ere | - | 12 | 28 | to | tse | see | see] = i 28] 25 | too | tao | aso | Sra | | Rockwell C | as | 20 | 10 | ro | wt | as TABLE 20-HARD DRAWN CARBON Bret Veuve “Senne wine, sxe 39 | 49 | iso | mo | 7 | so | Ste Uns Yooros $3] 80 | we | too | ss | | a 82] 85 | too | iso | Ss fe |e aaa Terao Svongth So | 100 | tos0 | soo | a2 | a6 Pr 88 | 0 | tae | to | | we | & oa Poe tom |e eo | 22 | seo | to | | te | — [as | so | 50 ss | 83 | im | te ny $a feet aoal ‘TABLE 2.11—CHROMIUM SILICON STEEL SPRING WIRE, SAE J157, | 55 1500 1650, aes Tersiesvera™ | wagon | Reston ; a Diameters MPa ter of Area | we [wa _[_ mm [ wo | Min Spring Quality Wire and Springs, and SAE J187, Oil a ea Tempered Chromium Silicon Alloy Steel Wire and 09 2050 2230 885 90.0 - 1 Springs. See Tables 2.10 and 2.11 for tensile strength and oo | Rw | gem [oss | oo | > hardness values of oil tempered wire. These alloys may ty | gow | gzig | aes | soo | } also be obtained hard drawn or annealed, and in com- ta | ao | so | sto | es | Inercal quality or valve spring quality. — . . ts | ao | 20 | ee" | — | 4. Stainless Steel Wire—SAE J230 and SAE J217 18 1980, 2160, 59.0 625 - 38 | ise | Zuo | so | eo | Stainless steels, SAE J230 in either UNS $30200 or 22 1940 2120 $85 61S - ' 320 | igo | two | So | os | is UNS $30400 grades, are produced to mechanical proper- 28 1910 3090 58.0 610 “6 i ties by cold drawing. In tis condition, these grades have 33 | tooo | domo | ors | io | a f ‘00d corrosion resistance and also good temperature eae resistance up to 260°C. SAE J217, UNS S17700, Spring we evils beast om Wire and Springs, covers a grade of stainless normally 4.0 1830 Ed cS) 4 535 al roduced cold drawn which, after fort ge hard- bot aa 7 2 330 i produced cold drawn which, after forming, i age hard se | teoo | igs | sos | sao | ened by heating to 480 + 5° for 1h. In this condition, so | mo | wo | eet | oe | it has mechanical properties equal to music wire. It has ey ic 1000 rH ee ae the cor \-resistant properti _ a 1886 s the corrosion sant properties of UNS $2020 an zo | 70 | woo | ws | eo |e j ition, preter resistance to relaxation up to about go | wo | we | ee ef 340°C. All metallic coatings, such as lead or copper, must 10.0 1640 1800 485 S15 40, t be removed prior to heat treatment. Springs made from 10 120 1780 a sa a } \Coalless wire must be passivated after heat treating to preterred sizes. } Sure maximum corrosion ressanee.Allof these stan- Tene srengh valves for itm ze may berate. ! 2.28VY TABLE 2.12 HARD DRAWN STAINLESS STEEL WIRE, SAE J200, TABLE 2.14—SPRING BRASS WIRE, SAE UNS 30200 (18:8) 1461, UNS 626000 AND UNS 627000 Tensile Stength® Frensile Strangtne TensitSength, MPa Diametere MPa Diameter MPa. Temper [ —— mm mm Mia Max Min [Max min [Max Hatthara 550 650 022 zea | 2440 160 | 1760 | 1950 Three Quarter Hard 625 7s 0.25 2020 | 2420 180 | 1720 | 3920 Hara 700 800, 028 2200 | 2400 200 | +680 | ta90 Bera Hara 800 300 030 ziao | 2380 220 eso | 1950 Spring 825 SI 035 2iso | 2350 250 | 1620 | 1820 040 2120 | 2320, 290 | i580 | 1700 045 2100 | 2300 300 | 1850 | 1760 050 2080 | 2280 350 | 1430 | i680 0s 2050 | 2250 400 | 420 | 1620 060 2020 | 2220 450 | 1360 | 850 070 1980 | 2200 00 | 1300 | 1500 0.80 1940 | 2140 550 | 1260 | 1460 0.30 1900 | 2100 eco | 1220 | 1420 1.00 1880 | 2080 700 | 1160 | 1360 110 wes0 | 2050 800 | 1100 | 1300 120 1820 | 2020 900 | 1000 | 1200 140 1800 2000 10,00 900 | 100 “Profored sizes ‘Tensile svength values for intermediate sizes may be interpolated, TABLE 2.13—HEAT. TREATED STAINLESS STEEL WIRE, SAE J217, UNS $17700 (17-7 PH) ameter | Tesi Seong oa Typical 050 2400 060 2380 0.80 2300 190 2300 120 2250 160 2200 200 2100 250 2050 3.00 2000 400 1950 5.00 1900, 6.00 1850 8.00 1800 10.00 1750 ‘Prelered sizes, 224 less steel spring materials are magnetic at spring hard- ness. Tensile strength and hardness values for hard drawn wire are shown in Tables 2.12 and 2.13. 5. Copper Alloys Spring Brass—SAE J461, UNS €26000 (10% Copper—30% Zine) ‘This material is most often supplied as strip or and is recommended for light-duty springs. Mechanical Properties are not usually related to size but rather to the terminology quarter-hard, half-hard, three-quarter-hard hard, extra hard, spring, and extra spring, which indi cate the mechanical properties available irrespective of ‘material section, as shown in Tables 2.14 and 2.15. Phosphor Bronze—SAE 3461, UNS C5100 and UNS ©5210 (80% Copper—8% Tin—0.3% Max Phosphorus) Sheet and Strip—This material is superior to spring brassin mechanical properties and equal in electrical con- ductivity. Mechanical properties are shown in Table 2.16. ‘Wire—Spring temper wire has the size-tensile relation- ship shown in Table 2.17, 7TABLE 2.15 — SPRING BRASS SHEET AND STRIP, SAE J461, UNS C26000 sevceson | fea, [__ tae Tenor | HE aan [me [ on [ a alae age aia Hard 030/070 | 490 | $60 | _ — | 7m] 73 ospow area rae ae Teele 3e0 ={= [8 [ 4 asm =le|-1~ a ase ‘Spring ‘oso/o.70 | 25 | 690 | — = 7 | 78 TABLE 2.16—PHOSPHOR BRONZE SHEET AND STRIP, SAE J461, UNS C52100 Tecate ae sie secion | Sere ‘Temper ral ‘MPa Min | Max | Min | Max Omwities [win [erp ne [peor op sla lola worn | 92-0 fas | ool a | a | eT basras Sle lala vor. atew|-|- oer. “1 ela ie osor.o | $85 | 60 | a | os | — | — 8B8f0% 21% |i | ow voy. =| m]-]= ews | 82, loolmlala| @] ® ts/0% S|"lale vo wv) w]-|- i Tyee | some | 828-5 | rs | 20] ag | ico | @ | & 58/09 Se] lis | a vay ce pm} —|— 9 2) | | a cx song | 225/;, | 700 | 000 | oe | or | @ | 158/09 e | 225TABLE 2.17—PHOSPHOR BRONZE WIRE AND ROD, SAE W461, UNS C51000 Diameter Minimum Tensile ae mm ‘Strength, MPa el over [| tu | wee | Ros | %in80mm = 60 | 1000 | 850 : 060 150 925 | 800 = 150 3.00 200 | 750 = 3.00 600 aso | 725 | a8 cosonyy 500 3.09 325 | 700 50 300 | 1200 725 | 625 90 Beryllium Copper—SAE 3461, UNS C17200 (98% Copper—2% Beryllium) ‘This material is commonly used because of its excel- lent fatigue resistance and its ability to be formed in a soft or cold worked condition and subsequently aged or precipitation hardened to spring temper at the relatively, low temperature of 315°C. Mechanical properties of the various mill tempers after heat treat, independent of size, are shown in Table 2.18, Silicon Rronze—ASTM B99, UNS 65500 (95% Copper—3% Silicon) There are several different chemical grades of silicon copper alloy wires. UNS C65500 strip is also used for flat parts requiring spring properties. They are less costly than, the phosphor-bronze group and have about the same tensile-cold worked temper relationships. The tensile strength of spring grade wire is 900 MPa minimum and 725 MPa minimum for strip and is a function of the chemistry of the alloy rather than size. 6. Nickel Alloy Wires, SAE J470 UNS N06600 (Inconel 600) (15% Nickel—16% Chromium—8% Iron) This is used chiefly because of its corrosion resistance and its strength and resistance to oxidation at elevated temperatures. Its nominal operating range is from a low of -40°C to a high of 345°C when springs are designed with average stresses, such as 380-585 MPa in torsion or 760-930 MPa in bending. As cold drawn, the tensile strength range is 1070-1280 MPa. The nominal stress relieving temperature is 425-480% UNS NO7750 (Inconel X-750) (73% Nickel—13% Chromium—1% Iron—2.5% Titanium—0.8% Columbium—0.8% Aluminum) This is an age hardenable alloy that has outstanding. resistance to relaxation at elevated temperatures up to 600°C. This grade is furnished in two tempers. The ower temperature aged product has a tensile strength range of 930-1140 MPa which after aging at 730°C for 16h will have a tensile strength of 1170 MPa and mé mum resistance to relaxation. If greater strength is desired, an aging treatment of 845°C for 4 h will develop 1520 MPa tensile strength but with some deterioration of relaxation properties. UNS NO4490 (Monel 400) (65% Nickel—35% Copper) This is used because of its high strength, ductility, and resistance to alkalis and acids. It is also the lowest cost nickel grade spring alloy. In the cold drawn spring grade, its tensile strength varies with the amount of cold work in the range of 900-1100 MPa, and it should be stress relieved after coiling at 275-315°C, It is best used for TABLE 2.18—BERYLLIUM COPPER, MECHANICAL PROPERTIES OF MILL TEMPERS AFTER HEAT TREAT, SAE Ja61, UNS C17200, Tonsie Ste —— eee stip Wire Hardess, Min Botore Hardening | “Temper {a wre | Min | Max | Min | mex | Ac| Rao | atsn Soliton Treated (A) at | 1140 | 1350 [ 1140 [1950] a6 | 58 | 70 Quarertard (%) | wT | 1200 | isto | 1200 | ta10 | 38] 58 | 79 HaltHard (2H) veHT | 1270 | 1480 | 1270 | 1480 | 39] 89 | 795 Tiree Quarter Hard Ge) war | — | _ | sai | 1520] 40] 60 | 80 Hard (H) wr_| aio | 1520 | — | — | ao} eo | 0 2.26‘moderately loaded springs which are subject to corrosive attack such as marine or salt water conditions. It is reasonably stable in the temperature range of -70 to + 200°C. UNS 05500 (Monel K-500) (6% Nickel—31% Copper—3% Aluminum) This is used in place of regular monel because it can bbe age hardened to provide higher mechanical properties. ‘When aged at 595°C for 16 h followed by slow cooling. to 480°C, this grade of wire has a tensile range of 1100-1240 MPa. It is nonmagnetic and can be used at temperatures to 235°C. Superalloys In addition to the above, several specialized grades of nickel or cobalt-base alloys containing varying amounts of chromium-tungsten-molybdenum and/or silicon have been developed. These materials, termed “‘superalloys,”” are all age hardening grades and have tensile properties in wire form that can be used to make springs that are dimensionally stable up to 815°C and at stress levels to 350 MPa. Details of these materials are given in SAE 3467, Special Purpose Alloys (“Superalloys”). Further information should be obtained from the manufacturers. 7. Spring Steel for Flat Spiral Springs Flat wire and strip are terms, sometimes used inter- changeably, which identify two different products. There is an area of overlap where a cross section may be Produced either as a flat wire or as a strip. The process- ing is different, and since this may have an effect upon the application, the two shall be described briefly below. Flat Wire Flat wire is produced by passing round wire through the rolls ofa flatening mil. The wire may be left with 4 natural round edge formed by the flow of material as it passes between the rolls, or it may be shaped by a Turks Head or edger to produce a flat or radius edge. In the hardened and tempered condition, the desired ‘mechanical properties fora given end use are obtained by quenching and tempering inthe strand. The required combination of stength, hardness, toughness, and duc- tility can be produced by varying the carbon content and the conditions of quenching and tempering. In ordering flat wire, it is necessary to specify the edge required, unless the approximately round edge produced by cold work is acceptable. 207 Strip ‘The ability of high carbon steel strip to attain high hhardnesses and elastic limits makes it suitable for many special applications. Spring steel strip is produced in a variety of chemistry combinations and mechanical properties. Selection of the proper grade can only be made after determination is made of the hardness, strength, and fatigue requirements, plus the ductility needed for form- ing. Spring steel strip is producee by subjecting hot rolled strip to a cleaning operation and a combination of cold rolling and thermal treatments. This process transforms coarse grained material of uneven dimensions and a rough surface to one that is capable of being formed, responding to heat treatment, and having uniform dimen- sions and a smooth surface free of defects. Selection of the raw material and the mill practice employed are key factors in determining the quality of the finished product, such as uniformity in response to forming and heat treatment, and ductility. Hardened and tempered cold rolled steel strip is produced to meet a range of hardnesses or ultimate ten- sile strengths. This operation is performed om the stri to secure the highest mechanical properties. Material in this conditior. is capable of only moderate forming. Hardened and tempered strip may be supplied in any of the following finishes: black tempered, scale-less tem- pered, bright tempered, tempered and polished, or tem- ered, polished, and colored. Untempered stee! strip may be supplied with different edges as listed below: 1, Round or square, a prepared edge. 2. Natural mill edge. 3. Approximately square edge produced by slitting. 4. Approximately round edge produced by edge rolling. 7 . 5. Approximately square edge produced by edge roll- ing or filing to eliminate slitting burr. 6. Square edge produced by edge rolling when width tolerances and finish are not as exacting as in Item 1. ‘Tempered strip is supplied with an edge as described in Item 1. 8. Hot Coiled Spring Materials Spring steel bars are available in carbon and alloy ‘grades. The bars are generally used in the as-rolled con- dition, but may be centerless ground before coiling. Table 2.19 lists some materials commonly used in hot coiled springs. Given a steel heat of minimum hardenability and a mild oil quench, the "H” steel sizes listed should‘TABLE 2.19-MATERIALS COMMONLY USED IN HOT COILED SPRINGS Tensile Properties Torsional Properties unmats [elastic timate | Elaste ApolcatonSaned Material Strength | Limit Moduius | Strength | Lim Modulus pa emanates MPa MPa & MPa. MPa o (Seg Chapter 2, = aoe MPa +4 MPa ‘Section &) Mae co Mine Mine Carbon stees| 'SAE 1088 and SAE 1095 1200 300 200000 780 520 76000 | upto 10 mm bar size Carbon Boron Stee! SAE 15B62H 1450 1280 200000 1000 70 76000 | upto 25 mm bar size Ally Steels SAE 51504 1450 1280 200 000 1000 720 76000 | up to 10 mm bar size ‘SAE 5160H 1450 1280 200000 1000 720 76000 | upto 20 mm bar size SAE 92604 1450 280 200 000 +000 720 76000 | upto 10mm ber size SAE 51850H 1450 1280, £200,000 1000 20 76000 | up to 20 mm bar size SAE 4161 1650 1280 200.000 1000 720 76000 | upto 60 mm bar size SAE 6150H 1450 1280 200000 1000 720 76000 | upto 101mm bar size ‘Generally accepted valves: reduced from nominal values of 205 000 MPa and 79.300 MPa to offset various factors relaled ty hol Goleg springs Carbon ste! minimum values are based on size shown. Gore hardness may be lower than surace hardness for size shown, re lel minimum values are based on a though hardened bar of 444 HB (1450-1650 MPa Uimate Tensie), Ret” SAE J4198, Typical 8, hardness ranges are 444-495 HB and 461-5 provide a minimum as-quenched core hardness of 58 Re. Sizes for the “H" steels are based on SAE J1268 JUN80. A minimum core hardness of 58 Re as-quenched is con. sidered desirable to produce a spring that can be fully preset to operate at maximum stress levels. An automo- tive suspension spring is typical of the high stress, mi ‘mum weight approach, ‘When springs are operated at lower stresses, a softer core can be permitted and larger bar sizes used. The car- bon steel sizes shown are based on this type of appl tion. Selection of a material involves many variables and Fequires a thorough knowledge of spring requirements, materials, and the heat treat system involved. 9. Corrosion Resistance To ensure that a new spring design will provide ade- uate performance, it is necessary to evaluate not only the mechanical requirements of the application, but also the environmental conditions in which the spring must function. Corrosion is the major environmental factor which will promote spring breakage. It destroys and removes metal from the surface of the spring by chemi- cal or electrochemical methods in an irregular fashion which causes an overall reduction in the spring stock size, compounded by localized areas of intensive metal removal. These effects seriously reduce both the static strength and the fatigue strength of the spring. Since the 228 corrosive losses reduce the spring stock size, the result- ing strength reduction will vary inversely to the section size and, therefore, gain increasing importance as the size decreases. To protect the spring against the corrosive environ- ment, one must either isolate the surface from the environment, nullify the environmental effects, or use a spring material that is resistant to the chemical attack of the corrosive agents. Isolation of the surface may be ied by removing the spring from the environ- ‘ment or covering the surface with a protective coating. Applied finishes are discussed in Chapter 1, Section 8. The environmental effects can be nullified on elec trochemical forms of corrosion by attaching to the spring sacrificial metal which is chemically more active than the spring metal. This can be done by plating the spring surface with a sacrificial metal, that is, cadmium or zine for steel springs, or by providing an electrical conduc- tive link between the spring and the sacrificial metal. ‘The most common corrosion resistant spring materials, are nonferrous metals or stainless steels. To be truly cor- rosion resistant, these metals must be selected on the basis of the corrosive environment in which they are to oper- ate, Among the nonferrous materials used are phosphor- bronzes, brasses, and copper nickel alloys. However, due to lower elastic properties, itis necessary to use a larger volume of these metals to give performance comparable to ferrous metals. For certain severely corrosive condi-tions, stainless steels offer very satisfactory performance ‘The materials have very nearly the same mechanical erties as the standard ferrous spring materials and ‘IT permit satisfactory design with the least amount of metal and a minimum of space. Because of economic conditions, i i not always pos- ble to-use ant-corrosive materials in spring parts sub- jet to such attack, It thus becomes necessary to analyze each application to determine the type and degree of pro- tection that is necessary to ensure adequate performance. 10, Material Handbooks Additional information on materials and further details of specifications can be obtained from the following sources: VY 1. Society of Automotive Engineers, Inc., 400 Com- ‘monwealth Drive, Warrendale, PA 15096. SAE Hand- book includes material compositions and properties. 2. American Society for Testing and Materials, 1916 Race Street, Philadelphia, PA 19103. Annual Book of Standards: Part 3 covers steel strip and wire, Part 6 covers copper alloys, Part 10 covers tests of metals. 3. Huntington Alloys Products Division, The Interna tional Nickel Co., Huntington, WV 25720. Handbooks of high-nickel alloys. 4. Copper Development Association, Inc., 405 Lex- ington Avenue, New York, NY 10017. Standards Handbook for copper alloys. 229Chapter 3 Helical compression and extension springs are normally cold wound when the material is less than 10 mm in diameter, and hot coiled when it is greater than 16 mm (Gee Chapter 4). When the cross section is between 10. mm and 16 mm both processes are common. A. HELICAL COMPRESSION SPRINGS 1. General ‘A helical compression spring is an open helical spring which offers resistance to a compressive force. Itis made in various forms and from different shapes of wire ‘depending upon a wide variety of uses. While itis neces- sary in some cases to employ square, rectangular, of spe- cial shaped wire, the use of round wire predominates in the manufacture of most compression springs. The most common form has the same diameter for its entre length dis known as ¢ straight spring. Variable diameter \worings are used to some extent, The main objective of a spring design is to obtain a spring which will be reasonably economical for a given application, will it into the available space, and will give satisfactory performance. The space governs the dimen- ional limits of operating length and outside and inside iameters. In addition, some other requirements, such as load, deflection, and maximum solid length, need to be known at the start. These dimensions, together with the load and deflection requirements, determine the stresses in the spring. ; ‘When a helical compression spring is loaded, the coiled wire is stressed in torsion, The stress is greatest atthe sur face of the wire; as the spring is deflected, the load varies, causing a range of operating stress. Stress and stress range govern the life ofthe spring. The higher the stress range, the lower the maximum stress must be to obtain com- parable life. Relatively high stresses may be used when the stress range is low or ifthe spring is subjected to static loads only. The stress at solid length must be sufficiently high to permit presettng if this manufacturing operation is included. However, ifthe spring is not preset, then the solid stress should be kept low enough to avoid perma- nent set, since springs are frequently compressed solid uring installation. ~ 231 Cold Wound Helical and Spiral Springs ‘The material selection depends upon the spring appli cation and such factors as cost, allowable stress, fatigue durability, corrosion resistance, and resistance to set or Toad loss at elevated temperatures. Ifthe space is limited, or if optimum fatigue durability is required, shot peen- ing and spring presetting may be considered as means of increasing the allowable working stress. Ifa spring is sub- ject to elevated temperatures and/or high level stresses, then presetting by either heat setting or hot pressing may be necessary to reduce the relaxation in service. See Chap- ter 1, Sections 4 and 5. Spring materials are fully described in Chapter 2, which should be referred to for material selection. 2. Spring Details Terminology and Types of Ends A typical helical compression spring is illustrated in Fig. 3.1 to explain the spring terminology. This figure shows the configuration of closed and ground ends of a com- ession spring. This minimizes the eccentricity of load- ig at the ends and reduces the installation space required. Typical types of end configurations which may be employed are illustrated in Fig. 3.2. The unground ends ‘may be used for reasons of economy, but they give eccen- tric loading with some increase in maximum spring wire stress and space required. The plain ends similarly pro- ‘duce eccentric loading and additionally present a handling problem due to springs tangling together. ‘Wire Diameter (d) ‘The desired size of wire will be deter tic limit and the modulus of the material, loads, range of operation, and type of application of the spring. Round wire provides the most economical compression spring design and should be used whenever possible. At times, however, it is necessary to resort to the use of square, rectangular, or flat wire in order to obtain the desired characteristics within a limited space. If the design necessitates a rectangular section, a rolled wire (with round edges) is more economical and can quite often be rolled from round wire in stock. For these wire sections, rectangular wire formulae are applicable, although the load capacity, rates, and stresses will be somewhat altered due to the large edge radii. ined by the elas-FREE LENGTH NS C-PLAIN FNDS GROUND Fig. 3.2—Typical ends of helical compression springs Coil Diameter (D) The following procedure is recommended when specifying the coil diameter in the itemized spring specifications. 1. Ifthe spring works over a rod, the minimum inside diameter of the spring should be specified. 2. If the spring works inside a housing, the maximum outside diameter of the spring should be specified. 3. If the spring is confined both internally and exter- nally, the limiting dimensions may be stated in addition, to the mean coil diameter. ‘Some consideration should be given to the spring index (C = ratio of mean coil diameter to wire diameter) in the design. An unusually low index is difficult to ‘manufacture, while a high index will cause large varia- in coil diameter. Indexes 3.5-15 are commercially \éactical to manufacture, but indexes in the range of $.5-9 are preferred, particularly for close tolerance springs and, ©) 232 10 TOTAL COILS SHOWN ‘GROUND SECTION D-PLAIN ENDS those subjected to cyclic loading. For details on coil diameter tolerances, refer to Part C of this chapter. The spring index is generally not selected at will as it depends upon other spring parameters. Considering only the active portions of a compression spring, there is a direct relationship between the ratio of maximum deflec- tion to solid length and the stress and index. This is expressed by the formula: Free Length (L,) The free length is an overall dimension measured par- allel to the axis of the spring when it is in the free or unloaded state. If definite loads are specified for the ‘manufacture of the spring, the free length should be an approximate dimension which may be varied to meet the load requirements. On the other hand, if a load with alerance is not specified, then it is advisable to specify 2 free length with a tolerance. Detailed information of ve length tolerances is given in Part C of this chapter. amber of Coils (N) ‘The number of coils should be specitied as a reference sure, and it should be stated whether it refers to active |.) or total coils (N,). The maximum number of coils Il be restricted when there is a solid length specifica im, When specifying the number of coils, it is advisa- to.use total coils rather than active coils. Total number coils is counted from tip to tip. On springs of great agth, it is advisable to specify the pitch or number of ils per 25 mm rather than the total number of coils. orings with closed ends or with closed and ground ends \ve one inactive coil at each end. Springs with plain ends € considered to have virtually no inactive end coils less they are fitted into specially shaped spring seats. otings with plain ends ground are considered to have out one-half inactive coil at each end; the actual num- xr depends upon the free pitch of the spring and the ickness at the tip after grind. Tolerances to number of ils are outlined in Part C of this chapter. lix of Coil In applications such as one spring operating inside vother, it is necessary to coil the springs so that the ices are in opposite directions, right and left. Ifa spring. Derates over a screw thread, the direction of the helix ould be opposite to that of the thread. In some stances, the direction of the helix may be chosen for \t number or source identification. In such special ses, the direction of coil helix should be specified; other- ise, it will be taken as optional. olid’ Length (L,) Itis desirable that the solid length be specified as a max- tum dimension, allowing the manufacturer any toler- tee under this dimension determined by variation in wire ameter and the amount of grind on the ends. oad (P) It s the force required to compress the spring to a speci- 2d loaded length. It may be used as a resisting force to 2sorb the energy of a blow or shock, or to exert a force in the case of a valve gear. In some cases, load is static id is used to exert force at a given point. Sometimes spring load is utilized for an energy source to actuate ts of a device rather than to offer resistance. ‘The axis of a helical compression spring should be posi- oned vertically when the spring is being checked for 233 load. The readings should be made during the unloading. which will produce a lower load due to the spring hysteresis, Load should be specified at a definite compressed length of the spring and not at some deflected distance from the free length, as the free length should be a vari- able dimension. Toierances should be applied to the load and not to the length whenever possible. Load tolerances are explained in Part C of this chapter. Rate (R) It is the change of load per unit length of spring deflec- tion as the spring is being compressed. Rate should be determined between 20 and 60% of total deflection when, test lengths are not otherwise established. If even coil spacing or pitch is used, a constant rate develops. In a compression spring with closed ends, the rate is not con- stant throughout the entire deflection of the spring. This is due to a portion of the active coil adjacent to the closed ‘ends closing down as the deflection nears the solid com- pressed length of the spring, thus increasing the actual rate of the spring due to the decrease in the number of active coils. It is possible to design a spring with a definitely increas- ing rate by using a variable pitch of coils or a variable diameter spring; however, a helical compression spring, ‘cannot be designed with a decreasing rate. Rate tolerances are outlined in Part C of this chapter. B. HELICAL EXTENSION SPRINGS 1, General ‘An extension spring is a particular type of helical spring, usually close wound, which offers resistance to a tension force applied through suitable end forms. Sim- ple hooks and loops serve the purpose most frequently. AA distinguishing characteristic of the conventional exten- sion spring is the presence of initial tension, a force in the close wound coils which must be overcome by a pull applied at the ends of the spring before the coils will separate. Music wire, hard drawn spring wire of various kinds, and pretempered carbon and alloy wire are the materials, ‘used in extension springs. Annealed or untempered wire which requires hardening and tempering after coiling is, rarely used for conventional extension springs, because during the hardening process the springs lose their initial tension.or HooK wine PENING DIAMETER, V -LENGTH OVER COILS) — INSIDE OUTSIDE DIAMETER DIAMETER TYPICAL ENDS USED ON EXTENSION SPRINGS © tm MACHINE CUT PLAIN ENDS. OK NN MACHINE HALF HOOK OVER CENTER @ Cll HALF LOOP OVER CENTER © Cl LONG ROUND END HOOK.OVER CENTER © =I SINGLE FULL LOOP OVER CENTER 8 =lll DOUBLE FULL LOOP ‘OVER CENTER © Cll FULL LOOP ATSIDE © ell SMALL EYE ATSIDE © fli SMALL OFFSET HOOK AT SIDE © —~<(lll CONED END W:TH ‘SWIVEL BOLT © all CONED END WITH SWIVEL HOOK Ocal THREADED PLUG TO FIT PLAIN END SPRING Oct OH Oo Vv-HOoK OVER CENTER Fig. 83Extension springs EXTENDED EYE OVER CENTER 234 DRAWBAR SPRINGTWiST LooP cross LooP Fig. 8.4—~Typical machine made loops 2. Types of Ends The sketches in Fig. 3.3 show some of the many types of ends which have been used on extension springs, Details such as hook opening or restreint of the loop within the body diameter should be specified on the draw- ing because the word descriptions in Fig. 3.3 are not universally standard. There is no limit to the designs of oops which may be developed for a particular installa- tion. In general, the machine hooks and machine loops are the preferred designs of those shown in Fig. 3.3. Many extension springs of small wire size can be looped automatically on a coiling machine. Fig. 3.4 illustrates typical loops which can be made complete on modern spring making machinery. ‘The cross loops made by this technique are free from the mutilated portions which reduce the endurance of this style of loop as made by other techniques. Stress Concentrations in End Coils ‘When sharp bends are made in forming the end hooks (or loops of extension springs, stress concentrations are produced which may cause failures, An adequate estimate f the stress concentration at the bend in the end hook ‘or loop may be found by multiplying the bending or shear stress obtained with the ordinary design formulas by the ratio re/t,, where ty is the radius of the center line of the bend and r, is the inside radius of curvature of the wire at the bend, as illustrated in Fig. 3.5. For best results, this ratio should not be permitted to exceed 1.25; in other words, the Uwice the diameter of the wire. Coned ends with swivel eyes, hooks or bolts, as illus- trated in Fig. 3.3, provide effective means for reducing the stress in the end coils of the spring. Another way to void stress concentration in the end coils is to obtain EXTENDED ‘cROss LooP 295 ENLARGED ‘TWiST LooP REoUcED ‘Twist LooP TORSION sTRESS ATA Fig, 8 5—Stress concentration at end hook bend aan extension spring action by combining a compression spring with two yoke-like drawbars, The limit imposed. upon deflection by the definite solid length of the com- pression spring prevents excessive extension and provides 4 substantial measure of safety for the spring in this arrangement. Extension springs may be made by screw- 8 threaded plugs into the ends of plain machine- cut springs, the threaded plug being equipped with a hook of substantial proportions to minimize the effects of stress concentration 4. Position of Hooks The position of hooks relative to each other can be in line, at right angles, or at any other angular position as required. The hooks on springs with many coils may change position somewhat in shipment, and the variation of position is also affected by other variations during ‘manufacture. Hence, in order to maintain the positionof hooks relative to each other, itis necessary to exercise special care in manufacturing, If the position of ends rela- tive to each other is important, the drawing of the spring should emphasize the importance of position by a state- ment as well as by pictorial representation. 