1992 IEEE FREQUENCY CONTROL SYMPOSIUM

of S- TCXOS
Fr equency Temper at ur e Behavi or and I ts Hyst er esi s
Usi ng New Shape GT Cut Quar t z Cr yst al Resonat or s
Hi r of umi KAWASHI MA and Ken j i SUNAGA
Sei ko El ect r oni c Component s Co. , Lt d.
Tochi gi - shi , J apan 328
Abstract
Thi s paper descr i bes f r equency t emper at ur e
behavi or and i t s hyst er esi s of sel f - t emper at ur e
compensat ed cr yst al osci l l at or s ( S - TCXO s)
usi ng new shape GT cut quar t z cr yst al
r esonat or s. T h e obj ect of t hi s paper i s t o
cl ar i f y t he f r equency t emper at ur e behavi or , i t s
hyst er esi s and agi ng char act er i st i cs of t he S-
T C X Os .
Fi r st , a S - T C X O i s successf ul l y ob-
t ai ned wi t h f r equency devi at i on of t 30/ - 150ppb at
best over a wi de t emper at ur e r ange of - 30 t o
t 85 "C, needl ess t o say, wi t hout any ext er nal
t emper at ur e compensat i on such as t her mi st er and
so on.
T C X O s i s ext r emel y smal l f or t he excur si on of
f a1 1- t emper at ur e to r i se- t emper at ur e, but occur s
somewhat for t he excur si on of r i se- t emper at ur e to
f al l - t emper at ur e. T h e var i at i on quant i t y f or
each S- TC XO i s, however , l ess t han 0. 38ppm
i n f r equency devi at i on f or t he t emper at ur e
excur si on of - 45 t o t 90 and - 45 "C, whose
excur si on shows t he out st andi ng hyst er esi s.
F i nal l y, agi ng quant i t y af t er t he pr e- agi ng
f or about t wo weeks i s al so f ound t o be 0. 33ppm
Second, t he t her mal hyst er esi s of S -
demands excel l ent f r equency t emper at ur e behavi or ,
and si mul t aneousl y, a f ur t her mi ni at ur i zed
osci l l at or i s r equi r ed wi t h l ow consumpt i on
power. As an osci l l at or sat i sf yi ng t hi s r e-
qui r ement , i t was al r eady r epor t ed[ l l , 121 by one
of t he aut hor s t hat a sel f - t emper at ur e
compensat ed cr yst al osci l l at or empl oyi ng a new
shape GT cut quar t z cr yst al r esonat or G- TCXO)
[ 31- [ 71 i s ver y avai l abl e. I n addi t i on, many
paper s wer e r epor t ed on t her mal hyst er esi s of
r esonat or s [ 8 1- [ 12 1.
The obj ect of t hi s paper is t o cl ar i f y t he
f r equency t emper at ur e behavi or , i t s hyst er esi s
and agi ng char act er i st i cs f or t he S- TCXOs.
F i rst , i t i s shown t hat a S- TCXO empl oyi ng a new
shape GT cut quar t z cr yst al r esonat or i s
successf ul l y obt ai ned wi t h f r equency devi at i on of
+30/ - 150ppb at best i n a t emper at ur e r ange of - 30
t o t 85 "C wi t hout any ext er nal t emper at ur e
campensat i on and i s ext r emel y excel l ent i n
f r equency t emper at ur e behavi or . Second, t he
t her mal hyst er esi s char act er i st i cs of f i ve
S- TCXOs wi t h f r equency devi at i on of l ess t han
0. 7ppm i n a t emper at ur e r ange of -30 t o t85"Car e
shown. A n exami nat i on of t he t her mal hyst er esi s
i s al so per f or med f or t he t emper at ur e excur si ons
of ( a> to (d):
(a) +25" C++90" C- - 45" C~+25" C
at a mean val ue of ei ght sampl es f or one year and
(b)
+25"C+ - 45C- +90" W +25C
- 45C- +9 0C- - 45C
+9 0 "C+- 4 5C- +9 0C
t wo mont hs, and af t er t he pr e- agi ng f or about si x
mont hs, t he quant i t y i s 65ppb at a mean val ue f or
(c)
ei ght mont hs.