5. Specifications The drawing specifications for an extension spring should include the coil diameter, the wire diameter, the free length, some reference to number of coils, and full particulars concerning the type of ends. Coil diameter is usually specified as outside diameter. Free length should ‘be measured inside hooks. Either the number of coils in the body of the spring or the length over the coils may be specified, but only as an approximate figure. In com- puting the length over coils, it should be recognized that there is always one more wire diameter in the length than the number of coils in a close wound spring. Full hooks and loops are the same diameter as the body coils, and the length required for two ends is approximately the same as two diameters of the body coils. When load is specified, the extended length inside hooks under the load should be given. The manufactur- ing tolerance should be applied to the load rather than to the extended length. Initial tension should be controlled by including it in the test load rather than specifying it independently. If load and tolerance are specified at a particular extended length, the free length inside hooks should be shown as an approximate dimension. Extension springs normally do not have a definite stop to their deflection such as the solid compressed length ‘of a compression spring. The drawing specifications, therefore, should include a statement of the maximum, ‘extended length which must be attained without encoun- tering permanent set, 6. Initial Tension ‘The initial tension that can be wound into an exten- sion spring depends upon the index and the winding tech- nique. Fig. 3.6 shows a range of practical values of initial tension in terms of stress. The corresponding load should be figured from the load-stress formula without any cor- reetion factor. It is difficult to coil springs with a very small amount of initial tension, because some springs will then have some open wound coils. Very high values of initial tension may require special coiling techniques and. may increase cost Normally, the initial tension varies slightly from coil to coil, so itis necessary to extend the spring an apprecia~ ble amount before overcoming all of the initial tens ‘A practical method of measuring the initial tension in 2 spring is as follows: 236 1. Extend spring some unit of length until all coils are open, then measure load. 2, Extend spring another equal unit of length and measure load. 3. Subtract the first load from the second load to ‘obtain the scale for the selected unit of length. Subtract the scale for the selected unit of length from the first load to determine the initial tension. ‘After separation of all the coils, the slope of the load- deflection diagram is the same as for an open wound spring. 7. Spring Design ‘The design of extension springs, apart from the need for making allowance for initial tension, differs little from the design of compression springs. The end construction affects the deflection of the springs, and spring calcula- tions should include an allowance for the effect of end deflection, In the case of full hooks or loops at each end, the total number of active coils for rate computations can bbe taken as the number of body coils plus one to allow tor end effect The stress levels for extension springs are substantially lower than the stress levels for compression springs. There are several reasons for this. In the first place, the tor- sional stress in the body of the spring must be low to avid high bending stresses in the loops. Second, extension springs are normally made without residual stress, so limiting factor is the torsional elastic limit rather than the high apparent stress which results from taking out set in ‘a compression spring. However, in cases where stresses are needed, it is possible to coil the spring with excess initial tension and then stretch it, removing initial tension and inducing residual stresses. If this practice is used, the remaining initial tension cannot be high, and it may be necessary to use coned ends or inserted hooks to avoid excess stress in the loops. Without residual stress, the maximum stress may be a compromise with the requirement for initial tension, Normally, springs are given a low temperature heat treatment to obtain the max- imum elastic limit, but this heat treatment may reduce the initial tension in an extension spring by 25-50% while increasing the elastic limit by a similar amount Still another reason for lower stresses isthe hazard that an extension spring may be extended beyond its design length either in assembly or by careless handling. Loops con the side will cause a higher stress than loops over cen- ter which apply the load axially. ‘Considering all of these uncertainties, a reasonable uncorrected stress level is 359% of tensile strength for fer- rous materials and 30% of tensile strength for nonfer~ rous and austenitic stainless steel for static or mild service. Prediction of fatigue life is very difficult because of fur-400 SPRING INDEX C. = 10 0 Pw B 70 z 2 & ew z = 0 2 2» 2 10 5 = STRESS OUE TO INITIAL TENSION adds “o Fig, 36Inal tenon in extension springs ther complications, such as surging caused by impact loads and increased stress which results from bowing of the spring, it it operates in the horizontal position. Exten- sion springs have many advantages, but experimental work and testing are necessary to utilize their maximum potential, 237 WHERE: S).4 = C. TOLERANCES FOR HELICAL COM- PRESSION AND EXTENSION SPRINGS 1. General The tolerances in this section apply to cold coiled springs made of round wire. When applied with soundjudgment, they represent economical limits for good com- mercial springmaking practice. Closer tolerances can be ‘maintained, but usually with increased cost. ‘The perfect universal tolerances system for cold coiled helical compression and extension springs has not been devised. That it will ever be is unlikely. ‘There are many reaons for this situation. One is the peculiar interdependence that each dimension of a spring hhas upon its other dimensions and upon its load- deflection characteristic. Another reason is that certain aspects of specifying spring tolerances can have a great influence upon the cost of springs. Ifthe designer is una- ware of them, or chooses to disregard them, he may design considerable extra cost into a spring without increasing its value. ‘i The wire which spring makers use in the manufacture of cold coiled helical springs is purchased to standards or specifications that include tolerance for the diameter and mechanical properties of the wire. Variation within the limits of these tolerances causes significant variation in the dimensions and load-deflection characteristics of springs. This variation in wire is one reason why spring makers recommend that designers apply realistic tolerance only to mandatory requirements and specify other require- ments only as approximate, suggested, or limiting values. The designer should specify either free length, solid length, or loads at specific lengths. If more than one of these is specified, the secondary requirements will have «a wider tolerance which includes the effect of the spring rate tolerance. When the above practice i followed, the spring maker has the opportunity to make compensating adjustments in the noncritical dimensions in order to maintain the mandatory requirements. Without this freedom for adjustment, the use of special controls or salvage opera- tions may be required, which would increase manuf: ‘turing costs The magnitude of the manufacturing tolerances required for a spring of a given design depends upon ‘many factors. Sometimes a customer allows insufficient space for a satisfactory spring. This results in a spring design that is overstressed and allows very little manufac turing variation. Tolerances in such a case are severely restricted by the application, and cost must become a secondary matter. Generally, cost is a major consideration, and springs must be produced in a most economical manner. Speci- fied tolerances need to be generous enough to allow acceptable springs to be made using ordinary methods. ‘Those types of wire made to close tolerances, such as music wire and carbon steel valve spring quality wire, tend to reduce variation in spring dimensions and permit the holding to closer tolerances than is the case when the com- mercial grades of wire are used. Compression springs designed to high stresses, within safe stress limits, tend to vary less than springs designed to low stresses. ‘The degree of nonconformance which the customer's receiving inspection will accept can be a vital factor in the cost-tolerance relationship. When the spring maker is aware of the customer’s inspection and acceptance practice, he can adjust his control measures accordingly to meet the required quality level at lowest possible cost. ‘The above is particularly true for large production runs which allow time for process refinement and the effec- tive application of statistical control. When this can be done, it tends to reduce variation in spring dimensions to the minimum, and very close tolerances can be met. Many drafting rooms use standard drawing forms hav- ing a tolerance box for machine dimensions. These box tolerances are almost always impracticable for springs. It is strongly recommended, when drawing forms with box tolerances are used for springs, that the tolerance box bbe crossed out and that realistic spring tolerances be speci- fied for mandatory requirements. Figs. 3.7-3.13 give tolerances for dimensions, load, and rate for compression and extension springs. ‘The diameter and out-of-round tolerance for the com- ‘mon grades of spring wire can be found in the SAE Hand- book under the section entitled Spring Wire and Springs 2. Coil Diameter Coil diameter tolerance may be specified om either the inside or outside coil diameter. The tolerances in Figs. 3.7 and 3.8 are functions of the wire diameter and the ratio of mean diameter to wire diameter. The tolerances are to be considered as manufacturing tolerances and do not take into account changes in diameter due to applied loads. Fig. 3.7 gives coil diameter tolerance for springs having wire diameters in the range 0.30-9,50 mm. Fig. 3.8 gives coil diameter tolerance for springs having wire diameters in the range 9.$-16.0 mm, 3. Free Length The free length tolerances specified in Figs. 3.9 and 3.10.can be maintained, in most cases, unless changes in free length are required to compensate for other varia- bles which affect load. Therefore, when a load with tolerance is not specified, it is advisable to specify the free length with tolerance for control and inspection purposes, Otherwise, it is recommended that the free length be a reference dimension. In the case of compression springs, the free length is an overall dimension measured parallel to the axis of the spring. For extension springs, the free length refers to the 2382.00 1.50 1.00 E é 8 3 os0 e & oso 3 F030 « E 3 Z 020 a = 5 8 0.10 0.05 0.40 0.50 INDEX (D/d) 8 5.00 10.00 WIRE DIAMETER, mm Fig. 3:7—Coil ameter tolerance—compression and extension springs for wire dlameters 0.30 to 9.60 mm. Round off index to nearest whole number Interpolale when the rounded.olt value Is an adé number. Use tolerance for 0.90 mm wire diameter when wire diameter 's ase than 0.30 mm. length inside to inside of the hooks. Free length toler- ances for compression springs (expressed as mm per mm of free length) are shown in Fig. 3.9 as functions of the number of active coils per mm of free length and the spring index. Free length tolerances for extension springs expressed functions of the free length appear in Fig. 3.10. 4. Load Load tolerances for compression springs, as functions of the nominal free length tolerance of Fig. 3.9 and the deflection from free length, are shown in Fig. 3.11 239 The normal load tolerance for an extension spring equal to the product of the appropriate tolerance factor from Fig. 3.12A and the appropriate multiplying factor from Fig. 3.12B, Tolerances determined in the above ‘manner are functions of the spring index, the ratio of free length to deflection, and the wire diameter. 5. Rate ‘When spring rate is a functional requirement, it should be specified with tolerance. Normal rate tolerances are shown in Fig. 3.13 as functions of the number of active coils.4.00 3.00 2.00 INDEX (D/d) 1.00 9.80 COIL DIAMETER TOLERANCE, = mm 0.60 0.40 0.30 90 100 11.0 120 130 140 160 160 WIRE DIAMETER, mm Fig 38—Cot iametertlerance—compression and extension springs lor wie diameters 9.5 to 16.0 mm. Round of index to nearest ‘ole number interpolate when rounded value Is odd numb. 16 0.040 4 ho 0.030 £ E fiz ee g 2 4 I oy : 48 oo ee = ES oo a GE a Ze t x Zo i 9 we oowk Zz c 0.009 - = gs 8 38 g g 8 £8 & NUMBER OF ACTIVE COILS PER mm OF FREE LENGTH Fig. 8 9—Free length tolerance—compression springs. Round off index to nearest whole number. interpolate when roundedot value is ~ edd number. These are tolerances for springs with ends closed and ground, For springs with ends closed ut not ground, multiply by 1.7 2.40| | | | | | FREE LENGTH TOLERANCE, + mm 10 & 8 288 8 FREE LENGTH (INS length tolerance—extension springs 6. Solid Length If the working range of a compression spring approaches solid compression, it is advisable to specify the solid length as a maximum. If the spring ends are closed and ground, the lowest specified maximum solid length should be not fess than the product of the maxi- ‘mum allowable number of coils and the maximum allow- able wire diameter. If the ends are not ground, the lowest specified maxi- mum solid length should be not less than the product of the inaximum allowable number of coils, plus one, and the maximum allowable wire diameter. When the wire diameter is less than 2.50 mm, itis advis- able to add to the solid length an additional amount equivalent to one-half wire diameter to compensate for thick ends which occasionally occur during production end-grinding operations. Failure to make this allowance could require a costly sorting operation after grinding, Finishes such as paint or plating, when applied to Springs, significantly increase the solid length of the 2a 8 & § $282 IDE TO INSIDE), mm springs and must be allowed for when calculating solid lengths. 7. Number of Coils Tolerance on the number of coils is given in Table 3.1 for compression springs and in Table 3.2 for extension springs. The tolerance is expressed in degrees as a funk Gon of the number of active coils. 8. Squareness and Parallelism of Ends ‘The squareness of the ends of compression springs hav- ing ends that are closed and ground should be within 3 deg of the spring axis when measured in the unloaded position. ‘The ends shall be parallel within a limit equal to twice the squareness of the ends.LOAD TOLERANCE, = % 160 FREE LENGTH TOLERANCE, + mm FROM FIG, 39 - aon & 8 £88 Bs 2001 DEFLECTION FROM FREE LENGTH, mm Fig. 311—Load tolecance—compression springs, +8. Enter chart from bottom with detection from free length to loaded length and from lett with fre length tolerance of Fig. 8.9. Round off percent load tolerance values 1 next larger whole number. interpolate when rounded-off value Is odd and between 8 and 25%, ‘TABLE 9.2—NUMBER OF COILS TOLERANCE OF EXTEN: SION SPRINGS TABLE 3.1—NUMBER OF COILS TOLERANCE CF COM. Tolerance, +409 PRESSION SPRINGS coe = ave Tolerance Active Coils Wound Wound Cols 09 7 s ol 310 s 410 45 90 For each additional 10 cols, ado 30 For each adkitional 10 coll, ade 15, 30 | 2.42TOLERANCE FACTOR, + PERCENT 3 Q 2 1 trom Fig. 3.128. L i il | | FREE LENGTH sh} DEFLECTION i 05 06 08 10 15 20 30 40 6.0 60 Fig. 312.8 fUcTIPLyinG FACTOR MIRE OIAMETER, mn 05 06 (0810 20 ao 4080 Fig. 3.128 243 LAL 80 100 INDEX (D/4) Fig.3,12-Load tolerance — extension springs, 96. To fin load tolerance, multiply tolerance facto rom Fig. 9.128 by multiplying factorS B & S & & S 2 < & 4 3 2 # < é NUMBER OF ACTIVE COILS -0—compression and extension springs. Rate and rato tolerance should be specitid only when rate is functional ‘of deflection over which tis to apply must be cary dented. Deflection ranges for rats conto; should fal within 20-60% limit of total deflection because rate is tkely fo be variable outside this range, D. HELICAL TORSION SPRINGS 1. General Torsion springs store energy or offer resistance to an applied torque when subjected to an angular deflection. This deflection Tesults in an increase in the number of coils in the spring and a relative decrease in the spring diameter. The term “torsion spring’” may be confusing, since compression and extension springs are subjected to torsional stresses, while torsion springs are subjected to bending stresses. ‘The contour of a torsion spring is usually cylindrical; however, it can be conical, hourglass, or barrel-shaped. The use of hovrglass or barrel-shaped torsion springs makes it possible £0 obtain an increasing rate. Assembly Of these springs over a rod of the same size as the inside diameter of the smallest coil results in some of the spring material becoming inactive during deflection, thus increasing the spring rate. Torsion springs may be either close wound or open wound. Open wound springs are recommended wherever possible, since close wound springs produce frictional 2.44 loads which are difficult to evaluate, When an open ‘wound torsion spring is used in a confined space, the length of this space must be adequate to accommodate both the increased number of coils due to winding of the spring and the variations in the original number of coils in manufacture, Close wound torsion springs increase in length as they are wound up. Clearance must be allowed for this increase in length and for variations inthe springs; otherwise, the springs will bind, and breakage or setting will result. Torsion springs may be coiled from annealed wire and fully heat treated after coiling for certain applications, particularly where unusually difficult end coil formations are specified. It is by no means uncommon to produce heavy torsion springs from large round or square bars which are coiled hot, as described in detail in Chapter 4 which relates to hot coiled springs. However, it is by far the most general practice to coil torsion springs from cither hard drawn or oil tempered wire. ‘Square and rectangular wire cross sections are fre~ quently used in torsion springs due to their greater mate- rial efficiency compared with that of round wire: However, the unit cost of square and rectangular wires is generally greater than that of round wire, and, in con-sequence, the use of round wire springs usually results in a lower net material cost. It is to be appreciated that “keystoning”” will resule when square and rectangular wires are coiled and that allowances for the effect should bbe made, as discussed in Chapter 5, Section 6. 2, Effect of Direction of Coiling and Benefi cial Residual Stress Torsion springs should be designed so that torque is applied in a direction which tends to reduce the coil diameter. Cold wound helical torsion springs made from, hard drawn or pretempered wire must be coiled to a smaller diameter than that desired in the finished spring, to allow for recoiling when released from the coiling arbor. This causes a tensile residual stress to be set up ‘on the inside of the coil and a compressive residual stress on the outside. If the spring is then loaded by a torque, 80 as to reduce the coil, it is clear that these residual stresses. will be subtracted from the maximum load stresses near the outer fibers. These beneficial bending stresses are analogous to the effect of presetting on tor- sional stress distribution over the cross section of helical springs. To take advantage of these residual stress pat- terns, torsion springs shoutd therefore be loaded in the same direction they were coiled. 3. Support of ‘Torsion Springs Torsion springs can be wound without appreciable set until the inside diameters of the coils are only slightly larger than the inside diameter to which they were origi- nally coiled before recoiling from the arbor. Winding to any smaller diameter would be equivalent to recoiling from a smaller mandrel than that used for the original coiling and would produce substantial set. For this rea- son, it is necessary to support adequately torsion springs which are coiled in such a direction that the diameter is reduced by the applied torque. The most common way to effect adequate support is to insert a rod through the spring, and experience shows that the diameter of this rod should be approximately 90% of the smallest inside diameter to, which the spring is reduced when under max- imum load or travel. Insufficient support on the inside is conducive to early failure. The size of the supporting rod, as indicated above, is affected by the decrease in the coil diameter due to the ‘angular deflection of the spring. The reduced mean diameter may be determined by the following formula: 245 where: D_ = free mean coil diameter D’ = reduced mean coil diameter N, = free number of active coils N%, = increased number of active coils Types of Ends and End Stresses The variation in torsion spring ends is almost limitless, but a few of the more common types are illustrated in, Fig. 3.14. The double torsion type of spring shown is, necessary for some applications, but it is expensive to ‘make. In many cases, it may be less costly to use two, springs instead, one left-hand and the other right-hand, When torsion springs are used with eyes at the ends or with bends off the coil, special care should be taken to keep stresses low. When bendsvare made at smaller radii and then loaded, the stress becomes tensile at the inside of the bend and may be much greater at that point than the stress in the coil. Due to the sharp curvature, the neutral axis moves in toward the center of the bend, and the tension stress becomes that of a cantilever load- ing multiplied by a constant greater than unity, depend- ing upon the degree of curvature. A close approximation, of the peak stress at such bends is given bi OH ‘SPECIAL ENDS Lif STRAIGHT TORSION DOUBLE TORSION Fig. 3.14 Details of helical torsion springs‘TABLE 9.3-HELICAL TORSION SPRING FORMULAE Rectangular Round Wire ‘Square Wire wire coo cor Eepe Feo FieOND oom 60M 8 = be Eo ew Ebe MO = SerEND PiBOND Feon,0 = axial dmension {= raga dimension where: S. = peak stress, MPa S =" stress from torsion spring ‘formula, MPa ty = radius of bend to center line of wire, mm 1 = radius at inside of ends, mm 5. Specifications Torsion spring applications are seldom alike and have no standard forms. It is recommended that all inqui and orders for torsion springs be accompanied by draw. ings clearly illustrating the types of ends, the diameter of the rod over which the spring must work (if one is used), the length of space availeble, and the position of the ends at maximum working deflection. In addition, the torque in newton-millimeters at loaded position and the angular position of the ends under this torque should be given. Particularly important is whether the spring should be left- or right-hand wound. 6. Design Formulae ‘The constants in the formulae of Table 3.3 are derived from the theory of the bending of straight bars. Tests may show appreciable differences from calculations as dis- cussed in Part D, Section 8 of this chapter. If the length of the ends is appreciable in comparison to the body of the spring, this length may be converted into the number of coils and included with N, in the oad deflection formulae. 7, Design Stresses ‘The maximum design stresses allowed for static load- ing with various commercially available spring materials, can be established as a percentage of the tensile strength of the material, as shown in Table 3.4. The end stress should also be considered in any torsion spring design, as discussed in Part D, Section 4 of this chapter, 8. Commercial Tolerances In the accompanying outline of tolerances for torsion springs, no load tolerances are listed. There are several reasons for this. Standards have not been set up as to what size arbor a spring of given ID isto be tested on. If and when the arbor size is agreed upon, then the amount of coil tension must be set, as this also affects friction and hence input and output of the springs. So far, no accepted standard test machine has been developed for universal use as has been done in weighing devices used on extension or compression springs. In order to test tor- sion springs, the vendor and customer must develop their ‘TABLE 3.4 ALLOWABLE DESIGN STRESSES AS A PERCENTAGE OF TENSILE STRENGTH FOR VARIOUS MATERIALS SAE No. Type of Wio 1% of Tonsila Seongth 32 Oi tempered aioy wire aa 1316 Oi tempered carton wire 76 ue Muse 8 ung Hard drawn carbon sie! co Hard drawn Stainless type 202 65 Phosphor Bronze spring wire 7 2.462.00 1.50 1.09) 0.80] 0.60 0.80) 0.40] 0.30 0.20 COIL DIAMETER TOLERANCE, + mm 0.10) 0.08 0.06 0.05 0.04 Sie ese g INDEX (D/d) 4.00 ges WIRE DIAMETER, mm. Fig. 8.15—Commercial col diameter tolerances for helical torsion springs with wire diameters 0.90 to 7.00 mm. Round off indexes 10 nearest whole number. interpolate when roundedoff valve Ie odd number own testing technique on the spring in question, making Such allowances for friction, end position, diameter, and Wire size as are inherent in the product. Such a proce- dure makes load testiag of each spring type come under a different method, In the present state-of-the-art, this hhave to be endured, as there is no remedy yet in sight, Coil Diameter Coil diameter tolerances can be specified on either the inside or the outside diameter of the coils, depending ‘upon the importance of the respective dimensions to the 247 uses. These tolerances apply to torsion springs made of 7.00 mm or smaller wire size on torsion coiling machines, They are given as functions of the ratio of the mean diameter D to the wire diameter d, as shown in Fig, 3.15. These tolerances are to be considered as manufac turing tolerances and do not take into account changes in diameter resulting from angular deflection Wire Diameter For wire diameter tolerances, refer to the SAE Hand- ‘book under the section entitled Spring Wire and Springs,Position of Ends ‘The ends or position of the arms with respect to the coil of a torsion spring are subject to the tolerances shown, in Table 3.5 up to and including a D/d ratio of 16. Fig. 3.16—Flat spiral springs TABLE 9. TOLERANCE FOR POSITION OF ENDS 2.48 ‘Number of Gail up 03 Over 3 thea 10 Over 10 tru 20 Over 20 ts 30 Over 30 Tolerance, = 409 8 0 15 20 2sE, FLAT SPIRAL SPRINGS 1. General A flat spiral spring consists essentially of flat spring material wound on itself with open space between the coils in the free position. These springs are fundamen- tally made to deliver a torque, but this may be translated to linear action by means of linkages. They provide a ‘compact means for storing energy and, therefore, are often used to produce force or motion in preference to other types of springs. Flat spiral springs are very widely used for locks and window counterbalances in automobiles and for main- taining pressure on carbon brushes of electric motors and generators. Typical examples are shown in Fig. 3.16. Another closely related type of spring is identified as aclock or motor spring. In this style of spring, the length of material is so great that itis confined in a case or drum, and a majority of the coils are contacting each other in the unwound position. As the spring is wound, the active length changes continuously so the load-deflection curve is nonlinear. The design of this style of spring is a spe- cialty which will not be covered in this manual. It is, recommended that the manufacturer be consulted in design problems of this spring style. Spring Materials ‘The material for flat spiral springs is described as flat wire in Chapter 2. By far the greatest proportion is made from oil tempered wire, although for certain applications, the springs are formed from untempered or annealed ‘material and are fully heat treated after forming. Spe- cial applications occasionally require the use of stainless steel, nickel alloys, or nonferrous materials. 