( d )
As a r esul t , t he t her mal hyst er esi s occur s some-
1. I ntroduction what f or t he excur si on of r i se- t emper at ur e t o
f al l - t emper at ur e, but i t s var i at i on quant i t y i s
ver y smal 1 and t he pr esent S- TCXOs ar e f ound t o
osci l l at or ( TCX0) t hat is used as sour ce of a have har dl y pr obl em on t he t her mal hyst er esi s,
f r equency st andar d of pr i nci pal l y handy- t ype even when t he f r equency devi at i on of 52. 5ppm i s
comnuni cat i on equi pment and consumer pr oduct s demanded i n t he t emper at ur e r ange of - 30 t o t 85"C.
Recent l y, a t emper at ur e compensat ed cr yst al
434
0-7803-0476-4192 $3.00 0 1992 IEEE
Fi gur e 1 r epr esent s a new shape GT
cut quar t z cr yst al r esonat or and i t s suppor t i ng
means. As descr i bed det ai l edl y i n t he r ef er ence
[ SI , t he pr esent GT cut quar t z cr yst al
r esonat or consi st s of a vi brat i ori al por t i on and
Fig. 2 Ci rcui t diagram of a S-TCXO.
OUTPUT
4c
3 . Ci rcui t Diagram and
Osci 1 lat i on Frequency
Fi gur e 2 r epr esent s a ci rcri i t di agr amof a
S- TCXO. T h e ci r cui t consi st s of a CMOS
i nver t er , dr ai n capaci t ance Cd, gat e capaci t ance
Cg, f eedback r esi st ance R f , dr ai n out put
r esi st ance Rd, vol t age sour ce V nn and a new shape
GT cut quar t z cr yst al r esonat or , and a si gnal
i s out put t hr ough a buf f er . Next , an
osci l l at i on f r equency i s der i ved f r om an
equi val ent ci r cui t of t he pr esent ci r cui t
di agr am, but t he der i vat i ve pr ocedur e i s
expl ai ned si mpl y because t he det ai l ed pr ocedur e
was al r eady- descr i bed i n t he ref erences[ l I , L21.
whi ch i s a coupl i ng r esonat or can be expr essed by
a summat i on of real par t Rei ( r esi st ance par t , i =l ,
2, i =l : Pr i nci pal vi br at i on, i =2: Subvi br at i on) and
i magi nar y par t Xe( r eact ance part). Ther ef or e,
t he ci r cui t i n I ;i g. 2 i s r epl aced by an el ect r i cal
equi val ent ci r cui t . Fr om t he ci r cui t , when
an angul ar f r equency i s t aken as w , Xe i s gi ven
as
1 ( 2)
A CMOS i nver t er f or a smal l si gnal can be __ 1 _. -~ +- 1 (14- Rei
expr essed by vol t age sour ce f l Vg ( f l : amp1 i f i ca- CI Cg Cd rd +Rd
t i on r at i o, Vg: gat e vol t age) and dr ai n r esi st -
ance r d, as i s wel l - known. Al so, t he i mpedance
of a new shape GT cut quar t z cr yst al r esonat or Mor eover , Xe f or a coupl i ng bet ween pr i nci pal -
435
and sub- vi br at i ons of t he GT cut r esonat or
whi ch i s t aken as a capaci t ance coupl i ng, was
al r eady- cal cul at ed i n t he r ef er ence[ 11. T he
r esul t i s, t her ef or e, empl oyed her e, al so, Rei of
equat i on( 2) i s cal cul at ed under t he condi t i on
t hat a coup1 i ng bet ween pr i nci pal - and sub-
vi br at i ons is ver y weak. Consequent l y, Xe and
Rei ar e, r espect i vel y, gi ven as
+- c , ( - 2 - 1 + - ) ) ] o2 c 2
CO w2 Cm
( i =l , 2) (4)
c 2
Rei =Ri (It L,
C L
where wI , wr : an ul ar f requenci es of pri nci pal - and
sui - vi brat i ons uncoupl ed
CO : shunt capaci t ance
Cm : coup 1 i ng capac i t ance
C,, Cz : mot i onal capaci t ance of
RI , Rz : seri es. resi st ance of pri nci pal - and sub-
pri nci pal - and sub- v i brat i ons
vi brat i ons
Accor di ngl y,
t he osci l l at i on f r equenci es f +_ ar e gi ven as