3. Design Formulae The formulae used for this type of spring are similar to those found under the torsion spring section: os7MLs _ 6M Eb? be Variations may be expected from the calculated deflee- tion values for several reasons. If the material is rolled from round wire so the edges have small radii, the load may be about 5% weaker than predicted. If the outer end is free to swivel about a pin, the load may be weaker than with a clamped end, If the active length varies as the spring contacts the arbor, or if the coils touch, the rate will increase. If only one load testis specified, itis cus- tomary for the manufacturer to vary the free position to compensate for these deviations. If two load tests are specified, it is necessary to-vary the developed length to meet a rate requirement. The arbor size and free outside diameter do not affect the load-deflection characteristics very much, but they do affect the total available deflection. If the arbor size large and the outside diameter is small, the overlaps may be so close together that very little deflection is avail- able. A rigorous equation for total deflection would be very complex. The following approximation equation is useful as a guide: Number =D. 2b. rant 7D, + D, ‘The equation assumes uniform spacing of successive coils and does not account for the extra thickness of mate- rial at the first overlap. Springs made from tempered ‘material normally have a slight increase in spacing as the coils get larger. This gives less total deflection than uni- form spacing. Therefore, the total deflection will be less than the formula indicates. Because of the large amount of residual stress, the apparent stress for static applications may be as high as the tensile strength of the material, Fig. 3.17 shows the stress versus thickness for steel springs with maximum residual stress, which is normal when the springs are made from tempered material, 2.49THICKNESS—mm 250 Fig, 2.17—Siress versus thickness for ste! springsChapter 4 Hot Coiled Helical Springs 1. General Definitions Helical compression and extension springs are normally cold wound when the material is less than 10 mm in diameter (see Chapter 3), and hot coiled when it is greater than 16 mm. When the cross section is between 10 and 16 mm, both processes are common, Hot coiled springs find extensive use in many fields of heavy industry and particularly in automotive, railway, agricultural implement, construction equipment, and ord- nance applications. Basically, they are designed accord- ing to the general helical spring formulas of Chapter 5, bbut some of their features require special attention. Spring stee! bars are available in carbon and alloy ana- lyses. The bars are generally used in the as-rolled condi- tion, but may be centerless ground before coiling. Table 2.2 lists some materials commonly used in hot coiled springs. 2. Heat Treatment Ina typical hot coil spring operation the bar is heated, coiled, and quenched in a continuous operation. Care ‘must be used to prevent excessive temperature before coil- ing, while maintaining adequate temperature for quench- ing. After quenching, the spring is tempered to produce the specified hardness. It is important that the finished spring have a fine grained structure and a surface free from excessive pitting, roughness, or other defects. Its sometimes desirable to heat treat a spring that has cooled below the critical temperature for quenching. This, may occur when a spring receives secondary operations after coiling or as part of a salvage operation. This is sometimes referred to as double heat treat and involves heating the spring to the proper temperature, quenching, and tempering. This process will produce satisfactory, springs, but care must be used to prevent excess scale and, decarb during two heating operations. Additional information can be found in SAE 5412, “General Characteristics and Heat Treatments of Steels" and SAE J413, “Mechanical Properties of Heat Treated ‘Wrought Steels.” 3. Shot Peening Shot peening is used to increase the fatigue strength of springs, thereby helping to reduce their size and cost. Refer to Chapter 1, Section 5 for details. Hot coiled springs need shot peening to remove scale in addition to its other benefits 4. Presetting Hot coiled compression springs normally receive a manufacturing operation commonly cailed presetting. ‘This process consists of coiling a spring to a free length greater than specified free length. Following heat treat- ‘ment, the spring at room temperature is compressed solid or to a specified preset length, to produce yielding. For high stress springs, the preset stress (corrected with Wahl. factor) is usually 1100-1150 MPa. Presetting is covered, in more detail in Chapter 1, Section 5. 5. Warm Setting In order to reduce the **sag”” or “settling” of helical suspension springs which occurs when they are subjected to vehicle loading over time, it has become common prac- tice to warm set—or hot press—the springs at an elevated temperature (usually in the range of 120°C-260°C depending on the particular spring design). One theory holds that the major benefit of this operation results from. 251‘TABLE 4.1—PREFERRED DIAMETERS OF ROUND STEEL BARS (ANSI B32.4) 90 | 1 | 90 | 5 as| 19 | 2 | 60 roo | 20 | 35 | es tio | 21 | 38 | 70 120 | 22 | 28 | 75 yao | 23 | 40 | a0 rao | 24 | a2 | a5 150 | 25 | as | 90 wo] 26 | a3 | 05 ira | 28 | 50 | 100 TABLE 4.2—CROSS SECTION TOLER. ANCES FOR HOT ROLLED CARBON AND ALLOY STEEL ROUND BARS Species Diameter,mm | _ Tolerance, | Out of Plus and Minus, | Round, Over | thw mm mm 8 013 020 10 018 022 15 O78 027 2 020 030 2 9023 034 30 025 038 35 020, os 40 0.35 052 60 0.40 060 80 060 0.90 100 0.80 120 TABLE 4.3-CROSS SECTION TOLERANCES FOR TURNED AND POLISHED OUND BARS Toerance, Minis On im spec Maxinan of Ske. amu ot | Carbon Range oy | camon ange | “Over ass Maximum ot | Meoour 10 0.55 incl ‘Also all Grades Cavon ange | CR" Fande | Arneaied or | Quorched and Gab ortece | O*C02810 | yess Releved | Tempore o Over | Thru in after Cold Normalized and Fnesing | Tompued betore Gi Fang Carbon Stet =] @] 0 O08 a0 on wo | | O08 ov oie os fo | so} O90 oe 1s oe 100 | tsa | - “O12 os oe 020 too | 200 | os O18 820 oe zo || O18 O20 oz 026 zs || 020 ae a5 2 toy Stet Cote —La[_o a8 om one Cut Leng =] @] 008 a0 oe ons wo | so} o oe os oe & | so} ov os oe 020 ro | to | as oe 620 O22 ts | 200 | oe 0 O22 as zo | Z| a0 oe 035 ae zs | | oe 5 028 030 252TABLE 4.4—CROSS SECTION TOLERANCES—STEEL BAAS (ROUND), CARBON AND ALLOY: COLD DRAWN, GROUND AND POLISHED AND TURNED, \ GROUND AND POLISHED actual size is to be specified. Generally hot rolled steel bars are specified for hot coiled springs. Table 4.2 shows tolerances for hot rolled steel bars in the range of diameters generally required. Other processed forms of aa round steel bars are available such as cold drawn, turned Speciied and polished, ground and polished, and turned, ground, Diameter, mm | Cold Orawn, | Turned, and polished. These are generally specified when close Ground and | Ground and Fi es ere ee tolerances are required; to control decarburization; for reduction or elimination of surface imperfections; and — | a] oo 0.03 availability in limited quantities. See Tables 4.3 and 4.4. 2 | 50} aoe 08 so] 75 | 005 0.05 7 | 100 | 008 0.08 Bar Length 100 } 150 a 0.10 150 | 200 7 out Bars are commonly purchased to exact length. Sug. ‘For nonesuiphuized carbon steels (steels specified Yo max sulphur limite under 0.08%) or for any stent quenched and tempered (Meat treated) normalized and tempered. oF any Sm lar double Weatment poe fo turing: the toler ance is inereased by 0.03 mm, an increase in the amount of strain hardening that occurs when the spring is stressed past the proportional limit (point ““A” in Figure 1.4), Increasing the temperature lowers the proportional limit to some stress lower than “A”, and therefore if the spring is still stressed to poTfit “B”, the amount of strain hardening that occurs is greater. This increase in strain hardening will reduce the dynamic or static settling (load loss) that occurs over the useful life of the spring. ‘A second theory is that a more effective beneficial residual stress pattern is set up over the bar cross section when a spring is warm set at elevated temperature. ‘Also refer to the general treatment of presetting at an elevated temperature which is given in Chapter 1, Sec- tion 4, oe It should be noted that a final (cold) presetting opera- tion is still necessary. In general, warm setting will decrease the load loss by ‘more than $0%, depending on the working stress level. 6. Specifications and Tolerances Bar Diameter Round bars are available as preferred sizes as described American National Standard, ANSI B32.4, “Preferred Metric Sizes for Round, Square and Hexagon Metal Products,”” Excerpts from this standard are shown in Table 4.1. Since spring performance is predicated on Sal factors including the bar diameter, adherence to referred sizes is not generally feasible and the desired 258 ‘gested sheared length tolerances for hot rolled carbon and alloy steel are found in Table 4.5. Length tolerances of special straightened machine cut bars that are approximately one half of the tolerances Table 4.5 may be found in ASTM A29. For other toler- ances consult the manufacturer, ‘TABLE 4.5—LENGTH TOLERANCES FOR HOT ROLLED CARBON AND ALLOY ROUND STEEL BAAS ‘Specified | Length Tolerance, Diameter, mm | Plus Only, mm ‘Over 15003000 over | thu | Tar 3000 — | 2s 2 20 2 | 50 we 50_| 100 2540 Coil Diameter The coil diameter can be expressed in terms of the mean coil diameter (D) which is used in the rate and stress for- mulae, However, coil diameter tolerances should be speci- fied on either the inside diameter (1D) or the outside diameter (OD) of the coils, depending upon the impor- tance of the respective dimensions to the user. Tolerances are shown in Table 4.6, based on coil diameter and sp length. For motor vehicle suspension springs, it is customary to specify the ID in order to facilitate the coiling of a family of springs on a single arbor. Where tangent tail ends are specified, additional clearance must be provided for the tangent tail end where the straight portion of the bar extends beyond the outside diameter, since the last 12-25 mm of coiled bar length does not continue to wind in a uniform diameter. See Table 4.6.TABLE 4.6—COML DIAMETER TOLERANCES For Free Spring Length, mm Inside or Outside Diamat Tolerance, Plus and Minis, mn For Specie or Computed Outside | Upto | Over over ver over Diameter, mm | 250 | 250 thru 450 | 450 thru 650 | 650 tha 850 | 850 thru 1050 750 0m 1100 | 08 13 25 36 45 Over 1100 tu 1500 | 1.3 25 36 46 56 Over 1900 thru 2000 | 25 56 36 56 be Over 2000 hry 3090 | 36 46 56. 65 65 Spring Lengths Spring lengths are to be measured after preloading (see “Preload Length”) as the distance parallel to the axis, between the end surfaces, or else between two reference points specified on the spring drawing, Free Length—Free length is the length when no exter- nal load is applied. When load is specified, free length is used as a reference dimension only. When load is not, specified, free length tolerance equals + (1.5 mm + 4% of free-to-solid deflection), Solid Length—(see also “Number of Coils””)—Solid length is the length when the spring is compressed with an applied load sufficient to bring all coils in contact; for practical purposes, this applied load is taken to equal approximately 150% of the load beyond which no appreciable deflection takes place. See Table 4.7 for the formulae to calculate the solid length. When nominal, solid length is specified, the tolerances shown in Table 4.8 apply. Preset Length—In the presetting operation (see “Presetting™) the spring is usually compressed solid. However, if the stress at solid length is so high that the spring would be excessively distorted, the presetting oper- ation may only be carried to a specified preset length. If more than one preset compression is desired, this must, be specified on the drawing. Preload Length—Preloading is the operation of deflecting the spring to the preload length in order to remove temporary recovery of free length before the spring is checked for load and rate. If the spring was preset during the manufacturing proc ss to the solid length, the preloading may also be car- tied to the solid length; but it may be restricted to a preload length slightly greater than the solid length, provided the maximum deflection during subsequent serv- ice will not go below the preload length. If the spring was preset to a specified preset length greater than the solid length, the preloading should be restricted to a preload length greater than the preset length. However, the preload length must not exceed the mini- ‘mum spring length possible in the mechanism for which the spring is designed. In suspensions, this is called the “length at metal-to-metal position.”” The metal-to-metal contact will occur in the suspension mechanism when rub- ber bumpers are disregarded. The spring deflection from the specified loaded length to the metal-to-metal position is called “Clearance.” Loaded Length—Loaded length isthe length while the Joad is being measured; it isa fixed dimension, with the tolerance applied to the load, Load Load is the force in newtons (N) measured on the load testing machine required co deflect the spring to the speci fied loaded length. It isto be measured during compres- sion of the spring (compression load) and not during release of the spring (release load),‘unless otherwise specified. With loaded length fixed, the usual tolerance for motor Vehicle suspension springs is expressed in terms of load equivalent to a deflection of +5 mm at the nominal rate. ‘Where the demand for greater accuracy warrants the cost of additional presetting or other operations, the load tolerance may be specified as low as + 1,50 mm at the nominal rate, In the springs for general automotive use, the load tolerance (with loaded length fixed) typically equals £(1.50 mm + 3% of free-to-solid deflection) at the nominal rate. This tolerance is limited to springs where the free length does not exceed 900 mm or six times the free-to-solid deflection, and is not less than 0.8 times the oD. Rate Rate is the change of load per unit length of spring deflection (N/mm). In the springs for motor vehicle suspension, the rate is expressed in terms of the load increase per 25 mim 254{ TABLE 4.7—FORMULAE FOR TOTAL COILS AND FOR NOM TABLE 4.8—SPRING SOLID LENGTH vac | aman Deaton ot Ree ane ee eee ele] ‘used to compensate for the cosine effect of the com 590 8 = 101 = tac engl. ‘The bracketed term inthe solid length formula or springs with two pita ends may vary betwoen (N,~ 0,90) and (Ne ~ 1.60). de ending on the pig detais deflection (N/25 mm). It is therefore determined as one half of the difference between the loads measured 25 mm. above and 25 mm below the specified loaded length. Tolerance is +3% with centerless ground or with preci- sion rolled bars, and +4% when commercial hot rolled bars are used. In the springs for general automotive use, the rate is determined between 20 and 60% of the total deflection unless otherwise defined. Typical tolerance is 5%. In non-critical applications, this may be increased to + 10%. Number of Coils, Total number of coils (N.) are counted tip to tip, active number of coils (N,) are specified as the number of working coils at free length. With increasing load, Ny may progressively decrease due to the “bottoming out” effect. Rate is inversely proportional to active coils and will therefore be higher than calculated when itis checked. ‘at heights or loads which cause significant bottoming out. Itis sometimes necessary For the designer to compensate for the bottoming effect by increasing the calculated num- ber of active coils. If no appreciable bottoming out occurs, the relationships between N, and N, are as shown in Table 4.7 which also gives the formulae for ‘nominal solid length. ‘Since nominal solid length may be exceeded somewhat by actual solid length due to manufacturing variation, ‘a frequent practice is to specify nominal solid length ‘together with a maximum solid length, as shown in Table 48, End Configurations (See Fig. 4.1.) Four types of configurations are used: ‘Taper Rolled—This end configuration is defined as having the bar end taper rolled prior to coiling, genet- 255 ally for 240 deg, to a point thickness (t) of approximately 33% of the bar diameter. After coiling, the tip of the tapered bar is to be in approximate contact with the adja- cent coil and must not protrude beyond the outside diameter of the spring by more than 20% of the bar diameter. A squareness limit of 3 deg with the axis of the spring is normal, but where finished end bearing sur- faces are required for critical applications, the ends may be ground. This grinding operation is performed pérpen- dicular to the axis of the spring helix. The resulting ‘ground bearing surface must not be less than two-thirds of the mean coil circumference, nor narrower than half the width of the hot tapered surface of the bar. This type of spring is used with fuil size flat seats. ‘The advantages of this end type are: a) Material savings due to elongation during taper rolling. 'b) Any partial number of total coils can be designated, because rotational orientation is not required. When max- imum stress must be reduced, a partial coil may be added rather than one complete coil (as in the case of a double tangent tail). This can result in a significant material savings. ‘) It requires less vertical space than tangent tail ends. Disadvantages: a) Taper rolling is an added operation requiring con- siderable equipment and floor space. 'b) Two taper rolled ends require additional labor costs to insure squareness of the ends with the coil center line. ‘The bar must be guided and oriented on both ends dur- ing the coiling operation, and this often results in signifi- cant serap oF rework problems. Pigtail —The pigtail end is defined as an untapered end coil formed substantially smaller than the central coilsTW PIGTAIL ENDS Fig. 41—Typical ends for hot coiled compression springs Of the spring and in such a fashion as to have the out- board bearing surface perpendicular to the axis of the spring helix. This end type is commonly used where the spring must be clamped to the seat to prevent any unseat. ing. It can also be used to fit over a piioted seat without clamping. The pilot seat for a pigtailed end is considera. bly smaller in diameter than for a non-pigtailed end and may result in a more efficient design. A. pigtail end requires the least amount of vertical space and provides for optimum preset conditions. Rotational orientation is not required, therefore, any number of partial coils can be designated. Further, it requires the Jeast amount of inactive material. When forming the pigtail, usually a straight end portion is required to facilitate coiling. The length of this straight portion is dependent on the manufacturer's equipment, but it can have an effect on the size of the pigtail. The major disadvantage of a dou- ble pigtail design is the need for additional equipment to form the second pigtail ‘Tangent Tail—This end type is defined as an untapered end coil formed as a helix having a pitch substantially equal to the bar diameter and its diameter the same as the central coils. To facilitate coiling, however, a straight end portion approximately 25 mm long is permitted to Project tangent to the coil diameter. The resulting hel cal end requires a spring seat with a corresponding heli cal ramp. The main advantage of this end type is that ‘tis the simplest and least troublesome method of coil ing a spring 258 The disadvantages are: 4) It requires angular orientation during installation, b) It requires maximum vertical space. ) requires a greater amount of material than the pig- tail or taper rolled ends ) With two tangent tail ends, any design change in the number of coils must be in increments of full coils. ‘The angular relationship between the opposite ends must be a constant for any one particular design application, Flat Tangent Tail-—This end type is similar to the tan- ‘gent tail end except the last 220 deg (minimum) is coiled perpendicular to the axis of the spring helix. To facili- fate coiling, a straight end portion about 25 mm long is permitted to project tangent to the spring diameter, This ‘end type has advantages over the tangent tail end in that it does not require a helical seat and orientation at instal- lation. It also reduces the space requirements. A disad. vantage is that it operates with a large amount of inactive material, Springs can be specified to have any combination of the four types of ends. The combination of two tangent tail ends may involve a complex arrangement for index- ing the spring seats, unless the design of every spring is. adjusted to an identical number of total coils. Spring ends and seats are usually so formed as to ren= der approximately two- thirds to one coil inactive at each end, Direction of Coiting For most applications, the direction of coiling is unim- ortant; however, right- hand coiling is preferred because ‘most spring manufacturers are so equipped. When direc tion of coiling is important, as in the ease of concentri cally nested springs, it must be specified for each component spring, maintaining opposite directions for adjacent springs. For tangent tail springs, the direction of coiling must conform with the installation conditions. Uniformity of Pitch The pitch of coils in a compression spring must be sufficiently uniform so that when the spring is com- Pressed, unsupported laterally, to a length representing a deflection of 80% of the nominal free-to-solid deflec- tion, none of the coils must be in contact with one another, excluding the inactive end coils. This require- ‘ment does not apply when the design of the spring calls for variable pitch, or when itis such that the spring can- not be compressed to solid length with lateral support. ‘When the design of the spring calls for variable pitch, or when it is such that the spring cannot be compressed to solid height without lateral support, the above require- ‘ment does not apply. ieConcentricity of Coils \& At free length, the center of all coils must be concen- ric with the spring axis within 1.5 mm. This axis is the straight line connecting the centers of the end coils. Squareness of Ends Unless otherwise specified, the tapered ends of any spring having an outside diameter to bar diameter ratio of 4 or more, and a free length to outside diameter of 4or less, shall not deviate more than 3 deg from the per- pendicular to the spring axis, as determined by standing the spring on its end and measuring the angular devia- tion of the outer helix from a perpendicular to the plate on which the spring is standing. In the case of a tangent tail end, the spring must stand on a seat with matching helical ramp. Tolerances for springs outside these limits are subject to special agreement. Closer Tolerances. ‘The above specifications and tabulations apply to i} seneral hot wound springs. However, where large volumes are required, such as automotive suspension coil springs, closer tolerances on some spring parameters are rossible. The high volume requirements permit the use ~f highly automated production equipment and tooling which can be precisely controlled. These same reasons + apply to the stzel mills which permit them to produce a high quality product to be used for the large volume requirements. These reduced tolerances apply io the fol- ‘owing spring parameters: | Coil Diameter—Inside diameters up to 140: +£0.75 |) mm, Inside diameters over 140: 1.25 mm. Bar Diameter—The tolerances detailed in the tabula- tion of hot coiled springs (Table 4.2) apply for commer- cial hot rolled bars: tolerance for precision rolled and centerless ground bars is +0.10 mm. Bar Length—The bars are commonly purchased to an exact length (for length tolerance see Table 4.3). Normal Lond—The usual load tolerance is + 12% of the specified rate or + 110N. whichever is greater. Where the demand for greater accuracy warrants the cost of additional presetting or other operations, the load toler- ance may be specified as low as 6% of the specified rate or £55 N, whichever is greater. Rate—Rate tolerance is 3% for springs made from precision rolled or centerless ground bars and +¢4% when commercial hot rolled bars are used. The rate is normally checked as the average between 25 mm above and 25 mm below the normal loaded height. 7. Other Hot Coiled Forms Large extension springs may be made by the hot coil- {ng method, but they cannot be made with initial tension. If (as is usually the case) it is desired t0 have the coils in contact at no load, then it is impossible to preset the spring and thereby induce a favorable residual stress pat- tern. Therefore, the design stresses for extension springs ‘must be considerably lower than for compression springs of comparable dimensions (see Part B, Chapter 3). Hot coiled helical zorsion springs should be open coiled for the reasons given in Part D, Chapter 3. A round bar, while theoretically not the most economical section, should be used if possible because of availability and facil- ity of spring manufacture. A square bar may be used, if necessary, to secure additional capacity. Design infor- ‘mation and formulae for round bar helical torsion springs are included in Chapter 3, Part D of this manual 287Chapter 5 1. Design Formulae for Round Wire Springs All spring design is founded on Hooke’s Law, which states that within the proportional limit of any material, deflection is proportional to the load. This means that if stress under any load within the proportional limit is divided by the corresponding strain under that load, the result will always be a constant value. This is indepen- dent of hardness or any other property of the material except the proportional limit In helical compression and extension springs, the wire is stressed in torsion when the spring is loaded. The basic design formulae for round wire compression and exten- sion springs are shown below: (Force N) ar spring rate N/mm) (Basie Stress MPa) ae 8DPK, ae Se (Corrected Stress MPa) By making suitable substitutions, the formulae may be combined or transposed to show other useful relation- ships. For example, the stress may be expressed in terms of deflection, rate, anid volume of active spring material. For springs made of steel wire (with G = 79.3 x 10° MPa) s = aser yf 2. Design Stresses (MPa) When a helical compression spring is deflected by axial ‘wading, the spring wire is twisted, essentially as a straight bar is twisted under torsion. The torque moment which 259 Design of Helical Springs produces the twist is the product of two factors: one is the load acting along the helical spring; the other is the ‘mean coil radius, or the distance from the spring axis to the center of the coiled wire. In the straight torsion bar spring of circular cross sec- tion, the twisting produces a shear stress which is uni: form at every point of the bar surface; but in the helical spring coiled from round wire, the stress pattern on the wire surface is more complex. There are two reasons for this occurrence. The torque moment results in a steeper twist angle for the short wire fibers at the inside of the coil than for the long wire fibers at the outside of the coil, and, consequently, it produces a higher shear stress at the inside of the coil. The axial load causes a direct shear stress which adds to the shear stress from the torque ‘moment at the inside of the coil, but subtracts from at the outside of the coil Under static loading conditions, the variations in stress over the cross section of the spring wire can be neglected. The standard stress formula can be used, as it will fur- nish an average stress over the entire surface of the wire. However, when the spring is subjected to fatigue load- ing, the standard formula is inadequate because it con- veys neither the higher local stress nor the wider stress range prevailing during each cycle at the inside of the coil. Fatigue failures are induced by a combination of high stress level and wide stress range, and it is a common experience that fatigue failures in helical springs occur at the inside of the coil. A failure at some other point can usually be traced to a local flaw in the wire surface or to severe coiling or presetting stresses. The maximum design stresses allowed for loading with, various spring materials which are commercially availa ble for helical springs can be established as a percentage of the minimum tensile strength of the material. Com- ‘mercial springs made with a moderate amount of preset- ting will give a reasonable factor of safety when stressed to the uncorrected stress levels given in Table 5.1. It has been found convenient to take into account the stress conditions at the inside of the coil by using a “cor- rected stress.”” This is obtained by computing the stress from the standard formula and then multiplying the result with a stress correction factor KW. The factor increases with greater curvature of the coiled wire. Greater curva- ture is equivalent to a smaller spring index, which is the ratio of mean coil diameter to wire diameter. The Wahl factor K, is most commonly used, and its values may be taken from Table 5.2.TABLE 5.1— ALLOWABLE UNCORRECTED TORSIONAL. STRESS AS A PERCENTAGE OF TENSILE STRENGTH FOR VARIOUS SPRING MATERIALS UNDEA STATIC LOADING. ‘SAE No. ‘Type of Wire 3 of Tensile Strength ‘S132/N157 Ol tempered alloy see! 50-55 36 Oil tempered carton steel 45.50 78 Music 40-48 ana Hara erawn carbon soe 42.48 30002 Stainless sleet @ Nonterrous metals 2 ‘When the design stress is calculated by including the Wahl stress correction factor, the stress values can be increased proportionally providing the spring index is between 4 and 9, For springs with smaller or larger indexes, the maximum design stress should be kept lower than that previously outlined ‘The recommended apparent stresses for compression springs may exceed the torsional elastic limits normally found in the wire. Nevertheless, they are applicable since, by suitably presetting, the spring manufacturer can induce offsetting residual stresses. Presetting is a beneficial oper- ation which will minimize the load loss in operation. A detailed discussion on presetting is given in Chapter 1 Values 20% lower than those previously given should bbe employed when designing a spring which is not preset. While the 20% lower values for nonpreset springs may seera to be unattractive to the designer, it should be real- ized that many springs are so designed because they can be produced at a much lower cost and they may perform fully as well as the preset springs. In general, helical compression springs should be designed so that the stress at solid compression is within the limits previously outlined. Any compression which falls short of deflecting the spring to its solid length will ‘not produce permanent set and therefore will not result in loss of load. Springs designed for long endurance life, such as valve springs, or for operation under elevated temperatures, require special considerations with the use Of design tables and diagrams given in this chapter, and as explained earlier in Chapter 1 3. Fatigue Life For those springs which are cyclically loaded, it is com- ‘mon practice to obtain basic mechanical properties from S-n fatigue data and from fatigue strength diagrams, as ‘explained by Figs. 1.6 and 1.7 in Chapter 1 ‘Three fatigue strength diagrams for typical round wire spring design applications are presented in Figs. 5.1, $.2, and 5.3 to aid in the actual design work in establishing the life cycle expectancy. These diagrams cover various spring materials having superior surface qualities, such as valve spring quality wires, and include the variables of presetting and shot peening. The Wahl correction fac- tor is used for all stress values referred to. ‘The allowable initial and maximum stresses are deter- mined by multiplying the appropriate stress factor Ks with the minimum tensile strength TS of the wire or bar used (o make the spring. The initial stress factor Ky is read from the 45 deg reference line, while the maximum stress factor K.; is read from the appropriate higher iagonel line depending on the life cycles desiced, The stress range is “established by the difference in the ordinates of the two diagonal lines. The fatigue strength diagram is then used cither to establish the permissible ‘maximum stress and stress range which will provide the desired life cycle expectancy, or else, to establish the life cycle expectancy for the maximum stress and stress range which must be considered for the spring application in question. TABLE 6.2—TABULATED VALUES FOR WAHL STRESS CORRECTION FACTOR K, ko = (+ S4] oo 00 on 02 03 04 os 06 o7 08 08 y 2 2058 1975 1.9051 8a 1.792 1746 1.708 1696 1.607 3 158015561833 1912 1493 1.476 1459 14301416 4 1404 139201381 13701960 1.381 1.342 13251318 5 1a 5.304 1297 1290 1284 2m 4273, 126201287 6 4259 1248 1243 1239 12951231 1207 12200 1216 r 1213 ¥210 12081203. 1.200 1197 1195 1189 1187 8 wee tea t79 77 i542 1.170 416501164 9 4162 1.160, 1158 11561155 1153 1151 1148 1146 10 513 tng 1140 1199 1138 1136 11931192 " 1131 900128127 4,126 1125 tiea waz 1420 2 arg, ate 1116 ris mie 13 ngs 13 WHO 4408 1.408.407 1406 1.108 1.105, 11031102 4 11021101 1100 099 1098 1.058 1097 10961085 8 10951098 092 093 1092 oat 1091 1.080 11088 6 088 1.088 087 1087 1086 1088 1.085, 084 1.086 7 10831089 1082 1082 10st 1081 1.080 1078 1079 8 10781078 10771077 107710761076 1075 tors 26008 SOLID STRESS 1 Ky} MAX STRESS AT STATIC SERVICE STRESS STRENGTH, ¢ s © & MAX. STRESS FACTOR Ks2 (MAX SH8ae ‘Go’ CYCLES WHEN SHOT PEENED (SOLID LINES) | 0.3 }—+ | | | \— CYCLES WHEN NOT SHOT PEENED (DASHED LINES) | 02 +t ee +7 v li By o1 a’ > - = | fo | 0 0 0.1 02 03 04 os 06 07 08 INITIAL STRESS FACTOR Ks1 (NZL SHEAR STRESS, MIN, TENSILE STRENGTH, Fig, 6.1 - Fatigue strength diagram for round wie helical compression springs which are not preset. All stresses are Wah! corrected. Diagram applicable to springs which are not preset and to the following materials: Music Stool Spring Wire and Springs-—SAE 417. Hara Drawn Carbon Steel Valve Spring Qualty Wire and Springs--SAE J172; Ol Tempored Carbon Steel Valve Spring Quality Wire ‘and Springs—SAE J351; Oi Tempered Crvomium- Vanadium Valve Spring Quality Wire and Springs—SAE J432 _ The formulae used as the basis of the fatigue strength 4, Examples of Spring Design Problems diagrams are presented in Chapter 1. When performing spring design calculations with the use of computers, these formulae may be used directly instead of the diagrams. Tables 5.3, 5.4, and 5.5 give the essential parameters Fequired for use with the formulae. Example 4.1: For a spring with a 28 mm inside diameter, using a 4.5 mm diameter hard drawn wire, determine the stress at a load of $30 N. D = 28 + 450 = 325mm SDP _ & x 32.5 x $30 = 481 MPa ad mx 4.509 Ps 261 MmTABLE § 3-PARAMETERS FOR COLD WOUND SPRINGS FOR FATIGUE STRENGTH OF FIGURE: 5.1-NOT PRESET Not Shot Peened Ke = 5600 & Cycles MaxKe Ke 108 300 326 108 420 260 107 480 243 ;_ = 8545 _M = 0.0000 0.6620, Ce A 8 Ke a aa) aya ors rae asi tres 303053603827 a | ge a 5B 3 Z 4 4 j BRS | th | iz oH \ ‘= CYCLES WHEN NOT SHOT PEENED (DASHED LINES) . —* w a | | a \E eyeues WHEN SHOT PEENED (SOLID LINES) | 5 | 2 0.2 + t — - ——| | 0.1 + 0 i | 0 01 02 03 04 0s 06 07 08 / INTIAL SHEAR STRESS INITIAL STRESS FACTOR Ksi (in their acho Fig. 5.2—Fatigue strength diagram for round wire helical compression springs which are preset. Diagram applicable to moderately preset springs and to materials mentioned In Fig. 9.1 262‘TABLE 5.4—PARAMETERS FOR COLD WOUND SPRINGS FOR FATIGUE STRENGTH DIAGRAM (OF FIGURE 5.2—PRESET Ky = 07000 G, = 7787 _M = 0189 Not Shot Peoned ‘Shot Peoned 0, 05758, Y'= 00557 Og = 0610, ¥ = 0.0500, ~ Gye Moxka KAKO 8 1 281 S10) eyer 450 peua0s eur 72 yeuaasolmerr one) io 640 ae a8 agra? 4732 or 620 2a 560486 8K3 MG i 5 Is 2 3 7 3 x Ye 7 8 | Q z UFECYCLES | | f _ | g os + : EXAMPLE 6.1—/ | & | 5 | z 02 : _ . O1 —+ | ~ | | | | 0 ° 01 02 0.3 04 0s 06 o7 0.8 INITIAL STRESS FACTOR Ks1 (ing nse swe Fig. 5:3 Fatigue strength di ‘epresenis B-10 fatigue lf, and i applicable to ot coed helical compression springs which are shot posned, preset, bon steel or alloy ste! bars IN. TENSILE STRENGTH, 1m for hot called round bar helical compression springs. All stresses are Wahi corrected, diagram nd using car 263TABLE 5.5—PARAMETERS FOR HOT COILED SPRINGS FOR FATIGUE STRENGTH OIAGRAM (OF FIG. 53-SHOT PEENED AND PRESET. Ky = 0780 C= 1808 C, = 0830 M= 00015 Y= 0.100 Gyclesn __MaxKee Ke A 8 joo ovis —~07a7 0798 0.008 10000 os1 OMG 0.828 (OLS 100000 0.845005 0.823 0.289 s000000 06170300 aaa 0.423 2500000 0.605 0.265 0.992 0.471 From Table 5.1, the allowable uncorrected stress for __Thus, if this spring were to be made with closed and SAE J113 hard drawn carbon steel wire is 42% of tensile" ground ends, the total number of coils would be 8.6, with loading. Since the tensile strength of 4.50 mm wire is 1380 one inactive coil at each end for closing. MPa min (see Chapter 2, Table 2.2), the allowable stress is 1380 x 0.42 = 580 MPa. Therefore, the stress of 481 is it ‘i MPa is satisfactory for static loading. 5. Design Stresses for Hot Coiled Springs Example 4,2: Using spring dimensions similar to those ; . in Example 4.1, determine the maximum load obtaina- High stresses can be used in design of hot coiled heli ble using an oil tempered carbon valve spring wire to _ Cal compression springs because presetting is employed attain unlimited life when the initial spring load is 270 and the heat treatment and shot peening processes pro- N and the springs are shot peened and preset, Vide the required control over the physical properties. Tensile strength of 4.50 mm diameter oll tempered car- Based on the material used and its heat treated hardness ‘bon valve spring wire is 1550 MPa min for SAE 351 (eee aid tensile strength, the spring can withstand a certain Chapter 2, Table 2.8), ‘maximum stress for a particular stress range and for an estimated number of fatigue cycles. Load loss or spring c =a BSL ay set, within permissible limits after a number of cycles of e” 45 loading is another factor to consides. High cyclic stresses ‘can result in appreciable spring set, so that unsatisfac- Se eee enced) tory spring performance may result even though the . DKwP _ 8 x 325 x 1.206 x 270 spring may not break in service. cl a x 450° The stress limits for hot coiled springs can be deter- ‘mined from the fatigue strength diagram given in Fi = 296MPa 5,3. This diagram is applicable to springs which are axi- 296 ally loaded without angularly aligned or offset spring Koi = fpgp 7 0191 seats. When the spring is subjected to excessive bow due to non-aligned spring seats, then the stresses should be Referring to Figure 5.2, for unlimited life when springs decreased proportionally. This diagram is applicable to fare shot peened and present, or using the formulae in springs which are preset and shot pened. All stress values ‘Chapter I together with the parameters in Table 5.4: are Wahl corrected. The diagram represents B-10 fatigue life or the number of cycles where 10% of the popula- Ko = 460 + 3429 Ks, = .460 + 343 x .191 = -526 tion is estimated to fail (see Section 7). The allowable intial and maximum stresses are deter- Since P, = Ks, therefore P; = 270 x .526 = 885N mined by multiplying the appropriate stress factors (Ku, Pho Ka 191 and Ka) by the minimum tensile strength for the bar Example 4.3: Using the data of Example 4.2, deter. used to make the spring. Basic mechanical properties for imine the numberof active cols required to give a Toad Lesa Se ane ee OE ooeee s explained in Chapter 1, and also in Section 3 of this ee ee ee chapter where application of the fatigue strength rams is explained. Fig, 5.3 is applicable to hot coiled springs and, in addition to stress factors Ky, and K,, it gives the maximum stress factor Kz for solid stress, preset stress, and for maximum stress at static service. Frequently, the solid stress may be the same as the preset stress, in which case the spring is compressed solid to per- form the presetting operation. In cases where the design 264solid stress is in excess of 1150 MPa, the presetting oper~ ation can be performed by presetting to a fixed spring length. 6. Fatigue Life of Hot Coiled Springs Fatigue testing is an accelerated method of examining springs for design adequacy and for quality control pur- poses. Fatigue life is expressed by the number of deflec- tion cycles a spring will withstand without failure. It can be estimated by the use of the fatigue strength diagram shown in Fig. 5.3. Re-examination of the design will be in order if the fatigue tests result in failures which are confined to one section of the spring ‘The procedure to establish suitable parameters for a fatigue test frequently requires the exercise of judgment when the spring application is known to involve random stress amplitudes. As an example, a suspension spring will undergo large number of cycles of small amplitude near the design load position without failure. Under greater amplitudes, the number of cycles without failure will be reduced, since the maximum stress as well asthe stress range are increased, and both are determining factors in the fatigue life of a spring. ‘The metal-to-metal position (vertical load limit) is fre- quently used as ¢ maximum deflection position of the spring in fatigue test; but in heavy truck spring designs this deflection is often considered excessive for the test, setup, as it is rarely reached in actual service. Then the 1um deflection position in the fatigue test may be ‘established at compressing the rubber bumper to between one-third and two-thirds of its free height during full ounce ‘The length of the test stroke is selected from experience. AA frequently used method of establishing the length of test stroke is to add to the compression stroke (from design load to load at maximum deflection position) one half ofthis deflection for the release stroke (from design to initial load). This practice may require modification in those cases where it would produce less then 100,000 ‘or much more than 250,000 cycles, according to Fig. 5.3. ‘There ae several considerations as to why the test setup should be preferred to be such that it will result in a Dreponderance of cycle figures which are nether 100 short ‘or too long. At higher stresses (shorter lives), the scat- ter ofthe cycle lives is theoretically reduced so that fewer ‘est samples will produce a given degree of precision in the estimated life of the entire population. However, lower stresses (longer lives) give more realistic results, since they duplicate more nearly the actual service con ditions and prevent spring settling during the test. Also, Comparisons between different groups of springs will be “more distinct at lower stresses, since different S-n curves tend to diverge the more they approach the fatigue limit DO (oF limited value of stress at which 50% of the popula- tion would survive a very large number of cycles, say n= 25 x 109, In order to establish the fatigue life cycles, which are acceptable in any spring design, itis desirable to have road, durability tests run over a prescribed course so that, fatigue life test data and actual road durability may be correlated. Example 6.1: To estimate the expected life of a hot coiled, shot peened, and preset spring which is to meet, the following specifications: End configurations: Taper rolled plus tangent tail Inside diameter: D, = 90.0 mm Installed load and P, = 6600 N at length: 260 mm, Rate: R= S1.0N/mm Metal-to-metal d= 45mm clearan Metal-to-metal limit $;_ = 1000 MPa stress: Min tensile strength of TS bar: Test stroke (compressed and released from installed): 570 MPa compressed = 60.0 mm released = 30.0 mm ‘The computations are as follows: Bar diameter selected: d= 15.00 mm Gat _ 76.0 x 10" x 15.00 "DPR 8 x 1050 x 51.0 Kw (for © = 105.0/15.00 = 7.00) = 1.213 8DKw _ 8 x 1050 x 1.213 oe 7 X 15.00) 101 % 15.00 x (8.15 + 1.67) + 15.0/3 = 153.8mm Ly + (PUR) = 260 + (6600/51.0) = 389.4 815 S/P 0.0961 L L Ps =P +R — Le) 12020 N Py = 6600 ~ 51.0 x 30.0 = 5070 Nand Pz = 6600 + 51.0 x 60.0 = 9660 N length toad stress. factor Tabulation t P Ss % Positions am N MPa Installed—! 2600 6600 ge O.404 Roleased—1 2000 50704870310 Compressed—2 200.0 9650928. 8at ‘Solid 1538 1202011850736 2.65 6600 + 51.0 (260 - 153.8)The factors (Ks) and Ky) may be used with Fig. 5.3 Since Kx to estimate the life cycles n to be approximately 2.5 x 10%; or the formulae in Chapter 1 can be used with the parameters in Table 5.5 to calculate directly the life cycles Y as follows: BY, (Ce/Key-¥ = (1-808/0.3591)!2"" = 250,900 cycles 7. Evaluation of Fatigue Test Results Ky = —Ku(Kse = Ks) eee It must be understood that the fatigue life cycles for aoe any group of springs will vary considerably even under 2 X 0.740 - (0.591 + 0.310) = 0.3591 closely controlled test conditions. Moreover, the average ‘TABLE 5.6 MEDIAN RANKS (PERCENT) FOR SAMPLE SIZES 1—20 II Sanne So 1 Rack Rank | Hg pp oe a | : ed oan Seen) ee eee eee aie eel ee ; RB RE om Bi GM Be ge ees i : ie BB a ee ee ee {i 3 e705s 72885 63528 © $5984 © 50000 45.1609 $ moo mist eras som) tat I~ 7 90572 79887 71.378 64.490 7 | ; foe gece ae | Se as | 10 93.303 10 : Rank ‘Rank | Son ek oe ow | 2 A $88 SR tel tee gee tee aime aa ue i 3 BE NSS ok ae HR ee BER ts! ae Seg i 2 2S Se eee Be eS ee ee ee ae gE | 3 ORME GES SSE SES Bet Ses cus met mee ems i § somo 5551250 goste caer atoms sam ant mace am \ § On SE gem SNE may ae im. mee aa ae i GS Bie pe gee a ge Bel ea ae! eet \ 3 Mer Se Eat Bae fee ee ee ee eee ee | o Gs RS BS SS Se; 80 Se Le as see esas going ose Taso sts atom clan seu? atmo seme 4 S7 BE Net wee eee oe a gm Be g Sis HE GE na, ee Bee eee ee a BS GM GH aoe Gee sm ea 8 ‘ ea Se Gn ae RE na Of * csieo wae es4i9 tron mae ‘ Be gig gen nee iH See ane aes 8 aie eee 8 SB 8 one 266TABLE 5.6 (CONT) ‘Sample Size Rank Rank Onder at 2 2 2 25 2% ar 2a 2 30 Order 1 3247 9301-2968 2847 2734 Bast asak 2a 2962 2A 2 764 7512 «71816895 BGS G72 G19 | 502257205122 3 12531-11970 1145810987 10.553 «10.159 97B1 9469.1 Bala 4 17209 15.734 15734 1506314492 ©—«19:942«13492 «12.858 «NSIT Nod 5 21890 20015 20015 19.192 18.435 «17.795 «17086 «1648315922 15907 6 28574 25.384 24297 29.209 2297921829 ara? © 20.010 19.928 «wea 7 31258 29.59 28580 27.406 26.324 «25.325 24.308 25597 22.795 21988 T B 9594334334 2863-31513 30269 29.120 28055 27.085 © DHT O281 9 40.629 38810 7.147 35621 34215 92.916 BNTI2 4089 29550 2aS7G |G 1 45.314 43.286 © A1431 39.729 -$B.161 36.712 95.970 34121 32GB BBT2 10 11 $0000 47,762 45.716 43.897 42.107 40.509 39027 97.650 96967 35.168 11 12 54.606 $2238 $0,000 47.985 45.054 44.305 42685 41.178 «30.775 BRMBA 12 13 59371 58714 54284 52.054 $0,000 48.102 46242 44707 43183 «ATTEO 13 14 64057 61.190 5856B 56.162 58.946 51.808 $0.00 48.296 465024505614 15 68.742 65.665 6285360271 $7892 55.695 53658 51.764 50.000 4895215 16 73426 70.141 «67.137 64979-1899 $9491 S7.915 55203 S348 Siete 16 ‘7 78.109 74.616 71420 ga487 © G5.785 63.287 60.973 58821 56.817 S4O8k IT 18 8279) 79.c83. 75709 72594. 69.730 7084 - 64.650 62950 60.225 $8.24 > 18 19 87.469 83561 79985 76701 73676 70880 68.289 «65878 636336159619 2 «92136. 88030 84.266 © 80808. 77.821 TAG7S «71.945 69.407 «TOI. BABGE 20 21 96753 92488 eBSe2 84912 8158578471 75.602 72938 70450 68.128 at 2 96898 92809 89.013 85507 82.265 79.258 76463 «73.857 T1424 2 2 97.031 99.105 89.447 86.088 BZ9Td © 79.990 77265 «T4719 «23 2 97.183 93377 89847 86.568 I5I7BDG72 TBO 2s 25 97265 93.628 90219 7.042 84078810925 26 97.363 93.861 90564 87.489 460226 a 97465 94.078 © 90885 8789527 Ba 97555 94.280 9118628 a 9768 ants 29 30 e776 30 life of the tested springs is not sufficient by itself tostab- _the sample of 99% were considered representative of the lish a judgment on the design, the material, or the produc- ton method which the springs represent. The relationship between the number of applied cycles and the percent- age of springs which failed at these cycles can best be ana- lyzed with the help of statistical techniques which will systematically deseribe the ‘dispersion’ or “‘spread”’ or “scatter” of the recorded test results. The extent of the scatter will depend upon the consistency of surface con- dition, fabrication, and the general quality of the springs which are tested. Sampling One of the main purposes of statistical analysis is to draw inferences about the properties of a large group (the ““population”) from the results of tests on @ small group (the “sample”. Ifthe entire population were tested, one ‘would not have to infer anything; one would know how the population reacted to the test. This would be called 100% confidence. If 99% of the population were tested, one would be 100% confident about that 999%. Also, if remaining 1%, one would be close to 100% confident of predicting the result if that 1% were tested. If only 58% are tested, one would know about that $M, but how much could one infer about the remaining 95%? Actually, if certain conditions of sampling are met, one can infer a sreat deal about the entire population from tests on small samples. What is required is a good, honest sample. The primary condition for a good sample is that it be taken at random under conditions which ensure that all springs of the population have an equal chance of being chosen. This is obviously impossible in the case where the sample consists of a few handmade springs of a design which has not yet gone into production. Only experience ‘can tell the engineer whether the various properties of the sample which can affect the test result (in regard to mate- rial as well as to production methods and controls) will also be present in the production springs (the “popula- tion"). This determination is outside the realm of statis- tics. However, statistical mathematics are based on the inherent assumption that the sample isa true representa- tive of the population. 267Distribution If the entire population were tested under identical test conditions, the results could be shown in graphical form, by arranging them in ascending numerical order and plot- ting the cumulative fraction (or percent) of failures over, an abscissa of “life cycles.”” A sample selected at ran- dom from this population can be expected to exhibit a similar distribution of fatigue life; the larger the number Of springs in the sample, the closer will be the similarity. It is possible to calculate the likelihood of similarity for samples of any given size. Tables are available based, on such likelihoods; Table 5.6 is an example. It presents the “median rank” of each test result for a sample size between I and 30. A rank is assigned to each individual test result corresponding to that portion of the popula- tion which itis most likely to represent. The median rank is used as an estimate of the true rank because it is just as|likely to be high as low. Table 5.6 lists percent figures. ‘A good approximation formula for the median rank (which may be used for larger sample sizes than those in Table 5.6) is: 100 x ene N+04 where: J = Position (in ascending order) for each test result in the sample N =Total quantity of springs in the sample ‘The median rank line constructed from such data predicts that certain percentages of the population will survive specific cycles-to-failure. But any such estimate ‘may err substantially on either the high or low side. The question is: How confident can the engineer be of such an estimate? When he has accumulated a great deal of experience in comparing the results of small samples which represent springs of different materials or differ- ent designs or different production methods, he may Judge, after contemplating two such median rank lines, that he should give preference to one set of springs over, the other because of its apparent superiority in fatigue life. However, when judgment based on experience is not considered adequate for a final decision, then it will be necessary to construct lines of higher confidence. In many cases 2 confidence level of 90 or 95 or even ‘999% will be required, so that there will remain only a 10 or 5 or 1% risk of the estimate being either too high or too low. For'a chosen confidence level, the life cycles of a given percentage of the population will be found within a certain “tolerance interval."” On the median rank ‘graph this may be represented by a “tolerance band” to either side of the median rank line. Wider bands indi- cate increased doubt about the line truly representing the population. With a given sample size, the bands will be wider for higher confidence levels. With a given confi- dence level, the bands will be wider for a smaller sample If a 90% confidence level has been chosen, the lower and upper limits to the estimated fatigue life distribution of the population can be shown by constructing *5t% rank” and ‘95t rank” lines (see “Theory and Tech- nique of Variation Research” by Leonard G. Johnson. Elsevier Publishing Co., 1964). The numbers 5% and ‘95% represent the chance of being either {00 high oF too ow in assigning the given ranks to the individual spr in the sample, and they are used to establish the limit within which the true population is expected to lie. For a concise summary of distribution mathematics see “Engineering Considerations of Stress, Strain, and Strength," by Robert C. Juvinall, McGraw-Hill Book Co., 1967. Weibull Plot Several systems of mathematically organizing the test result data have been established. In the past, the nor- ‘mal (or Gaussian) distribution has been most widely used. It is graphically represented by the familiar symmetrical beil-shaped distribution curve which is completely defined by two statistical parameters 1. The mieun life — In the test sample itis the sum of all the recorded test result values, divided by the sample size. Itthen becomes an estimate of the population mean fe. A population with normal distribution has the mean coinciding with the median (which is the middle result when all individual results are arranged in order of mag- nitude), and also coinciding with the mode (which is the cycle value at which the greatest number of failures occurs). 2. The standard deviation — It describes the scatter on either side of the mean. For the test sample itis mathe- ‘matically defined as the square root of the sample vari- ance (which in turn is the sum of the squares of the difference between each recorded test result and the mean, divided by the sample size minus 1). It then becomes an estimate of the population standard deviation. ‘When the test results are arranged in ascending numer- ical order, and the cumulative percent of failures i plot- ted (using median ranks) over an abscissa of life cycles on normal probability graph paper, it will be found that a straight line can be fitted to the results as long as the distribution is normal. This becomes an estimate of the population distribut ‘While the normal distribution has a number of attrac tive attributes and has been the subject of many publica tions, it must be recognized that in spring fatigue testing the results are usually not normal in that they cannot be 2.68plotted in the symmetrical bell-shaped distribution curve but in a skewed curve. This has led to other mathemati- cal formulations. In the automotive industry the Weibull ‘plot i used because it permits straight-line plotting ofthe cumulative failure probability versus life cycles on Weibull probability graph paper, even when the distr- bution is skewed. In the Weibull distribution the relationship between the number of applied cycles and the cumulative percent of failures at these cycles is expressed by a formula which uses three parameters: 1. The minimum life, which may or may not be 2er0. It is denoted by the letter “a.” Its generally assumed that a = zero because that is the condition for which the Weibull formula assures straight-line plotting on Weibull paper. 2. The Weibull slope, which is an indicator of the skewness of the distribution. It is called the “'shape parameter” and is denoted by the letter “*b.”” It also is a measure of the scatter of the distribution; a low slope value indicates a high degree of scatter, and vice versa. The slope is the tangent of the angle formed by the dis- tribution line with the abscissa on Weibull probability paper, when the scales are such that the distance representing the factor 100 on the (logarithmic) abscissa scale for life cycles equals the distance from 2.3 to 90.0% ‘onthe (log-log) ordinate scale for percentages of failure. In the Weibull distribution, the mean, the median, and the mode never coincide exactly. But when the Weibuil slope is within the range of 3.2-3.5, the differences are small enough to give the Weibull distribution an appear- ance of symmetry. The Weibull comes nearest to the nor- mal distribution when the Weibull slope equals 3.44 (thus representing an angle of 73.8 deg), because there the mean and the median have identical values ‘The more the slope value increases above 3.44, the ‘more the distribution curve will be skewed to the left (with a long tail to the left), where the mean is to the left of (or less than) the median, The more the slope value decreases below 3.44, the more the distribution curve will bbe skewed to the right (with a long tal to the right), where the mean is to the right of (or more than) the median. Fig. 5.4 shows a graph (for Weibull slopes I through 12) which locates the percent of failed springs at the life eycles representing the mean of the population. 3. The characteristic life, which is the 63.2% failure point for the population. It is called the “scale parame- ter” and is denoted by the Greek letter theta ( 6) 63.2 = 100 (1-1/e) where: ¢ = 2.7183 (the Napierian base) Example 7.1 (Fig. 5.5): Eight springs have been fatigue tested under identical conditions. The results are arranged in ascending order of failure cycles and are given rank 2.69 ‘order numbers accordingly. In this order they are assigned median ranks from Table 5.6 as follows: Orger ycles to Median ‘Spring No. Failure Rank € 1 1 000 30 a 2 931 000, 20.11 F a 114 000, 3205, H 4 195 000, 4402 & 5 135 000, 55,98 8 6 177 000, 6795 6 7 205 000 79.89 D 8 245 000 9170 ‘These points are plotted on Weibull graph paper in Fig, 5.5. Drawing a straight line of best fit through the ‘median rank points produces an estimate for the failure rate of the entire population with the parameters b = 2.4 and @ = 170 000. From Fig. 5.4 it willbe seen that for b = 2.4, the per- cent of failed springs at the mean is 52.7. At that failure level the median rank line in Fig. 5.5 shows 150.000 cycles. AC the B-10 life level (that is, at the number of cycles ‘where 10% of the population are estimated to fail), the ‘median rank line shows 66 000 cycles. Significant Difterences In most cases the fatigue testing of springs will be undertaken for the purpose of comparing different sam- ples, and the probability graph will be expected (0 con- vey information on the relative life distribution of the populations represented by those samples. For example, ‘the comparison may involve a sample representing a first design, and another sample representing a second design, ‘When the two median rank lines for the test data of the two samples are plotted on the same graph, they will, readily show if the second design promises some improve- ‘ment in fatigue life. However, in order to establish if there is a “significant difference” between the two designs, it will be necessary to find quantitative values for the degree of improvement (or degradation) between one design and, the other. The questions is this: How confidently can one say the limited test results indicate that the second design, assures an improvement in fatigue life for the entire spring, population? The answer depends not only on the amount of separation between the two plotted slopes, but also on the size of the two test samples. Furthermore, the degree of confidence in the super ority of one design over the other need not be constant {from one quantile level to another. For example, itis pos- sible to have a significant improvement at the B-S0 life (50% failure level) without any improvement at the B-10, life (10% failure level), or vice versa. This is partly due to differences in the Weibull stope, and partly due to thez NT Gedy ‘WEAN Fig. 54—Location of the ‘greater width of the tolerance bands at the lower quan- tile levels. Example 7.2 (Fig. 5.5): It has been proposed that the spring design represented by the sample of eight in Exam- ple 7.1 be replaced by a new design. Seven springs of the new design have been fatigue tested with results which, are shown arranged in ascending order and are assigned median ranks from Table 5.6 as follows: Od ‘ycies to Median Spring No. Failure Rank ™ 1 132 000 9.43 ° 2 195 000, 2288 kK a 233 000 3641 t 4 278 000 50.00 p 5 318 000 6359 J 6 5365 000 738 N 1 +440 000 9057 ‘mean for Weibull slopes 1-12 For this second plot it will be seen from Fig. 5.5 that the Weibull slope b = 2.7 and the characteristic life # 310 000; the mean life level is at 51.84% (see Fig. 5.4), therefore, the estimated mean life is 280 000 cycles. The estimated B-10 life is 138 000 cycles. ‘Since the estimated mean life in Example 7.1 was 150 000 cycles, the mean life ratio on the median rank lines is 280 000/150 000 = 1.87. This represents an esti- ‘mated improvement of 87%. The “‘confidence number” corresponding to this mean life ratio (that is, the probs bility that the true mean life ratio of the population greater than 1) is found by reference to the mean life nomograph (Fig. 5.6). 270EXAMPLE 7.1 [EXAMPLE 72 = 170 000: =EMEAN = 150 0005 30 40 50 60 70 8090 100 1000 CYCLES Fig. 55—Megian rank lines 200° 300400 500 600 800 1000 an3 ett oka WEIBULL SLOPE, b beet b+ 1 3. | 7-99.95 1a 799.5 99 foe { 13-- 95, + 14+. S00 15> 1.6 18 +60 20-- 8 g wu » 277 = 4 8 q ol 5: g x 7 3 z : 2 10 o 2 + 60 Connect total degrees of freedom with Weibull slope and locate intersection point on auxiliary line. Connect life ratic with intersection point and continue to intercept on confidence number. For unequal Weibull stopes perform operation for each ‘slope and average the confidence numbers so obtained. Fig. 5.6—Confidence nomograph at mean ie level 272 MEAN LIFE RATIO DEGREES OF FREEDOMWEIBULL SLOPE, b 15 +10 12 99.95 13 99.6 14 dies 15: 16. 18; 95 2 2 as = = = gon 7 2 a z S a a 2 3 z a 8 5 7 z 8 5770 x 4 g x 10) 8 z & a # 60 55 1. Connect total degrees of freedom wi Weibull slope and locate intersection point on auxiliary line. 2. Connect life ratio with intersection point and continue to intercept on confidence number. 3. For unequal Weibull slopes perform operation for each slope and average the confidence numbers so obtained. Fig. 6.7—Confidence nomagraph at 8-10 lite level 273 104, 103, DEGREES OF FREEDOMIn order to work with the nomograph it is necessary to establish the “degree of freedom” (= freedom of movement of the individual test results about a fixed ‘mean) for the samples representing the two designs. The degree of freedom equals (N - 1) where N is the size of the sample. Thus N,- 1 = 7 and N,~1 = 6, and the Product of the two, known as “total degrees of free- dom," is 42. ‘The nomograph furnishes the confidence number 98.8, for b = 2.4 (first design) and 99.4 for b = 2.7 (second, design). The average is 99.1, and this means that 99.1 times out of 100 the second design is superior to the first, design at the mean life level. ‘The confidence number corresponding to the B-10 life ratio will be found by reference to the B-10 level nomo- graph (Fig. 5.7). Since the estimated B-10 life in Exam- ple 7.1 was 66 000 cycles, the B-10 life ratio on the median, rank line is 138 000/ 66 000 = 2.10. The nomograph fur- -nishes the confidence number 91.0 for the first design and, 93.0 for the second design. The average is 92.0, so the second design is superior at the B-10 level 92.0 times out of 100. Thus, the confidence numbers obtained from the fore- going “significant difference” study indicate a certain superiority of the second design over the first. Quantita- tive values for the degree ofthis superiority are obtained by using the information from the nomographs on so- called ‘‘confidence interpolation graph papet”” or sim- ply “ratio paper’” which has a log-log ordinate for per- cent confidence and an arithmetical abscissa for life ratio Gig. 5.8). To obtain quantitative values at the mean life level, the mean life ratio on the median rank lines (in this case 1.87) is plotted on the ratio paper at the S0% confidence level and is connected by a straight line with the confidence number (in this case 99.1) at abscissa 1. The life ratio vvalues at other confidence levels are then found on this line. It will be seen that at the 60% confidence level the ratio is 1.75, so there are 6 out of 10 chances that 75% improvement will occur. At 90% confidence (which is fre- ‘quently used as a standard) the ratio is 1.32, so there are 9 out of 10 chances that 32% improvement will be realized, ‘The same procedure will establish quantitative values at the B-10 level. There the percent of improvement will always be comparatively lower for the same degree of confidence. This is due to the greater width of the toler- ance bands at the lower quantile levels. In this example, the life ratio at 90% confidence is 1.08, so there are 9 ‘Out of 10 chances that 8% improvement will be realized ‘Minimum Life Greater than Zero As stated earlier, Weibull plots are generally con- structed with the assumption that the minimum life is 274 zero. The sample data will then plot a straight line on Weibull probability graph paper. ‘When the minimum life is greater than zero, the sam- ple points can usually be fitted with a fairly smooth curved line, However, this would mean foregoing one of the major advantages of the Weibull process, which is to analyze the data with the help of a straight line even when the number of test results is small A relatively simple technique permits the test results, which indicate a minimum life greater than zero, to be converted to straight-line plotting on the Weibull graph, This requires that the curve drawn through the sample points be extended downward until the abscissa value which it approaches asymptotically can be approximately established as an estimate of the finite minimum life “a.” When this is subtracted from each of the plotted data points, it may be possible to fit the new points thus obtained with a straight line. If it develops that the new points.can still be better. fitted by a curve than by a straight line, then the estimate of the minimum life was incorrect. If the new line is still concave downward, the minimum life was estimated too small; if the new line is concave upward, the minimum life was estimated too large. A second (and possibly a third) estimate will then be required until the plotted points can be successfully fitted with a straight line. I¢ is well (o remember that before any one of the life cycle values found on this straight line is used for comparison with a corresponding value on any other median rank line, it must be increased by the ‘a’? value which was subtracted from the curved median rank line to obtain the straight line. Example 7.3 (Fig. 5.9): Seven springs have been fatigue tested with results listed below, arranged in ascending order and shown with their assigned median ranks from Table 5.6. The plotting of these values produces a curve which is concave downward. A tentative extension indi- cates ‘a’ to approximate 50 000 cycles. When the points are replotted with each result reduced by 50 000, they can only be fitted with a curve which is concave upward; therefore, a second attempt is made with a = 45 000, and this brings about a successful straight line fit. Order — Cycles wo Madian Cycle ess Cycles less No. Palle Rank’ 80.000" 45.000 1 85.000 943-35 000 40 000 2 110.000 2285 60 00065 000 3 135.000 ©3641 85 00090 000 185.000 50.00 105 000110 000 5 180.000 63.59 130 000 135 000 8 210.000 715 180000185 000 7250000 9057 200 000 205 000LIFE RATIO Fig. 5 8—-Contidence interpolation 27520 WEIBULL ‘SLOPE | | | eee | | | uireminuse” || LIFE | (000 evetes) i (1000 cytes) 230 40 60 60 70 6000100 a rr Fig, 5.9 — Weibul plot for springs with minimum ile greater than zero 2768. Rectangular Wire Springs ‘The use of rectangular wire sections (including the spe- cial case of square wire) is not recommended, except in ‘cases where space limitations make it necessary. Since these shapes are not produced in large quantities, they are difficult to procure. In addition, the square edge con- ditions which are prevalent in them tend to limit ductil ity and may thus promote breakage, Wire which is rectangular before coiling will upset at the inside of the coil and become trapezoidal in section after coiling. This limits the deflection per coil, as the solid height is predicated in the wire thickness at the inside diameter. An appropriate formula for the upset thick- ness (t) at the inside, compared to the original thickness (0 of the rectangular wire for (a) of the square wire], is: ©. wnt{t eK Sp where: D, = cutside diameter of spring D, = inside diameter of spring K = 0.3 for spring temper materials 0.4 for annealed materials Wire which is keystone-shaped before coiling over- ‘comes this difficulty but is costly and hard to obtain, par- ticularly in small quantities. ‘The principal object in using rectangular section wire is usually to procure maximum load capacity for the spring in a given space. Consequently, such springs are often highly stressed, The problem of insuring reasona- bly long service life becomes aggravated if the range of stress is high or if elevated temperature conditions pre- vail. In such cases, shot peening becomes a mandatory requirement. ‘The design formulae for compression and extension springs coiled from rectangular wire are shown as Goer DIN, K.Gbe DN, DP K,be The values for the deflection and stress constants are given in Fig. 5.10. They are approximated by the fol- Towing formulae: 0.430C, ~ 0.217 K, = defection constant = 9430Ce — 0217 ‘Cx + 0.200 0.685Cx + 0.355 Kz = stress constant reat S00 where Cy = b’t b= long side of rectangular cross-section wire short side of rectangular cross-section wire ‘The previously cited formulae are uncorrected for stress increase due to curvature. They are suitable for the basis of design for springs which are preset and then used for static loading. Coil curvature in rectangular wire springs alters both rate and maximum stress and needs to be con- sidered where fatigue loading is present, Factors for deter- mining corrected rates and stresses were developed by G. Liesecke and appear in graph form in spring design text- books. The reader is referred to these sources for these finer points in rectangular wire spring design. See book entitled “Mechanical Springs,” by A. M. Wahl. 9. Square Wire Springs Uncorrected stress — The uncorrected stress $ in a ‘square wire spring is obtained by assuming the spring to act essentially as a straight wire under torsion. This gives: 24PD 7 > Deflection — The deflection F for a square wire heli- cal spring is given by: 559 PDN, Gat oo SDN, F Pp Sprit =f. pring rate = = Corrected stress — The corrected stress S, which includes effects of curvature and direct shear and which should be used to calculate stress range for fatigue load- ing, is given by: S. = K's 1+ In this case, the spring index is C = D/a and K* represents the curvature correction factor. amrIEF LECTION CONSTANT. 2 Ganbre Fig. 510 — Constants for eorings of rectangular wire 10. Buckling of Compression Springs A helical compression spring which is made excessively long compared with its diameter can buckle in the same manner that elastic columns buckle under compressive loads. Regardless of the size or stiffness of the spring, the critical deflection at which buckling occurs depends on the free length, the slenderness ratio Ly/D, and the method of constraining the ends. Fig. 5.11 gives curves for the ratio of critical deflection to free length for springs supported with hinged and fixed ends. In springs with hinged ends, the free length must be ‘measured as the distance between hinges when the spring is free, not between the ends of the spring. This applies only when the distance between the hinges is not much different than the distance between the spring ends. If the distance is much greater or much less, the relations do not apply. These data are derived from calculations and are substantiated by experiments on springs with the conventional closed and ground ends and with reasona- ble concentric loading. Eccentric loading and poor end support will cause buckling at slightly lower deflections, Interference between coils will delay buckling and in some ‘cases will suppress it entirely. If a spring is compressed gradually beyond the buck- ling point, lateral deflections will build up while the axial rate decreases rapidly and then becomes negative. Sud- den collapse may not occur at the buckling point, and the spring may be used right up to the buckling limit. Where the slenderness ratio is very large, loading with- ‘out guides may cause complete collapse with the spring flying out transversely to its axis. If buckling cannot be avoided by spring design, provision must be made for proper guiding; care must be taken to prevent binding ‘on the guides. The lateral pressure on the guides increases with lateral deflections of the spring; therefore, the guideCRITICAL BUCKLING DEFLECTION FREE LENGTH RATIO: NOTE: BUCKLING WILL OCCUR IF CRITICAL BUCKLING DEFLECTION IS EXCEEDED, ed se HINGED ENDS FIXED ENDS 4 5 6 7 8 9 10 FREE LENGTH ae MEAN COIL DIAMETER Fig. 5.11 — Buckling erteria for compeession springs 279| clearances should be as small as is compatible with diameter tolerances and eccentricity of springs and spring seats. 11. Diameter Changes ‘The spring designer who must use springs in a confined space should be aware of the natural diameter changes, which occur in springs when they are compressed. These diameter changes are due to two combined effects. One is an increase in the diameter of a helical spring upon, compression when the coils, which were originally inclined, assume a position more nearly at right angles to the spring axis. The second effect is the result of the normal winding or unwinding of the spring during exten- sion or compression. In the case of the conventional round wire helical spring, this effect is one of unwinding during compression, Associated with this unwinding (or slight décrease in the number of coils) is an increase in the spring diameter. For springs using wire of other cross sections, this last effect may be zero or it may even be negative, resulting in a decrease in spring diameter dur- ing compression. ‘With a conventional round wire helical spring under compression, the combined diameter increase coupled. with the normal permissibte dimensional tolerances may be large enough to cause jamming of a spriug required to work within a tube or sleeve. Therefore, itis desirable to have a sufficiently accurate formula to determine the increase of the spring diameter during compression, ‘When spring ends are constrained from unwinding dur- ing compression, the relative increase of diameter is: aD oa = = 00s D7 00 where: AD = increase of mean coil diameter due to compres- sion from free length to solid length ‘When one or both ends of the spring are free to unwind freely without friction, the relative increase of the spring, diameter is nearly twice as great and is given by the expression: aD ouo[ = 08pd - one) Dt 2.80 For a normally supported compression spring, the diameter increase will be somewhere in the range between these two extreme cases, depending upon the friction con- ditions at the spring seats. For extension springs, the same formulae apply. A. ‘minus sign appears for AD, since an extension spring decreases in diameter during extension. The type of end hooks must also be considered, as it may or may not allow for winding of the spring during extension, 12. Conical Compression Springs Conical compression springs made with a uniform taper from end to end may be calculated with the use of the compression spring formulae by using the average or ‘mean coil diameter. The formulae will not hold after the spring is deflected more than that required to close the largest coil of the spring. Conical springs, like volute springs, are manufactured with variable col spacings. The pitch of the larger coils may be increased to maintain linearity closer to the solid length. Conical springs with a large taper from end to end are very much like volute springs in that the coils can telescope inside each other to give a markedly reduced solid length. In other cases, however, where the taper is small, the conical feature has little effect on the solid length which then becomes simi- lar to that of the straight helical compression spring. For volute springs, the reader is referred to the SAE Volute Spring Manual. Barrel-shaped or hourglass-shaped springs may be calculated as two conical springs mounted in series, each taking one-half the total deflection under a given load, 13. Relaxation or Load Loss in Helical Springs at Elevated Temperatures If a helical spring is compressed by a given amount between parallel plates at elevated temperature, it will be found that the load exerted by the spring will gradually relax or drop off with time. The amount of this relaxa- tion, or set, generally increases as the stress and/or tem- perature increases. Normally, the set is greater for long. periods of time than for short ones. Fig. 5.12 shows spring relaxation for various spring wires. For i tion on heat setting see Chapter 1, Section 4.INITIAL STRESS MPa INITIAL STRESS MPa INITIAL STRESS MPa SAE J178 MUSIC WIRE ‘SAE J157 CHROME SILICON WIRE 800 900 800 2 8 e Ly 700 e é 7 z 5) 5 e/ 500 2 4 { 400 L 02468 wit Wie 6 8 10 1214 1618 LOAD LOSS % LOAD LOSS % SAE J351 CARBON a SAE J132 CHROME VANADIUM. VALVE SPRING WIRE VALVE SPRING WIRE 900 900 200 5 wo 8 700 & 700 fe 5 a 600 Z 0 E 500 Z sot 420 coo LIL 024 6 8 101214 1618 024 6 8 10121416 18 LOAD LOSS % LOAD LOSS % SAE J230 STAINLESS STEEL WIRE SAE J470 INCONEL X750 WIRE 900 900 g 300 = 200 8 700 = 700 & 600 Z 00 E 500 2 500 400 400 0 2 46 8 101214 1618 04 8 12 16 20 2428 32 LOAD Loss % LOAD Loss % SPRINGS TESTED FOR 100h AT INDICATED TEMPERATURES. Fig. 5.12 — Spring relaxation for various spring wires 281Part 3 Design and Manufacture of Torsion Bar Springs SAE HS 796SPRING COMMITTEE H. M. Reigner (Sponsor), Eaton Corp., Engineering & Research Center J. F, Kelly (Chairman), Detroit Steel Products, Div. of Marmon Group K. Campbell (Vice-Chairman), Rockwell International Corp., Suspension Components Div. J. A. Alfes, Pontiac Motor Div., General Motors Corp. T. A. Bank, Firestone Industrial Products Co. J.J. Bozyk, Chrysler Corp.—Product Planning & Devel- ‘opment G. W. Folland, Rockwell International Corp., Suspension ‘Components Div. L. A. Habrle, Engineering Consultant R. E, Hanslip, Toledo Spring Co. D. J. Hayes, United States Stee! Corp. E. H, Judd, Associated Spring—Barnes Group, Inc. W. Mayers, Peterson American Corp. M. W. Mericle, Caterpillar Tractor Co., Materials Div, G. W. Myrick, XMI Tank System E. C, Oldfield, Burton Auto Spring Corp. W, Platko, Chevrolet Motor Div., General Motors Corp, G. L, Radamaker, Eaton Corp., Suspension Div. FT. Rowland, Registered Professional Engineer H. L. Schmedt, Caterpillar Tractor Co.. G. Schremmer, Schnorr-Neise Disc Spring Corp. K. E. Siler, Ford Motor Co., Chassis Engineering JE, Silvis, Winamac Steel Products Div—Norris Indus- tries B. Sterne, Bernhard Sterne Associates W. M. Wood, Associated Spring—Barnes Group, Inc. TORSION BAR SPRING SUBCOMMITTEE, K, Campbell (Chairman), Rockwell Intemational Corp., Suspension Components Div. R Siorek (Vice-Chairman), U.S. Army Tank Auto Com- mand W. Allison, Engineering Consultant J. Boayk, Chrysler Corp. G. Dentel, Chrysler Corp. RE, Hansip, Toledo Spring Co. J. Marsland, Chrysler Corp. W. Platko, General Motors Corp., Chevrolet Motor Div. D. W. Schumann, Ford Motor Co., Light Truck Div. A. F. Skover, Machine Products Co., Inc. B, Sterne, Bernhard Sterne Associates D. Tuttle, General Motors Corp., Oldsmobile Div. W. Young, White Motor Corp.‘TABLE OF CONTENTS CHAPTER I—INTRODUCTION]. . . CHAPTER 2—DESIGN CALCULATIONS T. Symbols Used in Formulae 2. Round Bar in Torsion +3, Rectangular Bar in Torsion... . . A. Straight Bar of Solid Rectangular Cross Section B. Straight Laminated Bar . cee 4, Torsion Bar Spring and Lever 5. Sample Computation . . . A. Requirements for Round Bar. B. Computation for Round Bar C. Computation for Laminated Bar D. Computation for Laminated Bar of Square Cross Section : 6. Operating Stresses . . ace cee A. Suspension Springs... . 2. oan B. Other Torsion Bar Springs . As CHAPTER 3—DESIGN OF END FASTENING] . 1 End Configuration. 5 A. Serrated End Connection . B. Hexagonal End Connection 2. Anchor Member . A. Serrated Anchor . B. Hexagonal Anchor 3. Transition Section . (CHAPTER 4—CONTROL OF ASSEMBLY POSITIO! T. Methods to Insure Correct Vehicle Height . ‘A. Infinitesimal Adjustment by Screw B. Adjustment by Vernier Steps . C. Equal Setrations at Both Ends D. Blocked Serrations at Both Ends E. Variable Height Systems... .. . 2. Marking of Windup Direction . . CHAPTER 5—MATERIAL AND PROCESSING]. . T. General Requirements. : Material. . . Value of Shear Modulus Upsetting and Machining Heat Treatment Straightening Shot Peening . . resetting. : Corrosion Protection. Special Testing . 2. Presetting. A. Load Deflection Curve B. Stress Distribution . Selection of Preset Strain . hrOmmpom>e 23 2B 2B 26 4 28 25 25 26 26 26 27 29CHAPTER 6—FATIGUE LIFE) CHAPTER 7—TORSION BAR SPRING APPLICATIONS 1. Square Torsion Bar for Boat Trailer ao 2, Hexagonal Torsion Bar for Truck Tilt Cab 3. Laminated Torsion Bar for Truck Tilt Cab 4. Torsion Bar With Integral Torque Arm... . CHAPTER §—STABILIZER BARS]. T. Applications. . 2. Design Elements. . . 3. Roll Rate Calculations 4, Stress 3. Design Example . APPENDIX|-SI UNIT AND U.S. CUSTOMARY UNIT CONVER- SION TABLE . 3 31 37 38 47 47 47 47 9 sl 32 i Na See aChapter 1 Introduction ‘Torsion bar springs are used in a wide range of installa- tions, from precision instruments through balance springs, to automotive and military tank suspension springs. The designs vary widely in efficiency of material utilization (tored energy per unit volume of spring material), in complexity, and in production costs. This Manual pre- sents information particularly pertinent to high efficiency springs as used on suspensions for surface vehicles. It is applicable, however, to many types of torsion bar springs for a wide variety of uses 34 The Manual has been revised to reflect up-to-date de- signs and manufacturing processes. It also incorporates the SI metric units in all definitions, design details, and specifications. In the International System of Units (S1), force is not defined by the action of gravity. Accordingly, in this SI (Metric) edition of the Manual, the kilogram (kg) is restricted to the unit of mass {in place of the Pound-mass, or pound avdp}, and the newton (N) is the Lunt of force [in place of the pound-force]. The millimeter (mm) has been chosen as the unit of length.Chapter 2 Design Calculations 1. Symbols Used in Formulae (for SI Units see ayaue = Angle @ when Punic is applied; positive as Appendix) shown in Fig. 2.1, rad B = Angle between reference line and lever at zero P = Force at end of lever, perpendicular to refer- Toad position; positive as shown in Fig. 2.1, ence line (Fig. 2.3), N rad = Static load (used for Pin suspension computa-. @ += = Windup angle when P is applied (= a + 8), tions), N rad a = Angle between reference line and lever when @sate- = Windup angle when Paaue is applied P is applied; positive as shown in Fig. 2.3, (© auunuc + B) (Fig. 2-1), rad rad Sjounce = Windup angle when the maximum operating JOUNCE POSITION he (SEE CHAPTER 2, TORSION BAR STATIC LOAD SECTION 5A) CENTER (LAYOUT) POSITION STATIC 1 REFERENCE LINE x T S(positive) REBOUND (NEGATIVE ‘AS SHOWN) h, (SEE CHAPTER 2, SECTION 5A) (POSITIVE AS SHOWN) REBOUND 6 POSITION ZERO LOAD POSITION WHICH WOULD BE REACHED IF “k" WERE CONSTANT ACTUAL POSITION, AT ZERO LOAD Fig. 2.1—Relation of various load positions at ond of lever 33load is applied, with the lever in jounce posi- tion (= ajounce + A) (Fig. 2.1), rad 4 = Diameter of round bar (outside diameter of tubular bar), mm 4 = Inside diameter of tubular bar, mm iameter of tapered bar at large end, mm jameter of tapered bar at small end, mm Thickness of rectangular bar or square bar, w = Width of rectangular bar (long side of cross section), mm. L = Active length of bar (defined in Chapter 3, Section 3), mm R= Length of lever (Fig. 2.3), mm f Deficction at end of lever from reference line ‘when P is applied, measured parallel to di- rection of P (= R'sin a); positive as shown in Fig. 2.3, mm f. «= Deflection at end of lever from reference line ‘when Pate is applied, measured parallel to direction of Prat (= R sin date); positive as shown in Fig. 23, mm fy = Deflection at end of lever from zero load posi- tion to reference line, measured parallel to direction of P (= R sin 8), mm f+ fy = Total static deflection at end of lever, from zero load position to static load position, ‘measured parallel to direction of Prati tm 6 Effective static deflection at end of lever, equal to load divided by spring rate prevailing at that load, mm T = Torque applied to bar (= P R cos a), N+ mm k = Spring rate at end of lever (variable), mea- sured parallel to direction of P, N/mm ke = Torsional spring rate of bar (= 1/8), N+ mm/rad y= Shear strain 30 28 26 zat 22 20 18 6 ool eee Fig. 22Saint Venant’s cootficionts for rectangular bar in torsion 34 G = Shear modulus (see Chapter 5, Section 1), MPa S, = Shear stress, MPa th Saint Venant’s stress cocficient (Fig. 2.2) Saint Venant’s stiffness coeficient (Fig. 2.2) GQ Load factor defined in GI Rae factor | Chapter 2 G Static deflection factor J Section 4 N= Number of laminae in laminated bar 2, Round Bar in Torsion ‘Table 2.1 shows the relation which exists for the follow- ing conditions: a) Bar is straight. ') Uniform cross section of solid cylindrical, tubular cylindrical, or solid tapered circular configuration, c) If tubular, inside and outside diameters are concen- trie, d) Torsionally loaded only. ) Tapered bars have constant taper, with de (at dis- tance x from end with diameter dain) ~ dain + (@ax ~ rnin) * x/L- When the three types are to have identical torsional rate ky and identical siress rate S./6, the following rela- tions exist between them: 1. The ratio L/d (L/dmin for the solid tapered bar) is the same for all three. 2. The volume (mass) of the solid cylindrical bar can be reduced by changing to one of the other two types: ‘a) The space requirements for the tubular cylindri- cal bar are larger than for the solid cylindrical bar because both d and L increase by a factor VIAT= GI However, these spacial increases are more than balanced for volume by the growing interior space as di (and there- fore the factor) are reduced. ') In the solid tapered bar, as the ratio dmax/4ain increases, dax increases against d in the solid cylindrical bar, but dain decreases by a larger percentage, and there fore L also decreases. The net result is that the volume ‘of the solid tapered bar becomes ever smaller as the ratio dnax/Gnin is increased. 3, Rectangular Bar in Torsion A, Straight Bar of Solid Rectangular Cross Section For a straight bar of solid rectangular cross section, loaded in torsion only, the following relations exis‘Windup angle (radian): oe SE “ptwG mtG Torsional rate (N+ mm/rad): ee 6 ntwo L ‘Stress (MPa): mew ‘Stress rate (MPa/rad): S _ mia 6° mL B, Straight Laminated Bar For a straight laminated bar (each of the Ny laminae having equal rectangular cross section, same material, same length), loaded in torsion only, the relations are: ie = BNLUWG TE _ TNLtG ee ° mL ‘When the bar has Nu: laminae of t; thickness, ws Ty torque, na: and ma, coefficients; also Nis lar te, Wa, Ts Taz and maa; also Nis laminae . . ‘Windup angle: ee ae mit?wG gitG TL _ m2SL Mat?wG M2 andso on Torsional rate: 1 6 Nu maiti?wiG + Nis met! weG + L : kr Nukn + Niskre + Stress: m202G mal = 1 0UG s= mab and so on ‘TABLE 2.1—FORMULAE FOR ROUND BARS IN TORSION Type of Round Bar Tubular uanty onal sit Wisp get radian ‘gle eC GET 25 2Sb ae ae Torsional T T Rate Ld ‘ ® No nirag roe | raat —esian) | 26 a poy me mg owe | sg | Pe edg iS. , Sn 20 ZL Flas FF ae cial eeye|peea a6 i” eae e ae verb Pass 35In actual practice, the dimensions for a laminated tor- sion bar can be established by the following procedure, considering the case where the complete bar is to have a rate kr and this cross section: From given data the maximum windup angle (ounce) will be computed at which all laminae may be subjected to the maximum allowable stress (Sujounce), with all lami- nae having the same active length (L), the maximum allowable thickness for each lamina is: ae, Swovnco L fom ns Broun F Any lamina thicker than tgs will be overstressed. Allaminae are wound up tothe sar Bae and WOU carty the same maximum stres if the thickness ofall Taminae could be held t0 tae, This would result in the rust effcient spring or in a spring of minimum weight. In most cases this isnot practical andthe thickness of the laminae must be adjusted to ive the required rate. Te should, however, be kept in mind that the stresses shonld beheld as close as possible to the allowable maxi mom stress in order to keep the spring weight to ami mom, The torsional rate of the complete bar is: er = Nukes + Nuakes +... Niskm In general, this result will not coincide with the desired value calculated from Reosa at+B ke = See Chapter 2, Section 4. If the rate as checked by load-deflection test turns out too low, another lamina will have to be added. If it turns ‘out too high, thinner laminae must be used, or the widths of the laminae must be reduced. It is important to remember that no laminae should ‘exceed tmax. but all laminae should be kept as close as possible to this thickness. The stress is not affected by fa change in width (w) as long as the ratio mo/n2 this holds true for most practical applications where (w/t) > 3.5, Consequently, changing w is a convenient way of rate adjustment if the widths of the laminae are not required to be held to existing commercial sizes. Changing the active length (L) of the bar, provided the specific design allows this, is also a means of rate adjustment; however, L also affects the stress, and the whole computation will have to be repeated. 4. Torsion Bar Spring and Lever Fig. 2.3 shows a combination of torsion bar spring ‘and lever which is frequently used in suspensions. In this, diagram the deflection f and the angles a and 8 are mea- sured from a reference line which is perpendicular to the applied load and passes through the center of the torsion bar. They are counted positive when their relations, to the reference line are as shown in Fig. 2.3. REFERENCE. LINE. Fig. 23-Torsion bar sping and lover ‘The load deflection characteristics of this mechanism are not linear but are given by the Following: pet Resa r <5 e=atB r emcee kr(o + 8). kt ath_ 0 wo Ser? = Bolo = Sa = aa facing ‘The function C; is plotted in Fig. 2.4 against the ratio F/R or sin a as abscissa, so that the curves represent load-deflection diagrams for the end of the lever. The rate will be a minimum when dk/df is equal to zero, or when =3sina OOF Feinta 1 362a} 2 2a 2a 2a) Lon FACTOR C)—= & i ; | AXING STAT DEFLECTION 4 ¥ It will be noted from the chart that C, becomes equal i 1P/df). Using the value of P above and the relation ‘ ky ‘ pat, @th | paar z\ dP _ kr tana, 1 | 88 Mla + pte + ( ge” Ble + Sa * weal Wl aP_ kf -@+ Ariane) ect amtond Rsina Fig, 24—Load factor versus deflection at Fg = Rema ap _aP da ‘ eee ee ai Lt @t Bina Re costa rektig -Eeg + (a + tana c= testa costa Fig. 2.5 shows the function C; plotted against the ratio a7— woK LetCeay TIN al Pee IE att ! | | at +; | 1 I] pa Mh t se a wo 4 ery 74 j [ent ware | ab cect [‘erataz} | . a hd b 7 2 a s 2 1 2 20" 30 40" so" oe desma Fg, 25—Rate factor versus datiction £/R or sin a as abscissa. It shows that the rate of such RG a torsion bar and lever is by no means constant, and that the minimum rate occurs at a position of lever center- G cos a line below the horizontal. When {/R is zero then for Tb me .0. at+f all values of B, Co ‘The rate will be a minimum when dk/df is equal to zero, of when C2 = 1/(2 sin? a + 1) only for negative values of sin a. The static deflection at any point is defined as 5 = P/k. Using the above values this becomes cosa 1 atB 8 + tna 38 Fig. 2.6 shows the function C plotted against the ratio £/R or sin a as abscissa, so that the curves represent a ‘comparison of static vertical deflection with the vertical deflection from the reference line. By definition, the ab- scissa values in Fig. 2.6 actually represent sin agave OF AYR. ‘The rate will be a minimum when dk/df is equal to zero, or when C = ~3 sin a.. ce] | 2 + i [sana rare | (ese3.ma]—F | a ‘fpvmetaae stave cerccron | t _ |—+— | = STATIC DEFLECTION FACTOR cy—= ss = a * i Fig. 26—Statc deflection factor versus deflection It is to be noted that the curves for minimum rate, 5, Sample Computation ‘maximum static deflection, and = 90 deg all intersect at /R = 10 of C= 30. II angles in previous equations are expressed A. Requirements for Round Bar in radians. In Figs. 2.4-2.6, the angle 2 is given in degrees, but for calculation of loads and rates from the constants Cy and Cy the torsional rate must be expressed in the Required: Determination of diameter (4) and active dimensions N + mm/radian. Where necessary, angles are __length (L) on a torsion bar spring of round cross section converted from degrees to radians by the relation 1 ra- for a passenger car suspension for the given values: dian = 57.296 deg as shown in the examples. Static load: Pyutse = 4000 N Rate at static load: k = 16 N/mm Length of lever: R = 400 mm Position of lever at static load: atte = 7 deg = 40.122 rad 39Deflection, static load to jounce position: he = 100 mm Deflection, static load to rebound position (with rebound stop): hy = 125 mm Stress at static load: Swotatey < 670 MPa Stress at jounce position: Swqounee? < 900 MP2 For recommended stress values see Chapter 2, Sec- tion 6. B, Computation for Round Bar The effective static defection at static load 8 = 4000/ 16 = 250 mm. Therefore 250 = 2 = = 000s . 400 2 R Position of end of lever above horizontal at static load {all positions shown in Fig. 2.1) R sin dye = $0.12 + 400 = +48.8 mm From Fig. 2.6 or the formula 0s aac = aes = ———oo Sete _ Co OS = Trane + A) + fa ane . 0.993 170.122 + A) + 0.123 2593 _ 9.123 = 1.466 O12 +2 0.625 1 B= Tyg 0 = 560 radians = 32 deg From Fig. 2.5 with f,/R = +0.122 and B = 32 deg; C= 1:10 KR? _ 16 + 400" CG 110 Therefore, ke = 2327000 N+ mm/rad Rechecking with Fig. 24 or formula we obtain _ Kx(@auue + 8) R 605 Gaui _ 2327000 (0.122 + 0.560) 400 + 0.993 4000.N ‘The lever angle at jounce position is obtained from fy the _ 48.8 + 100 R 400 Giounce = $21.8 deg = +0.38 radian sin ajounce 0.372 310 ‘The lever angle at the rebound position is obtained from fy — he _ 488 — 125 R 400 0.191 11.0 deg = ~0.19 radian ‘The torsional windup angle is 8 + aiounce = 32 deg. + 21.8 deg = $3.8 deg = 0.94 radians. ‘The maximum torque Tiounee = kr (B + Gyosnee) = 2.327 000 (0.56 + 0.38) = 2187000 N+ mm Then, 16 Trance _ 16 + 2187 000 p= Io Townes _ 16 + 2187000 _ 12 * T Sajouncer 3.1416 + 900 ug a= 24 Use d= 23 mm LPR _ 16 4000 - 400 Seats = — 8 7 3.1416 + 12 167 Cole To produce the required rate, the active length must _ E1416 + 280.000 + 76.000 32 + 2327000 Use L = 900 mm We shall sketch a line corresponding to B = 32 deg in Fig. 2.5 and mark the limits {/R = —0.53 at free position (7K = —0.19 at rebound £/R = 0.12 at static 1/R = —0.38 at jounce ‘The curve shows that the rate, going from static to jounce, builds up rapidly from Cz = 1.10 to a value of Cy = 1.60, of a rate change of 45%. The rebound rate first decreases and then builds up again. At the rebound position Cz = 0.96 indicating a decrease of 13% from the rate at static load. At zero load C2 = 1.40. This arrangement, with the radius arm at a small angle above the horizontal at the static load position, gives an increas- ing rate from this position to the jounce position which is desirable, The rate increase near the free position limits the deflection of the arm and thus facilitates assembly. ‘This spring requires presetting. (See Chapter 2, Table 22) For method of selecting preset angle, see Chapter 5.Computation for Laminated Bar If the active length of the round bar is found to be considerably in excess ofthe available space, then consid- eration may be given toa laminated bar. This will require more cross-sectional space, but much less length. Use of rectangular cross section material in the “as rolled” condition may dictate a slight reduction in maximum allowable sre. ‘A laminated bar with square cross section (Nut = w) can be fitted into a square hole anchor. A bar with stepped ‘width laminae requires a more complicated anchor con- figuration, but it has the advantage of reducing the com- pressive stresses between bar and anchor, since each step in the anchor takes only the torque of one or two laminae; also it improves the environmental clearance conditions 1 its cross section approaches a circle For this sample computation it will be assumed that 435% reduction in active length is Fequized (From 900 to 585 mm), that the stress at jounce position is to be limited to 860 MPa, and that w/t will exceed 4.0 so that "= 1 ‘The requirements for load, angular position, and deflec- tion are to remain unchanged: therefore, the torsional rate is to be maintained at kr = 2.327 000 N + mm/rad. ‘Then tar = 7 Simul: =» 800 + 585 87 en ms 0G 0.940 = 76000 D. Computation for Laminated Bar of Square Cross Section For the square cross section (Nu = w), and assuming ty = 03: ky were mewe we [Ee oe THO 0.3 » 76000 39708 = 244 Using tw = Nt? = 244 and tax = 7.04 Nucmiey = 244 qon = $92 Let N= 5; then: = 244/5 = 48.8 10 5+ 70 = 350 0.33 — (0.2/5) = 0.333 ~ 0.040 = 0.293 ant 0.29: 76.000 343+ 35 385 = 2290000 (which is within 14 % of the required rate, thus satisfac- tory) = 86» 22 Sinn = 860 + 25 = 855 MPa (This solution is satisfactory.) With density of steel = 7850 kg/m Mass = 7.85 + 10" + $85 + 35? = 5.6ke, Required Clearance = diagonal of the'square L414 + 35 49.5 mm 6. Operating Stresses A, Suspension Springs Suspension springs are generally loaded in one direction only and, as can be seen from Table 2.2, they are usually shot peened and preset. The bars are subjected to both processes because this enables the springs to operate at the high stresses required of them. Without shot pening and presetting, the bars would encounter early fatigue failures and would suffer excessive settling. Both shot pening and presetting are further discussed in other sec- tions of this manual. It can also be noted that research and development is pushing the upper edges of Table 2.2 upward. As materials and processes are improved, added classes can be expected. Extension of applications to reduce weight and size will follow as experience in the higher stress classes is gained, ‘Classes 1250-1100—Operation of torsion bars at these stress levels requires close attention to material and pro- duction quality as the critical flaw size becomes smaller in inverse proportion to the operating stress level. With proper precautions in design and manufacturing, the ex- pected limited spring life can be extended to acceptable levels. Only minimal applications have been logged in these classes, primarily in advanced military vehicles. Ex- tension to commercial applications will bear close cost- ‘benefit analyses because of the costs involved in the design ‘and manufacture of material in these stress ranges. With proper handling, very extended lives have been displayed by early production material{ \ ‘TABLE 22ROUND TORSION BAR CLASSES or Tove Brecon Te Sea weed | car? | sating cass | sverstiax | elite | Sea stot Damo | Danese wo | "taba | Gotacton | By | reverse | recta | preset | uptsaiinm | upieinm | Typical Apleator zo | 00) owe [ve | veo | Yen "oso Ce a eo [ [tomer | vee [mp vee | ve —— foo Paw tome Pre ot ee eS SRE ay a oe a rereeo Lnneye wees wo [00 | 2a [vw | v6 [v™ | teteoe Paeeoe! caps 800, 800 a4 Yes - Yes Yes 15621 41500 Truck suspensions nese =} eros | taeseo Fasetn | tae bars 700 | 700 [24 | =] Pomote | Aavoaures | — | o1bo00 sco | sc0 | 2m | — | pomsorw | Asvoaurea | — | ggigog |__| Hakhcovers tos on i040 sso | s50 | ama | — | romaate | arveawes | — | St 10200 * Torsion bars in class 550S are normaly produced from shallow hardening carbon steel grades; they cannot be expected to have the depth of harden which can be achieved wih alloy steels. *(ESA)—Electro Slag Remett 3.12iid et is oS mae RAM nk Ci sete G ‘Class 1000—Bars operating at these stress levels have now been in production for over 20 years. Operation at these stress levels implies limited spring life, but with proper attention to design and manufacturing detail ceptable extended life can be obtained. Particular empha- sis on minimizing operating stress range can do much to assure extended life. By referring to the conservative “Life Test” diagram, (Chapter 6, Fig. 6.1) it will be ob- served that a shot peened and preset bar operating be- tween a maximum stress of 950 MPa and a minimum stress of 170 MPa is estimated to have a minimum fatigue life of about 60.000 full cycles. On the other hand, a bar operating under a stress range reduced from 780 to 600 MPa (between a maximum stress of 950 MPa and ‘a minimum stress of 340 MPa) will have a minimum life of about 140.000 cycles. Class 900—This might be called a typical passenger car torsion bar spring. A bar required to have a life in the laboratory in excess of 100.000 cycles, with a maxi mum stress of about 900 MPa and a'minimum of about 275 MPa has proved very satisfactory in the field. From the diagram, a minimum life in the laboratory of about 120000 cycles would be obtained. ‘Class 800—This is a typical truck suspension torsion bar spring. To achieve the longer life required of a truck spring and to take into account the harder usage, the maximum stress is limited to 800 MPa. According to the chart, 9 truck spring designed to a maximum stress, of 800 MPa and a minimum of 200 MPa should have a minimum life of 250.000 cycles in the laboratory. In the discussion of fatigue life, it was noted in each ccase that the expected life was that to be obtained in the laboratory. In the laboratory, the bar is usually cycled. between the metal-to-metal jounce and rebound positions, ‘but in actual service, very seldom is the bar called on ‘to withstand these extreme cycles. A formal road test program or a record of failures, or lack of failures, should 3.13 ‘be made to attempt to find a correlation between actual service life and laboratory life. After a correlation has ‘been established between actual service and laboratory fatigue life, the laboratory testing becomes a quality con- trol tool ‘Classes 1000, 900, and 800—Rectangular and Lami- ‘nated Bars—Since most bars of rectangular cross section, are fabricated from “as rolled” stock, their inferior surface finish, as well as the secondary stresses that occur in ssuch bars, when twisted, must be taken into account. A. reduction of the maximum operating stresses (Table 2.2) by 5% is, therefore, recommended when using Class 1000, 900, and 800 steels B, Classes 700 and 550—Other Torsion Bar Springs ‘Counterbalance Springs—Counterbalance springs run the gamut of stress ranges. Fortunately, they do not usu. ally have the service life required of them that suspension, springs do; for example, 10.000 cycles of a tilt cab bar represents about 30 years life at the rate of one cycle per day. Generally, counterbalance springs are under load most of the time and are only relaxed when the mecha- they are counterbalancing is operated. Because of this condition, the maximum stress is usually limited to about 700 MPa. Counterbalance torsion bars sometimes, ‘because of space limitations, have to operate at stresses ‘over 700 MPa and it is a practical solution to preset the spring, Stabilizer Bars—Stabilizer bars are generally designed to give a certain roll resistance, and the stress is of second- ary consideration. Because a stabilizer bar is a part of the suspension and is subjected to reverse loading, it should be shot pened if some areas are stressed in excess ‘of 700 MPa, but it need only be shot peened in the highly stressed areas.Chapter 3 End Configuration A. Serrated End Connection Of all known arrangements, serrated ends permit the smallest end diameter. Satisfactory static strength may be obtained even if the diameter of the serrated end is, small. However, experience indicates that the minor diam- eter of the smallest serrated end should not be less than 1.15 times the diameter of the bar. Refer to ANSI B92.1 (latest revision) for spline and serration design and inspection. ‘The serrations themselves must be designed with two features in mind: ‘Stress Concentration at the Serration Root—At the point where the anchor first contacts the serration (next to the transition section) the bar stil is under the full windup torque. Even with the enlarged end diameter, the stress at the bottom of the serration root may be ‘wearly as high as that in the body of the bar. Added to \Cois a stress pattern due to the localized pressures on, the serration flanks, so that the total stress here may be very high. ‘This stress concentration can be minimized by provid- ing the largest root radius compatible with the anchor serration contour. The standard serration pressure angle (at the pitch diameter) is 45 deg which is large enough to permit effective shot pening. Surface finish of the serra- tion contour is usually not specified because the methods. of forming produce an acceptable finish. ‘Compressive Stress on the Flanks of the Serration— ‘A design that is favorable for the first condition (large root radius) reduces the flank area, and therefore increases, the pressure on the flanks. If this pressure is too high, fan excessive local permanent deformation will occur somewhere near the contact area, initiating a fatigue frac- ture. Experience has shown that a compressive stress equal to approximately 140% of ultimate tensile stress is permissible when it is calculated for full length contact, ‘on approximately 25% of total serrations in accordance with the hertz equations for stress due to pressure between curved elastic bodies. Formation of serrations may be done before or after heat treatment depending on manufacturing procedure. However, it should be noted that the serrations will be 315 Design of End Fastening subject toa final acceptance gage. The possibility of serra- tion distortion exists when heat treatment follows the serration forming operatio ‘Serrations shall be parallel to the centerline of the bar in accordance with tolerances specified on the individual part drawing, ‘A slight change of serration size is produced by shot peening operations. Assembly problems sometimes result when a serrated bar end is machined to the maximum size, as defined on a drawing, and gaging is conducted directly after the machining operation with shot peening following later. To insure absolutely proper assembly of serrated torsion bars into their anchors, it is necessary to conduct composite “go gage” checks of the serrations after all shot peening is completed. External and internal serrations when new will not show full contact. During the first few load applications, the high spots will take a local compressive set and the required contact area will be established. On parts which have been fabricated with reasonable accuracy, the perma- nent set required to produce this contact is not large ‘enough to cause damage. Both external and internal serra- tions deform elastically during load application, and the contact will therefore be limited in length. Experience hhas shown that it is satisfactory to make the length of the serrations equal to 0.4 times the major diameter. B. Hexagonal End Connection Hexagonal ends for round torsion bars can be produced without machining. They offer advantages in high volume production where upsetting machines in various degrees of automation can be utilized. The usual practice is to centerless grind the bar stock, ‘cut to a predetermined length, upset the ends toa hexagon shape, and finish with a coining operation to improve the flatness of the sides, There is no need for further machining or grinding. Part numbers, arrows or letters indicating the direction of rotation, and other coded data are usually put in during the upsetting operation. The transition section, which is being formed by the plastic flow of the material, is left in the “as forged” condition. ‘The slightly rougher surface of the transition section is not detrimental as long as no pits or inclusions are present. The ratio of width across flats to bar diameter muststrkerunne | \ ‘. caoove 7 ENAMELING L cove ronstee. wear /L enoove 120 0€6° 6008 Fig. 3.1—Typlcal design of hexagonal end connection be sufficiently large to preclude high stress in the hex section which would cause end failures. Ratios of 1.4, 1.3, and 1.2 for classes 1000, 900, and 800, respectively, have proved satisfactory in laboratory and field testing For classes below 800, a minimum ratio of 1.20 is recom- mended, A typical design is shown in Fig. 3.1. The dimension. ‘across the hexagon flats has an cveral! tolerance of 0.40 mm. A minimum corner radius of 0.75 mm is specified, to go with a 1.00 mm maximum corner radius in the anchor. To insure against the possibility of the comers being too round, itis necessary to specify a dimension ‘across the corners of the torsion bar hexagon. Otherwise, ‘a.cam action by the hexagon in the anchor is conceivable which may cause a bursting of the anchor. 2. Anchor Member A. Serrated Anchor Stresses in the torsion bar anchor or hub will not be as high as they are in the bar, although high local stresses exist. The internal serrations must also be designed with generous fillets in order to avoid high stress concentra- tions. Shot pening is not necessary. ‘The high pressure angle required for a durable serration design will produce large radial components of the flank pressures, and, therefore, a tendency to burst the anchor, as if it were loaded by internal hydraulic pressure. Ade- quate material must be provided behind the teeth to with- stand this internal pressure ‘The internal serrations should be longer than the exter- nal serrations and so positioned that they will overlap 316 the external serrations at both ends, regardless of toler- ances and assembly variations. ‘The hardness of the anchor can be considerably lower than that of the bar. In many installations it has been held below Rockwell C 30. If broaching is scheduled to follow heat treating, the hardness should not exceed Rockwell C 36 in consideration of tool life. (See ANSI B92.1.) B. Hexagonal Anchor AAs in the serrated anchor, local high stress areas exist in the hexagonal anchor. They are located at the corners of the internal hexagon, and unless sufficient wall thick- ress is maintained, the anchor will split through the cor- ners 180 deg apart. This will be particularly true if the torsion bar has overgenerous radii at the comers. The bar will then act as a cam and cause a wedging action The internal portion of the hexagon anchor can be formed by broaching or extruding, but the parallelism of the sides must be held within 0.004 mmm of length ‘A surface finish of approximately 1.9 jum or less should be maintained. The hardness of the anchor should be substantially beiow the hardness of the torsion bar. Anchor hardness (6 internal faces) of Rockwell C 25-30 has been found satisfactory. This lower hardness permits the relatively rough finish on the bar (generally in the as-forged condi- tion) to “seat” into the anchor faces. Otherwise, the bar tends to bear only on the high spots. This introduces. serious stress concentrations in the bar which have been found to lead to premature failure. However, the hardness ‘cannot be substantially less than Rockwell C 25 or settling of the bar end into the anchor face will change the relative position of the anchor and torsion bar appreciably.‘The length of the internal hexagon should be such that the torsion bar is in full contact with the anchor. The transition portion must be protected from corrosion if the anchor extends beyond the bar so as to form a pocket, for the retension of corrosive elements. ‘A maximum clearance of 0.50 mm across the flats has ‘been found to provide the necessary clearance between the bar and the anchor. The minimum clearance is deter- ‘mined by the relative position of the anchors and the desired ease of assembly. 3, Transition Section ‘The transition from bar diameter to end diameter should be gradual in order to keep stress concentrations to a minimum. A taper of 30 deg included angle and a fillet radius of 1.5 times bar diameter have been found tobe satisfactory. Some users have specified an arc instead ‘of a uniform taper in blending the end diameter with the bar diameter. A radius of four times the bar diameter hhas been found to be satisfactory. ‘With splined torsion bars where thin-walled anchors are utilized, and especially where a tubular anchor struc- ture is used to reverse the direction between a solid torsion ber and a tubular torsion bar, additional transition con- figuration controls are required which are not shown in Fig. 3.2. Specifically, the sharp edge produced at the june- ture of the trensition radius or taper and the end diameter ‘ust be radiused. Experience to date has shown a radius Of three times the spline depth is adequate. ‘The active length “L” of the torsion bar spring is not immediately apparent on the spring drawing which gener- ally notes the overall length (L,,) and the lengths of the two end sections (Lena). Adjacent to each end section there is a “transition section” of length 1 which may bbe considered to consist of one inactive portion close to the end section of diameter D, and one active portion close to the active bar diameter (d). This portion has : ole fees we ne expen} | 317 the same flexibility as a bar of diameter d and an “equiva- lent length” te. See the sketches in Fig. 3.2. The type of transition section (uniform taper or arc, Fig. 3.2) affects the flexibility of the bar. ‘When the transition section has the form of a uniform taper where the length 1 can be readily computed or measured, the equivalent length 1. equals the product of 4 and a factor Q which is found in Fig. 3.3 at the intersection between the appropriate D/d value on the abscissa and the Q curve (see right hand ordinate scale). When the transition section has the form of a circular are with radius +, the length J cannot be readily estab- lished on the drawing, but it can be computed as the product of the bar diameter d and the factor 1/d which in turn is found on the left hand ordinate scale in Fig. 3.3 at the intersection between the appropriate D/d value ‘on the abscissa and the appropriate r/d curve (with inter- polation between two curves if necessary). Then the equiv- alent length 4, equals the product of | and a factor V which is found in Fig. 3.3 at the intersection between D/d and the V curve (see right hand ordinate scale). ‘The active torsion bar spring length is: Loi — 2Lnea — 24 + 24, 20h) 21 ~.Q or 210a-%) ‘The derivation of the Q and V factors follows. ‘The windup angle of a short element of length dx and diameter y under an applied torque T is _ 21 dx oa rye From this, the windup angle of a uniformly tapered section can be calculated by integration as re ih ze * tbat Ga) + (Ga) ] Suse = tenons tev tener oF rransirion—f— Fig. 8.2—Transition with uniform tapor (et; and with circular arc (ight)#/d (RELATED TO r/d) For the tanion section of length & between the active bar of camater snd the native and of slemeterO (D = root drnete forthe serrated end; ‘Ds invrbedclametar for hexagonal end), the equivalent ative langth ya diameter dis V- when transition ian are of racusr, € whantanition lea uniform taper 10 08 168 06 FACTORS Q AND V 14 oO 08 o6 os Md Fig, 83—Torsion bar springs, transition data 318i] | where: Oper = Windup angle of a uniformly tapered section d= Diameter of small end, equal to diameter of bar D = Diameter of large end 4 = Length of taper This is correct for any taper angle. Comparing this ‘equation with the equation for windup angle in Chapter 2, we see that the tapered portion has the same flexibility as a bar of diameter d and a length 1, tT ilba* Ga) + Ga) |-eees where: 4, = Equivalent active length of transition section at ber diameter d From the formula for windup angle d 8, the windup angle of the transition section formed as a circular are ‘ean be found by integration as ane aaa* ops a) D>, (PY a+ 0? + 15(2) ‘Comparing this equation with the equation for windup angle in Chapter 2, we sce that the transition section, 3.19 hhas the same flexibility as a bar of dismeter d and a length Je Dy (2) #402 +52) wen (Baa | =v re 39 It should be noted that this equation is developed for ‘4 parabolic arc rather than a circular arc because of ease of calculation. This license introduces an insignificant er- We can consider the value Je the equivalent active length of the transition section. Fig. 3.3 gives the value Of this equivalent in a convenient form. The total active length of the bar will be L = Length between transition sections plus equivalent active length of both transition sections. ‘A transition section formed by a circular arc may: be specified by the use of the following formulae (Fig. 3.3). gs 4 £ s@-1 a~ Vale For a given bar diameter and end diameter, the length of the transition section will be directly specified by the radius of the arcChapter 4 1, Methods to Insure Correct Vehicle Height A. Infinitesimal Adjustment by Screw This system is desirable where vehicle height is impor- tant and where a definite preload is necessary to prevent reversals of load. The screw adjustment is being used to overcome tolerance stackups of mating parts, variances in Bt between external and internal parts, variances in load to be carried, and also, to overcome settling problems Fig. 4.1 shows two typical torsion bar adjusting mecha- nisms using hexagonal torsion bar anchor ends to transmit torque from the torsion bar to the anchor housing. A screw operating a lever on the anchor housing converts the torque into a force and lever arrangement. Spherical oF cylindrical seats must be used to toggle the bolt in ‘order to eliminate any possibility of bolt bending. During assembly, the parts are positioned near the de- ~d free position, the torsion bar is assembled to the \.-chor housing which is attached to the reactionary mem- ber, and the torsion ber is wound up to the desired angle. Variances ofthe adjusting mechanism range from bolts Control of Assembly Position being used in tension, to bolts being used in compression and a number of other combinations. Provisions for protecting the assembly from rust are commonly used. B. Adjustment by Vernier Steps. ‘This system is desirable for experimental installations where the static lond or standing height may need to be changed. It is subject to errors and, therefore, not desirable in large production. ‘The two ends of the bar are made with different num- bers of serrations; for example, 55 serrations on one end, '57 serrations on the other. Each step of adjustment is, then equal to (360/n,) — (360/ns) deg, where n; and nz are the numbers of serrations at the two ends. For convenience in assembly, the end with the smaller ‘number of serrations should also be smaller in outer diara- eter, so that it will pass through the bore of the large internal serration. In assembly, the lever must then be located in the de- sired free position (zero load position) and the bar turned. until a position is found at which both serrated ends will, enter. crossmemaen Fig. 4.1—Two typical torsion bar adjusting mechanisms a2t. Equal Serrations at Both Ends In this system, the smallest step of adjustment equals 360/n deg, where n is the number of serrations. The as- sembly tolerances will usually be smaller than 360/n, so that adjustment during assembly is not possible. ‘The internal serrations in lever and anchor must be broached in relation to fixed reference lines and the bar must be so machined and preset that the serrations at ‘both ends are in proper relation. The permanent set after the first windup must be measured, and the angle of suc- ‘ceeding windups adjusted correspondingly. This has been done in large production runs without difficulty. During assembly, the lever is positioned near the de~ sired free position, and the bar is inserted, forcing the lever to the correct position. This type of assembly is used for volume production where vehicle height consid- ‘eration and settling problems are of relatively. minor im- portance. D. Blocked Serrations at Both Ends ‘This system can be applied to (Section C) in order to insure correct assembly. One tooth is removed from each external serration and one space is omitted in each inter- nal serration (by removing one tooth from the broach, see Fig. 4.2). These blocking serrations are kept in proper position in anchor, lever, and bar, so that only the correct, assembly position is possible. ‘Where bars with clockwise and counterclockwise windup are used, the blocking serrations can be so ar- ranged that bars will it only in the correct station. E, Variable Height Systems ‘These systems fall into two broad categories; mechani- cally actuated and hydraulically actuated. It is usual to apply one of the systems outlined in Sections B, C, and/ = sors Gecrwans | 4 gr sie Fig. 42—Applcaion of missing tooth and missing tooth space for indexing or D to assure proper system assembly and component alignment. Beyond that, proper vehicle elevation and/ or attitude may be manually or automatically controlled ‘or a combination of these. Power supply and control system configurations also present endless combinations and can involve many me- chanical and hydraulic subsystems outside of the scope of this Manuel. 2. Marking of Windup Direction Even where blocked serrations are used, bars should bbe marked for direction of windup to avoid mistakes in fabrication and assembly. This should also be done on bbars which are not preset, but will be stressed at or near the recommended operating stresses. Such bars will take a small permanent set and must be reinstalled in a station with the same windup direction if they have been removed for servicing. 3.22Chapter 5 Material and Processing 1, General Requirements Selection of material and proper specification and con- trol of processing are vital factors in the production of torsion bars, Applications involving high stresses necessi- tate strict hardenability and quality requirements for the material and suitable control of decarburization, heat treating, shot peening, and presetting, plus quality assur- ance provisions, even to the extent of fatigue testing of samples of each production lot. From this ultimate cate- gory, torsion bars vary all the way to parts using heat treated carbon steels at low stresses involving no special requirements or processing. The following brief discus- sions and tables are presented as basic information to assist the designer in detailed specifications. They will also serve to direct attention to certain aspects of process ing which are of eritical importance ‘A. Material ‘The hardenabilify of the specific material selected for a torsion bar must not only guarantee that an adequate surface hardness can be achieved, but must also insure that the proper center hardness t0 achieve the optimum residual stress distribution—as determined by expeti- ence—will be obtained. Since the surface of a spring is of major importance, the quality of steel ordered should reflect experience gained under service conditions. Bars which operate at a high stress level may require purchase of Magnaflux Quality steel (Aircraft Quality ifan aireraft application is specified), which details special steel testing practices such as magnetic particle inspection. ‘Also, rigid inspection standards for finished springs should be specified as the application requires. For bars designed to operate at somewhat lower stresses, it is recommended that a standard of surface quality be estab lished by negotiation between the manufacturer and the steel producer, and that surface inspection standards be made as definite as possible to avoid misunderstandings. Centerless grinding of hot rolled steel may be utilized for two primary purposes. The grinding insures the toler~ ances required on the bar diameter; also the amount of steel ground off may be suficient to improve surface con- ditions Table 5.1 suggests hardenability specifications for com- 3.23 ‘TABLE 5.1—RECOMMENDED MINIMUM. HARDENABILITY Aoproximate Dia. | Minimum Hardenabity, Jominy ‘of Body, mm_| Oistance in mm at 65 Rockwall C 2 05 ae 130 5 160 6 190 mon sizes of alloy bars. To achieve a 90% minimum martensitic structure, itis necessary to select a steel that will meet these minimum hardenability criteria. Typical hhardenability by grade can be determined from Hardena- ity Bands for Carbon and Alloy H Steels, SAE Hand- ‘book, J1268. The depth of hardness for a given steel may 'be improved by utilizing a strong oil quench with violent agitation, The extent to which the required improvement ‘can be achieved is usually determined by experiment. Grade selection of carbon steels is usually based on carbon and manganese content, with standard SAE 1065, (G 10650) through SAE 1095 (G 10950) analyses gener- ally chosen; these are usually fine grained killed steels. ‘Standard Hardenability Bands for Carbon H Steels, SAE Handbook J 1268. ‘Table 5.2 gives recommendations for magnetic particle inspection specifications. Category A is in the form cur rently used for heavy duty class 1000 applications. See Table 2.2 in Chapter 2, Section 6A. Category B represents the limited stated requirements ‘TABLE 5 2—RECOMMENDED MAGNETIC PARTICLE INSPECTION REQUIREMENTS Requrement Category (Appticable to Hardened Springs) a ‘Ro evidence of Grcumlerentalindications ong tudinal cracks, heavy seams, continuous hight Indeations, heat checks, laminations, oF flakes. gt longitudinal indications wai a ‘Greumferental width of 8 men shal ot exceed 2 loegiuginal length of 100 mm indivaualy ‘ano/or collectively oF a8 agreed between see Suppliet and user Noevicence of creumlerental indications, lang ‘usinal cracks, or heavy Seams, {As agrees between supplier and manulacture.for highly stressed bars when extensive negotiations have taken place. Similarly the lower stressed bars designated as category C generally do not have written drawing re {quirements and the supplier agreements are less compre hensive in scope. B. Value of Shear Modulus ‘The value of the shear modulus “G” recommended for use in design and stress calculations is 76.0 X 10° MPa for steel. While somewhat lower values have been used in the past, recent tests confirm the recommended figure, which has become accepted as standard for harden- able carbon and low allow steels in many modern refer- tence works and handbooks. Since tests are partly depen- dent on shape, the values of G thus obtained show some fluctuation, but for torsion bars above 16 mm diameter and flat bars requiring hot forming, the suggested figure of 76.0 X 10? is applicable. 79.3 x 10° MPa is recom- mended as the value of the shear modulus for round torsion bars below 16 mm diameter. The shear modulus is also known as the modulus of rigidity and torsional modulus of elasticity. Identical values of G should be used for torsion bars, with and without presetting, A substantial number of windup tests show that the rate of the torsion bar—and therefore G—is noticeably reduced s an immediate con- sequence of presetting (while all dimensional values re- ‘main unchanged in the presetting operation). However, the greater part of this loss is restored within relatively few days, and the remaining difference between the orig nal and final G values is less than the normal variation between products of successive heats of steel, or between successive tests on the products of any one heat of steel. C. Upsetting and Machining. For torsion bars with machined ends (such as serra- tions) and machined surfaces in the A and B categories, itis recommended that the material either be die forged or upset from bar stock. The die should be designed to result in uniform material removal in machining over the body and ends of the bars, A microscopic study of the longitudinal c1oss section of the upset or forged ends should show a grain flow structure following the contour of the bar. ‘When bars are upset, itis usually considered advisable to normalize them. If experience has indicated that the heating for forging or normalizing can lead to cracking, ‘especially with high hardenability materials, the cooling should be controlled. This may be accomplished either by slowing down the cooling rate near the exit end of the furnace or by covering the heated portion. Since the surface must be of superior quality, it is necessary not 324 only to provide adequate stock for removal, but t0 achieve ‘a quality surface in machining. It is generally recom- mended that highly stressed parts be ground rather than merely turned, ‘No circumferential grind or tool marks should be evi- dent. Regardless of machining methods, which are op- tional to the manufacturer, there must be no stress risers, in the surface. These generally result from cutting or grinding blends, tool tears, or similar faulty machining conditions. Table 5.3 suggests machining allowances based on the diameter of the raw stock. For torsion bars of other than round section, it is not a general practice to machine cor grind the surfaces, but itis desirable to limit decarburi- zation and surface imperfections by agreement between steel supplier and manufacturer. For certain applications, such as deck lid springs, even the normal hot rolled sur- face may suffice. Because of the cost of surface removal, ‘each application should be surveyed individually for stock removal requirements. Also, for very high volume appli- cations, special surface quality negotiations with suppliers may result in less allowance than would be otherwise necessary. ‘TABLE 5.3-RECOMMENDED MINIMUM STOCK REMOVAL ‘iametral Alowance, mm Dia, of Boay, mm | Classes 700-800 | Ciasses 1000—1250 To. Over Include 00 — 125 075 150 125 — 190 105: 224 19.0 — 250 140 3.00 250 — 375 170 375 375 — 500 224 5.00 500 — 630 2.80 630 63.0 — 900 555 8.00 ‘Most automotive torsion bars have hexagon forged ends that are neither ground nor machined. Experience has shown that such torsion bars manufactured of SAE 15B62H (H15621) and SAE 5160H (H51600) steels re~ quire no annealing after upsetting and forging prior to heat treating. D, Heat Treatment In heating for hardening, bars should be adequately soaked prior to quenching. Highly stressed bars should bbe heated in a controlled atmosphere furnace, in a neutral salt bath, or by electrical methods to minimize pitting, scaling, and surface decarburization. This is mandatory if either the bar body or serrations are finished beforeheat treatment, Quenching should be done in oil at a temperature of 40—80°C with rather vigorous agitation Bars should be removed from the quenching medium while still warm and transferred immediately to the tem= ppering furnace to avoid cracking. The actual process must be checked metallurgically to insure that the hardening transformation is complete. To avoid the necessity of straightening, bars are frequently heated and quenched while hanging in a vertical position or quenched in a machine which spins them during the cooling. Periodic checks should be made of the as-quenched hardness in cross sections of the bars as well as of the tempered parts. ‘Adequate metallurgical tests should also be performed to check for retained austenite, decarburization, and proper microstructure. The presence of decarburization, retained austenite, or nonmartensitic structure will result in decreased fatigue life. Hardness of highly stressed bars (Class 800 and above) is generally specified as Rockwell C 47-51. Steels of supe- rior surface and internal quality as processed can be used. at values up to about Rockwell C 54. Ordinary quality steels intended for lower stress (Class 550) applications may be as low as Rockwell C 40. It is desirable to specify, fas high a hardness as possible consistent with quality so as to minimize permanent set in operation, but not 0 high that excessive breakage ie encountered. Hardness testing on the surface of torsion bars must be avoided because any indention becomes a stress riser. Sample coupons or sacrifical bars may be included in the heat treat lot for testing purposes. E, Straightening For most applications, the straightness of a torsion bar is not critical, and requirements would be on the order of perhaps 3 mm Total Indicator Reading (TIR) in 300 mm of length. Itis essential, however, that steps be taken to avoid cracking should straightening be undertaken. In the case of hardened alloy steel springs, it is recom= mended that straightening be performed while the bars are still hot from the tempering operation. F, Shot Peening ‘Springs should be shot peened to the maximum obtaina- ble intensity consistent with cost limitations. In the in- stance of the most highly stressed bars, a minimum inten- sity of 10 C on the body and 7 C on the serrated ends is desirable. Because of the necessity of peening the root radii of the serrations, a smaller diameter shot must be used, than would normally be economical and consistent with the body peening requirement, As a general rule, the shot diameter for reaching into the serration root radius should be approximately one half the radius. One common practice is to peen the spring with a mixture of 60% large and 40% small diameter shot (by weight) using $550 and $280 nominal diameter steel shot, for ‘example, Intensity measurements in accordance with SAE 3443 should be made at regular intervals. The common requirement is for 90% minimum visual surface coverage. It is the intent to provide complete “visual” coverage fr even, in most cases, coverage in excess of 100%, ob- tained by repeated slow passes through the peening equip- ment. Reduced pening intensity even to the extent of 12 A minimum is frequently specified on automotive applica~ tions because of the added cost of obtaining the very high intensities required for critical military usage. If ex- perience so indicates, limited or no peening on the rela- tively lightly stressed ends of some commercial bars may bbe acceptable of pening may be limited to stressed areas such as bends. Recent work in the 1000—1250 Classes of high stress level series torsion bars has indicated the desirability of ascertaining the quality of the shot peened surface through residual surface stress measurements, Use of high speed, x-ray diffraction equipment allows the collection of a sta~ tistically significant number of distributed readings from ‘production sample item. Desirable residual stress levels ccan be ascertained during the development program of the particular item as it is dependent on stress level im- posed and expected cyclic life history. G, Presetting The preferred method of presetting bars is to twist them through a specified angle, release and twist at least ‘two more times to the same end position. The twist angle should correspond to about 0.022 radians strain and the resultant set should not exceed 0.008 radians strain as discussed in Chapter 5, Section 2. Set in excess of the latter amount indicates the physical properties of the bar are subnormal. In large volume operations with hexagon tends, it is often more satisfactory to establish an angle for each lot which will give the desired set condition, and twist one time to the value established. It is generally considered desirable to peen bars prior to presetting, and most processing schedules show that sequence. Nevertheless, the two operations have been suc cessfully used in reverse order. Since the aim for pening. and presetting is the same—to secure favorable residual stresses—the sequence can be influenced by peening inten- sity and bar diameter as well as the possible obliterating effect of one operation on the other. Practical consider- ations concerning processing requirements and preset an- gle tolerances may frequently be the deciding factors for the choice of sequence. 325| H, Corrosion Protection Since corrosion results in pits and other surface imper- fections leading o reduced fatigue life, it may be desirable that precautions be taken in this regard depending on the material and application. Corrosion protection should ‘be accomplished by adequate cleaning followed immedi- ately by applying an acceptable primer. Materials and procedures as defined in Federal Specification TT-C-490 for cleaning and pretreatment methods, TT-P-636 for al- kkyd coatings, and TT-P-1757 for zine chromate primer, have been found satisfactory. Commercial practice has been to apply one or two coats of inexpensive baked enamel or epoxy paint following the cleaning operation. Serrations and other exposed areas should be coated with 8 good grade of No. 2 grease or equivalent. When addi- tional protection to the surface of the bar is required, it is recommended that a double layer of plastic tape be applied. Both vinyl and polyethylene materials have been used. Other materials may be used depending upon the nature of the environment. Rubber sleeves and plastisol coatings have been used, but care must be taken that deterioration of the coating will not permit hidden corro- I, Special Testing [As jis evident from the previous discussions, a wide variety in detailed processing can be employed to produce satisfactory parts. To provide the most economical torsion bars of suitable quality, itis necessary to define require- ments accurately and completely and then substantiate the selected processing by adequate fatigue testing. To establish the confidence limits inherent in a given material and process, or to prove the efficacy of process ‘control, itis necessary to perform fatigue tests on the finished parts. Detrimental residual stresses from im- proper heat treatment or inadequate peening will readily show up in reduced fatigue life, as will poor surface, im- proper presetting, or other defects. It is common practice to test at a higher loading than expected in service to provide an expedited test, as discussed elsewhere in this ‘Manual. Since the fatigue testis the ultimate test of spring quality, it provides the basic data for evaluation of pro- cessing changes and variables. Corrosion testing to evaluate coatings and protective covers at the attaching end may be carried out in conjunc- tion with fatigue testing or as a separate procedure. The following test procedures have been established by SAE, all of which are published in the SAE Handbook: SAE 1406, Methods of Determining Hardenability of Steels SAE J417, Hardness Tests and Hardness Number Con- versions SAE J419, Methods of Measuring Decarburization SAE 1420, Magnetic Particle Inspection SAE J421, Cleanliness Rating of Steels by the Magnetic Particle Method SAE J443, Procedures for Using Standard Shot Peening Test Strip SAE 3448, Surface Texture It will be noted however, that no standard fatigue test- ing procedure is given and it is believed important to impart a note of caution when establishing such a proce- dure. By use of the information contained in this Manual, the engineer can design a test which is meaningful and capable of accomplishment. 2. Presetting A. Load Deflection Curve Presetting is an operation which allows the designer to use, in his calculation, operating stresses higher than the initial elastic limit of the material. It increases the load capacity of the torsion bar in the direction of preset; but reduces it in the opposite direction. The operating torque must always be in the direction ofthe preset torque and should never be reversed. Torsion bars used for clock ‘wise and counterclockwise operation should not be preset. An exception may be made for bars highly stressed in one direction only. A preset in this cirection will benefit the life of the bar even if this results in an actual stress increase in the opposite Cow stress) windup direction However, care must be taken that this actual stress does not exceed the safe limit Theory and practice of the presetting operation have ‘been well established for torsion bars of circular cross section, and the following paragraphs deal with such bars. While the presetting operation can and should also be applied to rectangular torsion bars, the limits of presetting ‘windup and the amounts of allowable set strain on these bars have not been as thoroughly explored. Fig. 5.1 shows the load deflection diagram obtained during the presetting of a torsion bar spring of approxi- mately 58 mm diameter and 1930 mm active length. The loading begins at point 0 and progressing to point I shows a linear increase in torque with windup angle. The slope Of the line O-L indicates ai torsional rate of 765 10° N + mm/deg. After point I, partial yielding occurs and the rate of torque buildup decreases up to point II which represents the maximum specified presetting torque. Upon removal of the load, the bar returns from point I to point III elastically and along a straight line. The distance (III on the abscissa represents a pemanent set of approxi- mately 30 deg. The slope of line II-III now indicates a torsional rate of 740 X 10° N + mm/deg. This drop of torsional rate immediately after presetting appears in most 326TORQUE - 108 N.mm ‘cases but it is of a temporary nature, in other words, 40 60 80 100 WINDUP ANGLE - DEGREES Fig. 5.1—Load-dotlection diagram during presotting the curve O-I-II is identical with that in Fig. 5.1, but the apparent drop in shear modulus will persist for a short time only. See Section 1 of this chapter.) At point U, all the material in the bar was either in the yield range or below. Since the yield point is now at point 1, any subsequent deflection from point IT, for example to point 1V, will be elastic. Then, if the operating torque is lower than the presetting torque, the maximum load torque in service will correspond to a point such as point 1V, and the yield point will not again be reached. Yielding does not occur instantly; therefore, if point IV js to be very close to point Il, the bar should be held at point II for several seconds or it should be loaded to that point several times. B. Stress Distribution ‘The mechanism by which the bar produces the extra load carrying capacity is illustrated by Fig. 5.2. Here the coordinates have been changed to stress versus strain. Strain 7, shown along the abscissa, is determined from y= 04/2 L. Nominal stress S,, shown along the ordinate, is derived from S, = 16 T/z d® This stress has no physical significance above point I, because it is calculated by a formula which is true only. ‘when Hooke’s law holds—that is, below the elastic limit. But the actual stress at the surface can be derived from curve O--II by a method shown by G. B. Upton! as follows: Through a point A on line I-II draw a vertical AD. Lay a tangent to curve I-II through point A, which inter- sects the Y axis at point B. Drop vertically from point A by a distance equal to 1/4 OB to point C. The actual 2G. B, Upton, “Materials of Construction.” Timoshenko, “Strength ‘of Material,” Part 2, should also be const 3271200 1000 800 STRESS - MPa 2 400 200 02} 4] 08 og ° 0: 0,005) (0.010 0.015, 0.020 0.025 STRAIN Fig. §2—Stress-train diagram during presetting stress at the surface of the bar is equal to DC when the nominal stress equals DA and the strain equals OD. ‘Curve I-V is drawn by connecting a number of points. found in this way and is the actual stress-strain curve for a point at the surface of the bar. "We sce that the material has strain hardened from 800 MPa at point I to 910 MPa at point V, which explains part of the increased load capacity. But even more is due to a change in stress distribution. If the material of the bar is homogeneous, any other clement must have the same stress-strain curve as a parti- cle at the surface, Also shown along the abscissa is a linear scale, with the maximum strain OE equal to unity (upper scale. It is generally considered proved that during ‘twisting of a round bar, the cross sections remain plain ‘and undistorted, so that the strain at any point is propor tional to radius. Therefore, at full preset, when the surface is at strain OE, a point at 0.6 radius is under a strain of 0.6 times OE and its stress is shown by the point of curve I-V above that abscissa. ‘The curve O-L-V then shows the actual stresses under full preset windup plotted against radius as abscissa. Lin- car stress distribution would give stresses along the straight line O-V, and the excess of stresses above this Tine accounts for the increase in stored energy. “The ordinate E-I shows the nominal stress at the sur- face when the bar is under full preset windup. When the bar is released, this nominal stress becomes zero, and the actual stress will be reduced by stress E-IL. A trapped stress equal to E-ll minus E-V will then remain at the surface. Similar reasoning applied to points below the surface shows that the trapped stresses at any radius are indicated by the difference between curve 0-F-V and the straight line 0-1. Thus the beneficial trapped stresses for this example do not go any deeper below the surface than where the diameter is 759% ofthe bar diameter which is the point of intersection of O-I-V and O-II. ‘The trapped stresses have been measured experimen tally and found to agree with this diagram. 240. Fuchs and R, L Matton, “Measurement of Residual Strat in Torsion Bar Springs” Proceedings ofthe Society for Experiment Stress Analysis, Val. 4 No. I 328C. Selection of Preset Strain In Fig, 5.3 the curve O-L-Il of Fig. 5.2 is replotted on different coordinates. The abscissa is the same as in Fig. \52 except it is now called “applied strain Ya.” Plotted as ordinate is “elastic recovery strain Yer" which is equal to the nominal stress (curve O-I-II in Fig. 5.2) divided by the shear modulus. The difference between applied strain and elastic recovery strain is the permanent set strain Yat For the range 0-1, the elastic recovery strain is equal to the applied strain, since we are operating elastically ‘or below yield point I. For range Il, the elastic recovery strain is less than 0,025 0.020 015 010 ELASTIC RECOVERY STRAIN - 79, 0.005 0.008) ‘0010 the applied strain or, in other words, the bar has taken fa permanent set. The difference between applied strain ‘and recovery strain is the permanent set strain Yor Yoo ~ Yer “The ordinates of any point along the diagonal O-I-VI fare, by construction, equal to the applied strain as ex plained for the range 0-1. Consequently, the vertical dis- tance between curve 0-F-IL and diagonal 0-I-VI represents the set strain Yee "The nominal stress, identical with that shown in Fig, 5.2, has also been plotted along the ordinate. In this form, the diagram applies to any straight bar of circular cross section, made of the same material and hheat treated the same way, as the original test bar. For 1000 € € a Z ow E 4 z z coo E 3 2 0 20 0.015 0.020 0.025 APPLIED STRAIN = Yap) Fig. 5 9-Elastic recovory of a round bar 3.29 Sennen EEEany other material or heat treatment, a new load deflection diagram similar to Fig 5.1 must be established by a simple torsion test. From this data, the corresponding elastic recovery diagram of Fig. 5.3 can then be constructed. ‘Any point along curve I-II shows the elastic recovery strain or nominal stress which is produced by presetting. up to the corresponding applied strain. The diagram may therefore be used to determine the proper presetting for ‘anew design. Assuming that Fig. 5.3 represents the elastic recovery characteristics of our material and heat treat- ‘ment, we proceed to select the proper presetting strain. Due to production tolerances in chemical composition and heat treatment, the elastic recovery will vary in a range limited by the two dashed lines. If, for example, we want a bar to withstand an operating stress of 1000 MPs nominal, we should preset to a strain Yap = 0.022 so that the minimum expected nominal presetting stress. S, becomes approximately 1070 MPa or about 7% above the operating stress. With this applied strain of 0.022, the set strain should not exceed 0.008. It should be kept in mind that this quantity depends entirely on the chosen. applied strain, which in turn is determined by the maxi- mum expected operating stress. The set strain, as shown by Fig. 5.3 is always determined by the difference between applied strain and elastic recovery strain. Any set strain higher than this difference would indicate that the torque capacity of the bar has been exceeded and it is evidently inferior to the original test bar used to establish Fig. 5.3. Such a bar would be subject to continuous settling and eventual failure. The presetting, therefore, can be used as a final check on the quality of a given material and heat treatment. ‘The relation between strain and windup angle is given 3:30 by @ = 2L-y/d (radians) @ = 114.6 L y/d (degrees) If presetting is used on through hardened bars, there is no reason to stop the process short of an applied strain Of about Yap = 0.022, even if the operating stress is less than 1000 MPa. At this strain, the presetting penetrates to about one-half diameter so that three-quarters of the material is loaded to yield stress. Presetting to a larger applied strain will increase the torque capacity very little more, but the chances of damaging the material are in- creased. Therefore, an applied strain of Yap = 0.022 is recommended for torsion bar springs made of through hardened material of a hardness near Rockwell C 50. Using this applied strain of 0.022, the windup angle 6 = 2.52 L/d degrees. Shallow hardened bars may be preset only to slightly above the yield stress of the mate- rial, Torsion bar springs of the size described in this chapter hhave at maximum windup a stored energy of the order ‘of 16.000 J. This is sufficient to impart a velocity of 900 km/h to a 0.5 kg fragment. The danger to personnel is, therefore great if the bar breaks during presetting, even if a fragment is accelerated only by a fraction of the total enengy. Bars may crack during presetting due to inclusions or seams, and the presetting fixture should therefore contain guards which will protect the operator.Chapter 6 Fatigue testing is an accelerated method of examining springs for design adequacy and for quality control pur- poses. Fatigue lifes expressed by the number of deftection ycles a spring will withstand without failure. It can be tatimated by the use of the fatigue strength diagram shown in Fig. 6.1. Reexamination of the design will be in order if the fatigue tests result in failures which are confined to one section of the spring. ish suitable parameters for a fatigue test Frequently requires the exercise of judgment when the spring application is known to involve random stress amplitudes. As an example, a suspension spring will undergo a large number of cycles of small amplitude near the design load position without failure. Under Fatigue Life ‘greater amplitudes, the number of cycles without failure will be reduced, since the maximum stress as well as the stress range are increased, and both are determining factors in the fatigue life of a spring. Fig, 6.1 shows @ relationship between the initial and ‘maximum stress in the test cycle, and the estimated mini- ‘mum number of cycles to failure, for a solid round alloy steel torsion bar, heat treated to Rockwell C 47-51, shot pened, preset, then subjected to unidirectional loading ‘only, as are most preset bars. ‘The chart covers fatigue tests under this loading condi- tion, with the specified stresses maintained throughout the test. It provides a ready means of estimating minimum fatigue life for any selected method of testing or a means ESTIMATED MIN, CYCLES FOR SHOT PEENED AND PRESET SPRINGS OF H51600 (SAE 5160H) vo [TTT Teta E s es ce se 0 ae srness(wrs) 100 4 INTE MA RANGE, CYCLES E ees 3 is 3 zo ="s 53 # | iii 7 INITIAL STRESS- wes Fa 81—Teon rpg 331TTASLE 6.1—MEDIAN RANKS (PERCENT) FOR SAMPLE SIZES 1 TO 30 Sample Size Rank Rank Oder 1 2 3 « 5 6 7 8 ° 10 Order 1 50000 29.289 20620 15910 12.945 Y910 9428800 Pate 1 2 7o7Nt $0000 38573 91.981 6445 22BA9 © 20.113 17962282 3 79370 61427 50.000 42.141 36412 32.052 gare 25.8573 4 84090 68.619 57859 $0.00 44.015 390895510 5 87055 73555 63.588 55984 $0.00 5169S 6 aaog0 77.181 67.948 60691 Saas 7 oosr2 79887 71.376 s4490 7 8 91700 82038 741az 8 8 92587 83774 9 10 9290010 ‘Sample Size Rank . Rank Ower 1 2 13 “ 6 16 7 18 19 20 Order iy a ee) ee a ee 2 1479 959812579 11702-10940 10270-9478 tsk Ber? «82512 3 23578 21669 ©2005 te? «17-402 16.965 15.422 14581138277 4 92380 29758 27528 25.608 23939 22474 «21.178 20024 8988 TBO 5" a1fe9 37853" 35016 92575 30.48z BSED 25940 ASAI A154 2OETS 6 80.000 45951 42508 39544 36.967 34.705 32.704 30921 20.922 27880 7 Sarr §4089 50.000 46.515 43.483 HOBe3 | 3H4E9 © 4G.371 BAAN) TST 8 67420 62.147 $7492 52485 50.000 46.941 44294 ata2a 3966037710 B 9 76421 70.242 Gage © GDASE —SBSIT 53.059 $0000 47274 ©4830 BSG 10 85204-78331 72472 67.425 63033 59.177 55.768 52726 50.000 47542 10 11 99892 86.402 79.985 74.992 69543 65.205 61531 58.177 55.170 S248 HT 2 94987 e742 av353 76061 Tian! 67296 63629 60.30 57.3742 13 94808 86.208 82568 77825 73.060 69073 G5309 62289 «19 “4 95.169 89.060 89698 78821 74520 «70878 «6720514 15 95.484 89.790 84578 79976 7586 72.119 1S 16 95760 90222 as4i9 810117703318 7 96005 90849 86.173 8194917 18 96222 91322 868838 19 wae 91.749 19 20 9659420 of selecting a desired stress range for the fatigue test. It recommended that this fatigue setup produce at least fan average of 50 000 and preferably an average of 100 000 cycles. ‘There are several considerations why the test setup should be preferred to be such that it will result in a preponderance of cycle figures which are neither too short nor too long. At higher stresses (shorter lives), the scatter of the cycle lives is theoretically reduced, so that fewer test samples will produce a given degree of precision in the estimated life ofthe entire population. However, lower stresses (longer lives) give more realistic results, since they duplicate more nearly the actual service conditions ‘and prevent spring setling during the test. Also, compari- sons between different groups of springs will be more distinct at lower stresses, since different S-N curves tend to diverge the more they approach the fatigue limit (or limiting value ofthe stress at which 50% of the population would survive a very large number of cycles, say n= 10 000 000). In order to establish the fatigue life cycles which are acceptable in any spring design, it is desirable to have road durability tests run over a prescribed course so that fatigue life test data and actual road durability results| may be correlated It must be understood that the fatigue life cycles for any group of springs will vary considerably even under closely controlled test conditions. Moreover, the average life of the tested springs is not sufficient by itself to estab- lish a judgment either on the design or on the material ‘or on the production method which they represent. The relationship between the number of applied cycles and the percentage of springs which failed at these cycles can best be analyzed with the help of statistical techniques which will systematically describe the “dispersion” ot “spread” or “scatter” of the recorded test results. The extent of the scatter will depend upon the consistency ‘of surface condition, fabrication, and the general quality of the springs which are tested The following paragraphs have been more fully treated 332TABLE 6.1—(CONT0) Sample Size Rank Rank Ower at 2 a 24 25 Ey a 2 ey 30 Order fey Se er en ee eg 2 7a6e 7512 «7181 G895 6236572 «6.199 S22 720 Sa 3 12531 11970 11458 1098710559 10.159 9.781 949594 Bete 4 17208 15734 15.734 15088 14.492 19.942 13402 12.958 12517 sate 5 21890 20015 2001518192 18.495 «17.735 «17086 «1648315922 SBT 6 26874 25:84 24297 25.299 22979 21.829 20742 20010 19.208 18691 7 31258 29.859 26580 27408 © 26.324 25.325 24908 23597 22.795 «21988 8 35949 34.394 32.869 31513 30.269 20.120 2805S ©2700 «S| EzeT 8 40.629 3a810 37.147 35621 34.215 32.916 31712 30593 29550 28STE 9 10 45314 4328641431 39,720 Set) 36712 35970 BAI21G2958 3187210 11 $0000 47.762 45.716 43837 42.107 40509 99.027 97.650 36.267 95.168 11 12 54686 52298 50.000 47946 45.054 44.905 4208S 41.178 «30775 Jeaee 12 13 S971 56714 S428 52054 $0.00 48.102 46942 © 44707 «431894176013 4 64057 61.190 $8568 56162 53.948 51895 50.000 40.296 48.502 4505614 15 68.742 65665 62.859 60271 57.892 55605 53658 51764 80.000 489525, 16 79.426 70.141 67.137 6437961699 59491 57.915 $5203 S948 S1eK8 16 17 7409 74616 71420 gB.aB7 65.785 62287 © G09} BRAN © $8817 SHSM 7 18 82791 +. 79.089 75,709 72594 69.730 67.084 G4EI0 62.980 60225 582018 19 8749 89.561 79.985 76701 73.676 70.880 G28 © SATA GIGI? © «6159619 2 © 92.196 88030 84.266 80.808. 77.621 7407S 71.945 © —GUAOT —GT.O41 4832-20, 21 96.753 92.488 88542 84912 81565 78471 75.602 72995 70.450 68.128 «2 2 96098 92.809 8901985507 82.265 79.258 76463 73857 Thane 22 2 97.091 93.105 83447 86.05 © az 79900 77.265 74719 «23 2 97153 93377 89847 86.563 83517 80.672 TOM «24 2 97265 93628 90219 e7042 840788130925 26 97.369 93.861 90564 87.483 axo0s 6 27 97465 94.078 90885 © a7.a96 «27 28 97555 94280 9118628 a 97638 9446829 50 s77i6 30 in Section 3 of Chapter 8 of the Manual on Design and It is possible to calculate the likelihood of similarity Application of Leaf Springs, SAE HS J788 APR80 and the interested reader is referred to that Manual One of the main purposes of statistical analysis is to draw inferences about the properties of a large group (the “population”) from the results of tests on a small group (the “‘sample”). It is, in fact, possible to infer a ‘reat deal from the small sample when it has been prop- erly selected. Only experience can tell the engineer whether the various properties of the sample which can alfect the test result (in regard to material, o production ‘methods, to controls) will also be present in the produe- tion springs (the “population”. This determination is out- side the realm of statistics. However, statistical mathemat ics are based on the inherent assumption that the sample {sa true representative of the population If the entire population were tested under identical test ‘conditions, the results could be shown in graphical form by arranging them in ascending numerical order and plot- ting the cumulative fraction (or percent) of failures over ‘an abscissa of “life cycles.” A sample selected at random, from this population can be expected to exhibit a similar distribution of fatigue life; the larger the numberof springs in the sample, the closer will be the similarity for samples of any given size. Tables are available based fon such likelihoods; as an example, Table 6.1 presents the “median rank” of each test result for sample sizes between I and 30. A rank is assigned to each individual test result corresponding to that portion of the population which it is most likely to represent. Table 6.1 lists percent figures, ‘The median rank line constructed from such data pre- dicts that certain percentages of the population will sur- vive specific cycles-to-failure. But any such estimate may err substantially on either the high or low side. How ‘confident can the engineer be of such an estimate? When judgment based on experience is not considered adequate for a final decision on this question, it will be necessary to construct lines of higher confidence. In many cases, a confidence level of 90 or 95 or even 99% will be required, so that there will remain only a 10 or 5 or 1% tisk of the estimate being either too high ‘or too low. For a chosen confidence level, the life cycles of a given percentage of the population will be found within a certain “tolerance interval.” On the median rank graph this may be represented by a “tolerance band” to either side of the median rank line. Wider bands indi- 3.33cate increased doubt about the line truly representing the population. With a given sample size, the bands will bbe wider for higher confidence levels. With a given confi- dence level, the bands will be wider for a smaller sample ‘Several systems of mathematically organizing the test result data have been established. In the automotive indus- try, the Weibull plot has found wide acceptance because it permits straight-line plotting of the cumulative failure probability versus life cycles on Weibull probability graph Paper. In the Weibull distribution, the relationship between the number of applied cycles and the cumulative percent Of failures at these cycles is expressed by a formula which uses three parameters: 'a) The minumum life, which may or may not be zero. Itis denoted by the letter “a.” When a = zero, the sample data will plot a steaight line on Weibull paper. Otherwise, the points will plot a curved line, but this can be converted to straight-line plotting by a relatively simple technique which is detailed in the Leaf Spring Manual SAE HS ‘3788 APR&0). ') The Weibull slope, which is an indicator of the skew- ress of the distribution. It is called the “shape parameter” and is denoted by the letter “b.” It also is a measure of the scatter of the distribution; a low slope value indicates fa high degree of scatter, and vice versa. The Weibull ‘comes nearest to the normal (or Gaussian) distribution when b = 3.44. ‘The Leaf Spring Manual shows a graph (for Weibull slopes 1 through 12) which locates the percent of failed springs at the life cycles representing the mean of the population, ‘) The characteristic life, which is the 63.2% failu point for the population. It is called the “scale parameter’ and is denoted by 0: 632 = 100(1 ~ 1/e) where: € = 2.7183 (the Napierian base) Example I: Eight springs have been fatigue tested under identical conditions. The result are arranged in ascending ‘order of the failure eyeles and are given rank order num- bers accordingly. In this order, they are assigned median ranks from Table 6.1 as shown in Table 6.2 ‘These points are plotted on Weibull graph paper in Fig. 6.2. Drawing 2 straight line of best fit through the ‘median rank points produces an estimate for the failure rate of the entire population with the parameters b = 2.4 and @ = 170 000. For b = 244, the percent of failed springs at the mean is $2.7 according to the Leaf Spring Manual. At that failure level the median rank line in Fig. 6.2 shows 334 TABLE 62 Order ‘ycles to Medan ‘Spring No. Faure Flank € 1 61 000 8.50 A 2 31000 20.1 F a 114000, 52.05 4 4 135.000, 402 c 5 155000, 55.98 8 6 177000 67.95 6 7 205.000 7389 ° 8 245000 9170 150 000 cycles. At the B-10 level (that is, at the number of cycles where 10% of the population is estimated to fail) the median rank line shows 66 000 cycles. Tin most cases, the fatigue testing of springs will be undertaken for the purpose of comparing different sam- ples, and the probability graph will be expected to convey information on the relative life distribution of the popula- tions represented by these samples. For example, the com parison may involve a sample representing a first design, land another sample representing a second design. ‘When the two median rank lines for the test data of the two samples are plotted on the same graph, they will readily show if the second design promises some im- provement in fatigue life. However, in order to establish if there is a “significant difference” between the two de- signs, it will be necessary to find quantitative vaiues for the degree of improvement (or degradation) between one design and the other. The question is this: “How confi- dently can one say the limited test results indicate that the second design assures an improvement in fatigue life for the entire spring population?” The answer depends not only on the amount of separation between the two plotted slopes, but also on the size of the two test samples. ‘The degree of confidence in the superiority of one design. over the other need not be constant from one quantile level to another. For example, there may be a significant improvement at the B-50 life (50% failure level) without any improvement at the B-10 life (10% failure level). This is partly due to differences in the Weibull slopes, and partly due to the greater width of the tolerance bands at the lower quantile levels. Example 2: It has been proposed that the spring design represented by the sample of eight in Example 1 be re- placed by a new design. Seven springs of the new design hhave been fatigue tested with results which are shown arranged in ascending order and assigned median ranks from Table 6.1 as shown in Table 6.3. For this second plot, it will be seen from Fig. 6.2 that the Weibull slope b = 2.7 and the characteristic life @ = 310 000; the mean life level is at 51.8%, therefore the estimated mean life is 280 000 cycles. The estimated B-10 life is 138 000 cycles. ‘The Leaf Spring Manual explains the method by which‘92170 000 = 2310 609 Fig, 62—Modian rank lines 3.35TABLE 69. ‘Order ycies to Medan Spring No. Failure ‘Rank w 7 132 000 249 ° 2 195 000 2285 K a 233.000 361 t a 275.000 50.00 e 5 315.000 59 4 6 5385 000 7718 N t 449.000 9057 the amount of improvement over the original design (Ex- ample 1) can be reduced to percentage figures at certain degrees of confidence. The result is in this case that with 90% confidence a 32% improvement can be expected at the mean life level, an 8% improvement at the B-10 level 3.36| | | | | t f + yl k y dl Chapter 7 Torsion Bar Spring Applications 1, Square Torsion Bar for Boat Trailer ig. 7.) Fig. 71—Square torsion bar, boat taller Specifications Installation: Length of lever: Static load: Static deftection: Position of lever at static load: Angle of lever at jounce position (limited by component interferences): Maximum torsion bar length: Maximum torsion bar size: Boat trailer suspension: ‘two transverse bars, one fon each side of vehicle; see Fig. 7.1 for schematic layout R= 450 mm ounce = —12.5 deg L = 550 mm (= 25 x 25 = 625 mm? Material: UNS 41500 (SAE 4150) Hardness: 415-461 Bho Shear modulus: 76.0 x 10? MPa Determined from Calculations Total deflection to jounce position = 175 mm Phounce = 3182 N Calculation Determine static torque: Pyatie X RX 0S due 1000 x 450 x 0.9239 31.5 X 10°N + mm f+ f = Rsina +R sin B = 100mm Trae Therefore sin g = 100 + 450 X sn 225 deg 6049 B= 37.2 deg Total deflection to jounce position =.R sin & + R sin djance 450 x 0.6089 ~ 430 X 0.2164 = 175 mm Torsional spring rate: Tae _ S315 x 10° B+ Qastic 14.7 N-mm = 566 x 66x 4 3240 X 10° N + mm/radFrom kr = as we obtait Trounee = 56 600 (37.2 — 12.5) = 1398 x 10°N » mm Tiounce = Pounce X RX 08 djounce 1398 x 10° 450 x 0.9763 ON Pounce From Fig. 22: q2 = 0.208 3 = 0.141 ‘Assume t = 20.0 mm r 3240 x 10° = 529 mm a Tate _ 8315 X10 Taaue _ $315 * 10" M ‘mat? 0.208 x 20) ~ “0? MPa Steaua = = 36:600.37.2 ~ 12.5) 0.208 x 20° _ 1398 x 10° 1.664 X 108 840 MPa, Note: This high stress is considered ucceptable in thi ~ application as experience has shown that the jounce posi tion is attained infrequently during operation. 2, Hexagonal Torsion Bar for Truck Tilt Cab ig. 7.2) Specifications Installation: Hexagonal torsion bar for cab- over-engine truck cab counter-balance; torque re- quirement and maximum in- seribed diameter given; se- cured to chassis on one end, other end torqued. Torque required (T): 2050 x 10° N + mm under tilt 1880 10° N ~ mm under tilt 8 (Gee Fig, 72) Windup angles: a = 235 deg B = 215 deg Torsional rate desired: (2050 + 188010 235 + 215 N-mm ke = 6 = 873 x 10 N- mm = 5000 x 10 rad Maximum bar size al- towable: 28.0 mm across hexflats Material UNS G51600 or G41400 (SAE 5160 or 4140) Hardness: 415-461 Bhn Maximum allowable stress in either di- tection (S;): 725 MPa Shear modulus (G): 76.0 X 108 MPa ‘Data Determined from Calculations Bar size selected: L = 683 mm active length S, = 698 MPa under 2050 x 10" N - mm torque N= mm ‘Actual torsional rate (kz) = 87.3 x 10 NM Calculation Using the formulae for stress and for windup angle with an hexagonal bar from Timoshenko:t T ie 0.188 d? where d’ = diameter of inscribed circle Therefore: T oa, 2080 x 10 ores x 725 = 15.04 x 10° = 247mm a Select 25 mm hexagonal bar (4 = 25 mm) TL : O.11Sd*G "Timoshenko, “Strength of Materials” Part 2, Chapter VI 3.38ae 1600 +— TORQUE REQUIRED om 5 ACTUAL- 6 vo | 1oRauE From NW) HEXAGONAL TORSION BAR TORQUE (103 N.mm) b 8 800 ~1200 1600 =2000 0 10 2 30 40 CAB TILT, DEGREES Fig. 7.2—Hexagonal torsion bar—truck tit cab 339Therefore: ous d*G VU kr QUIS x 25¢ 76.0 x 10° 3000 > 108 = 683 mm Then: 2050 x 10 ee 0.188 x 259 CLs) 3. Laminated Torsion Bar for Truck Tilt Cab (Fig. 7.3) Specifications Installation: VY Torque required (T) per side: Windup angles: Torsional rate desired (kr) per side: Length per side (L): Material: Hardness: Maximum allowable stress in either direction Sd: Shear modulus (G): Torsion bar for cab- coverengine truck cab counterbalance; torque requirement and bar length given; torsion bar secured to chassis at cen- terline, torqued at both ends. 1835 < 108 N « mm under 770 x 10° N - mm under tit B (see Fig 7.3) = 39.5 deg = 0.6894 rad B= 15.5 deg = 0.2705 rad (1835 + 719)108 395 + 155 = 47.36 x 10° N+ mm/ deg 2714 x 109 N > mm/ rad 700 mm G51600 (SAE 5160), leat spring stock 415-461 Bho, 725 MPa 760 x 10? MPa Data Determined from Calculations Width (w) of all laminae Number of laminae (Ni) = 50mm 2 Thickness of laminae () = 9.50 mm Maximum working stress Ss) Actual torsion rate (kr) 1 MPa 47.43 X 10° N + mm/deg per side or 94.86 10° N - mm/deg total Catto ea 8 BEES ry G 760 x 10° = 25.000 (required) 5 = Bate mL Assuming m2 = ns (Fig. 2.2) ms x 700 ap KI = 9.69 mm 76.0 X 10° X 0.6894 2 ta ‘Assuming w = 50 and t = 9.50 (nearest standard leaf spring size below 9.69) w/t = 50/950 = 5.26 ts = 0.292 = 25000 awe 25000 Ne oxp7 x so xo ~ 1997 Use Ni ‘Then se BE O22 X 9.59 X50. 16 X 10 a = 2718 x 10° N - mm/rad = 47.43 x 10° N + ‘mm/deg (Satisfactory because itis within 0.15% of the specified rate) mate _ aL (1/72 = bi Fig, 22) 894 X 9.5 X 76.0 X 10° _ = ail = Til MPa 4, Torsion Bar with Integral Torque Arm The example which follows is similar to torsion bar springs used in the front suspension of some compact and mid-size passenger cars. Two bars are usually so ar~‘4000 3000 Nef tora Torque Required = 20 E 2 % ACTUAL TOTAL mm TORQUE FROM 3 1000 LAMINATED « TORSION BAR e 500 * a ° -600 —4 1000 1500 x ° 10 20 30 40 60 0 CAB TILT, DEGREES AD 700 mm +] 2 LAMINAE - 50 mm WIDE, 9.50 mm THICK, Fig. 7--Laminated torsion bar—truck tit cab ranged that their torsionally stressed portions are posi- tioned transverse or perpendicular to the centerline of the vehicle. The torque arm portions are substantially Parallel to the centerline of the vehicle. The example is, simplified in that only two diameters are used in the tor- sional part of the spring. In practice, both ends are ex- truded to at least three different diameters to improve the efficiency of the springBar deflection: 4 = Static deflection of bar P, = Static load on bar (approximately perpendicular {0 arm in design position) ka = Rate of bar from load Px Pa oD po ke . if rs -. lity freore-- 1 a a -«} 2210, + [aaron (vt = at) Gh feta top. fy [poe tet sa-e-ewr 85] + e+ o@t- olf wa), — ae Gh fo+et qe 1 —" a Where Loatbtett (ea. 3) ic Clee ae (eq. 4) ahh (ay wn () oo 11 = Correction for end transition (see Chapter 3, Section 3) Load on bar at jounce Py = Pat Bika, (eq. 6) {8s = Jounce travel when ay (angle of arm travel con- sidered to be small) Windup angle Pa 1 180 ge = BA xg t= HO 1 or Ro ane ae 16P FN “= cag +R — Nea 6) sin a + ([Bip + weno x ang] + [RU ~ 00s 6) + N sin oF} (ea. 9) 3.42 Stress at cushion point M=P@ +>) T= p@ — aye 5 = 16P [arortery ae + @ ~ ann} (a. 10) Stress at step point “D Pla + b + e(1 — qd} por ae isP s kx HE {fa ++ ec -a] +t +b +e — a ha »] _ . (eq) K, = Stress concentration factor at step ~ 1.1 Design of bar as shown in Fig. 7.4: Material: SAE 15B62H (UNS H15621) Hardness: 444-495 Bhn Preset: 7% over max load Shotpeened: 2 passes at 90% coverage Dimensions: (Ref. Fig. 7.4) a= 122mm b= 108 mm € = 90mm f= 595 mm 915 mm 420 mm, 75 mm 4, = 32mm 4; = 285 mm = 63.5 mm 207 x 102 MPa G = 79.3 x 10° MPa 9s 4 = gp gg = 133577 2 a= (Gs) = 1, = 51471 mm Jy = 102943 mm* n= 05D (min) + (1 Vt = 05+ 12d: +02 +09 dy = d (0.60 + 0.18) = 285 - 078 =22 mm with D/d = 1.20 (min) vo =08ine rate of bar from eq. 2: Determis ‘bar from eq. 2: oe eee 03: 1 1 ke 3 207-10 - S147 3577 033577 a2 — 1225 » lw =n ae wae + + [108 + 90 + (122 + 108) - + (15 — 595 + 1.60 + (595 — 22)] ai a2 ‘A-POINT OF LOAD APPLICATION (C-SUPPORT CUSHION: E-ADJUSTABLE ANCHOR A JOUNCE [—-JOUNCE TRAVEL, ANGLE a< — 7 STATIC ow <| WINDUP . ANGLE 2 6 S Ar LOAD - ZERO VS Fig, 7.4~Torsion bar wit integral torque arm 3.431 : = 12 io [eo 319635 1g 230% = 1.33577 + 199.7807 os] 0.33577 + __ 161516 HS ita aie3.38 - 1 1 (16.25231 + 4.78423): 108 (74.088 ~ 1.815848): 10° + 0.33577 31 9635-10" + 19.7843 - 10° + 11148. (2.27215 + 62.65164) + 10° + 0.022087 319635 + 10 : 134.92379 sraess + 0.022057 = 0.004221 + 0.022057 = 0.026278 ky = 38.05 N/mm Determine load capacity of bar by rearranging eq. 11, and using stresses for point “D” at jounce load: S) = 1240 MPa 930 MPa 1 [122 ¥ 108 + 901 — 1.33577] Fea Foe + 0g — STH 7 ae — as SIS 799.7807 + [199.7807 + 161 S16 5123 815 = 199.7807 + 448.8077 Pq = 7900 + 930/1240 = 5925N 1900 N Allowable Travel: 8; = 7£ (Py ~ Pa) 1 Fygs (7900 — $928) = 51.9 mm 5925 ~ 38.05 ” @20* = 122 0 aaa Windup angle: aw ‘Stress in Bend of Bar Point “B” ‘The maximum value of principal stress was found to be at @ = 22 deg into bend (See Fig. 7.) oot (2) = nt ar N = 420 — 75 tan (73.1/2) = 364 mm Using a 9 Stes a une 16 0 Ff 5 = [ites 08 va x in22 de % inane} + (f= 4 (15 — 368 tan 22 deg) sin 22 dg] + [15(1 ~ 008 22.dég) + 264 sin 22 ea} + [265.59 + (5.461 + 13635997] 930 MPa Stress at Bushing Point “C” Using eq. 10 Stress at jounce 16 x 7900 , = 16x 1900 + ((122 + 108° sy = SR ltt22 + 108) + (¢ . + (420% = 1225)9) = 851 MPa ‘The allowable stress at points “B” and “C” depends fon the margin of safety required in this section of the bar. 3.44STRESS AT POINT "F” ON ARC OF BEND MOMENT ON RADIAL SECTION AT . om

CHAPTER 7—PHYSICAL APPLICATION CONSIDERATIONS T, Desirable Pneumatic Spring Mounting Geometries ‘A. Reversible Sleeve and Reversible Diaphragm Types - B. Convoluted Types... CC. Paralelogram Suspension Linkage D. Multi-Axle Pneumatic System . 2. Undesirable Pneumatic Spring Mounting Geometries APPENDICES ~~ A, SI Units and Symbols for Terms Used .... B. Glossary of Unique Terms and Phrases Used in the Manual C. Data Record Form for Pneumatic Spring Applications D. Rate and Frequency Calculation Form for Pneumatic Spring Applications . S‘This manual has been prepared to assist the engineer and the designer to have a better understanding of the basic principles, types, and uses of pneumatic springs. In addi- tion, it has been designed to serve as a useful guide in the selection of pneumatic springs for specific applications. The need for such a manual has been dictated by the popularity ‘and growth of pneumatic spring applications. “The use of pneumatic springs in commercial applications thas been a relatively recent development, but the idea of such a spring is not a new concept. “The earliest available records of a practical approach to pneumatic springs is a patent granted to John Lewis in February 1847. Prior to 1910, Benjamin Bell was engaged inexperimental work on sleeve-type pneumatic springs with various piston shapes. The work and ideas put forth by these two gentlemen were made possible through the combined efforts of an Irishman and an Englishman some 300 years before that time. Robert Boyle, the Irishman, published a paper in 1660, “New Experiments—Touching the Spring of Air”, setting forth the law, “absolute pressures and volumes are in recip- rocal proportion when temperature remains constant”. Eigh- teen years later, Robert Hooke, the Englishman, formulated the relationship between force and elongation of elastic solid ‘materials ‘The first serious work toward adapting the pneumatic spring to the automotive industry was carried out by the Firestone Tire and Rubber Co. in the early 1930's. By 193: experimental Buick and Plymouth cars were equipped pneumatic springs. These were soon followed by installa- tions on other vehicles, such as Studebaker, Chrysler, Ford, Lincoln, etc. In spite of this interest, the pneumatic spring 44 Chapter 1 Introduction and History ‘was not adopted in the late 1930°s primarily because of costs ‘and the tremendous improvements being made with steei springs and suspensions in general In 1938, General Motors Corp. became interested in a new suspension with pneumatic springs for its buses. Work- ing with Firestone, the first buses were tested in 1944 and first production was realized in 1953. This breakthrough triggered the growth and development of the pneumatic spring into the many new fields and applications that are in use today. In addition, new concepts in pneumatic springs hhave evolved to meet the requirements of these new and changing needs. During the mid-S0's Goodyear developed and patented the rolling lobe type air spring. The rolling lobe air spring load-deflection characteristics can be greatly influenced by piston contour. “The pneumatic spring has been able to make inroads into uses formerly reserved for more conventional and better known springs because of some unique characteristics and versatility. The more common and better known advantages of pneumatic springs are: Controllable spring rate Adjustable load capacity . Simplicity of height control . Reduction of friction 5. Nearly constant frequency with respect to load varia tions Why and how pneumatic springs can offer these advan- tages and meet the needs of modem industry will be de- scribed in the following chapters which cover: basic princi- ples, types, design problems, and special uses of pneumatic springs.Chapter 2 Basic Principles of Pneumatic Springs 1. General Discussion ‘The pneumatic spring is basically a column of confined gas in a container designed to utilize the pressure of the Bas asthe force medium of the spring. The compressibility of te gas provides the desired elasticity for suspension use, The pneumatic spring's ability to support a mass depends upon its effective area, which is a nominal area found by dividing the load supported by the spring by its gas pressure at any given positon. The effective area is a function of deflection. Whether it remains constant, increases, or de- creases is governed by the design of the spring and its components. The spring rate is the result of change in effec- tive area and the change in gas pressure as the spring is deflected. The gas pressure varies with the speed and mag- nitude of deflection; for a unit of deflection, the pressure and, therefore, the spring rate will be different for isother- mal, adiabatic, or polytropic processes. Preumatic springs provide an adjustable spring rate, ad- justable load carrying ability, simplicity of height contol, and a low friction action. Pneumatic springs are adaptable for light or heavy suspension appli 2. Compression Processes For a specific spring design, the minimum pneumatic spring rate occurs under isothermal compression conditions and the maximum spring rate occurs with adiabatic com- Pression. The polytropic rate varies between the isothermal and the adiabatic. The isothermal rate results when all the heat of gas compression escapes so thatthe gas remains at constant temperature. The isothermal rate is approached when the, spring is deflected. very slowly to allow time for the heat to escape, the gas temperature remains constant, and the gas pressure rise is minimum. ‘Adiabatic rate occurs when all the heat of compression is, retained within the gas. This condition is approached during ‘apid spring deflection when there is insufficient time for the heat to be dissipated. The higher temperature of the gas results in a higher gas pressure and, therefore, a higher spring rate. ‘When the heat of compression is partially retained within the gas, a polytropic rate results. This occurs during most ‘normal spring deflections and produces neither isothermal iabatc rates, although in normal use itis much closer 'o the adiabatic situation. 3. Constant Volume Pneumatic Springs Pneumatic springs which maintain a relatively constant Volume at a given operating height regardless of static load. 43 Or gas pressure are referred to as constant volume pneumatic springs and are the more common type in use at this time. At a given height, the load-carrying ability and the spring rate are varied by changing the pressure ofthe confined gas. With this type of spring, an external source of compressed {225 is needed to maintain the spring ncight as the load on the spring is changed. The natural vibration frequency of the ‘constant volume pneumatic spring remains more uniform with changes in load than does the natural frequency of the. ‘constant mass pneumatic spring. 4. Constant Mass Pneumatic Springs Pneumatic springs which use a fixed mass of gas as the clastic medium are constant mass pneumatic springs. A ‘given amount of gas is sealed in the system and remains constant for all conditions of load or deflection. As the load on a constant mass pneumatic spring is in- creased, the gas volume is reduced and the spring rate in- ‘creases. Conversely, when the load on the spring is reduced, the gas expands and the increased gas volume results in a reduced spring rate. Thus, the natural vibration frequency of a suspension system using a constant mass pneumatic spring. increases as the load on the system increases, Pneumatic springs which are not connected to 2 gas source with height control or other valve arrangements, often called “locked in” systems, are classed as constant ‘mass pneumatic springs, as are most hydro-pneumatic springs. 5. Basic Cylinder and Piston Springs A. Cylinders with pistons can be used as preumatic springs but they have several major drawbacks: (1) Sliding friction transmits significant forces through the spring. Short impulses are especially detrimental, (2) It is difficult to maintain zero gas leakage past the Piston and rod seals for the desired life of the unit, (3) Clevises are required at the top and bottom for most ‘mountings. (G) The effective area cannot be manipulated, (5) The piston and rod guide present wear problems. B. An advantage is that high operating pressures may be used, and the unit can combine load-carrying and damping functions,