subst i t ut i ng equ. ( 3) i nt o equ. (l),
* - I
(5)
where,
) f , 2 ('.'fI=01/2T) f , ' Z = ( I + -~
i a
r l l +a- h
Ks2=Km2 ( l+-
1 +U
>2
1 +a-h - .. .~
h=
Rf(rd+Rd) w2CgCd
a=Co/CL r,=Co/Cl
b=Cm/CL r2=Co/C2 Cm2
h 2 - C 1C 2
Equat i on (5) gi ves t wo osci l l at i on f r equenci es,
and shows an osci l l at i on f r equency f + of
pr i nci pal vi br at i on f or a posi t i ve si gn i n f r ont
of t he r oot and al so an osci l l at i on f r equency f-
of sub- vi br at i on f or a negat i ve si gn. I n addi -
t i on, subst i t ut i ng equ. (4) i nt o equ. (2) and
t aki ng (Co/ Cd)2<<1 i nt o account , l oad capaci t ance
C, i s r ewr i t t en as
)
1 R i
L A - ( t +
rd +Rd
2 Ri
(6)
Cd rd fRd
As is appar ent f r omequ. (51, i t i s easi l y pr ed-
i ct ed and under st ood t hat f r equency t emper at ur e
behavi or var i es wi t h CL, because fk i s gi ven as a
f unct i on of Ks and al so t he Ks i s gi ven as a
f unct i on of t he CL.. By det er mi nat i on of an
out i mumval ue CL, a S- TCXO ext r emel y excel l ent i n
f r equency t emper at ur e behavi or i s obt ai ned
wi l l be descr i bed l at er.
4 . Results and Discussion
Accor di ng t o t he f ol l owi ng pr ocedur e,
as
t he
r esul t s ar e expl ai ned and di scussed. F i rst ,
el ect r i cal equi val ent ci r cui t par amet er s of
pr i nci pal - and sub- vi br at i ons f or new shape GT
cut quar t z cr yst al r esonat or s mount ed on cer ami c
pedest al s t hat ar e used i n t hi s st udy, ar e shown.
Second, f r equency t emper at ur e behavi or of S-
TCXOs wi t h t he af or ement i oned r esonat or s i s shown
and t he S- TCXOs ar e f ound t o have a f r equency
devi at i on of l ess t han 0. 7ppm i n a t emper at ur e
r ange of -30 t o t85"C. Thi r d, t he t her mal
hyst er esi s char act er i st i cs f or t he obt ai ned S-
TCXOs ar e exami ned f or t he t emper at ur e excur si ons
of four sequence. Also, t he t her mal hyst er esi s
of a novel GT cut r esonat or wi t h no pedest al i s
compar ed wi t h t hat of t he GT r esonat or s wi t h
t he pedest al . Fi nal l y, l ong t er m agi ng
char act er i st i cs one year ar e shown over.
4. 1 El ectri cal Equivalent
C ircui t Parameters
Tabl e 1 r epr esent s el ect r i cal equi val ent
ci r cui t par amet er s of pr i nci pal - and sub-
vi br at i ons f or new shape GT cut quar t z cr yst al
r esonat or s t hat ar e used i n t he pr esent S- TCXOs.
Sampl es No. 1 to No. 5 ar e empl oyed f or t her mal
hyst er esi s char act er i st i cs and sampl es No. 6 to
No. 13 ar e empl oyed f or agi ng char act er i st i cs.
A s i s obvi ous f r omTabl e 1, t he var i ous val ues
ar e suf f i ci ent l y sat i sf act or y as a r esonat or .
Her e, capaci t ance r at i o ri i s def i ned as Ti =
C, / Ci ( i =L 2).
436
Table 1 Values of el ectri cal equivalent ci rcui t
parameters used in the S-TCXOs.
0.5-
2. 46
No.3
...
0 0 .. ..............
Prin. 2 0997 68.5 0.98 5.86 415 189
Sub. 1.8940 451 2.73 2.59 938 72
1 3 2. 43
0.5-
R o r
-0.5
Lrl
4.2 Frequency Temperature Behavior
No.4
...... ..............
......
..........
F i gur e 3 r epr esent s t he f r equency t emper -
at ur e behavi or of f i ve 5-TCXOs empl oyi ng new
shape GT cut quar t z cr yst al r esonat or s. Load
capaci t ance CL i s so sel ect ed t hat t he f i r st
or der t emper at ur e coef f i ci ent a t of pr i nci pal
vi br at i on r eaches zer o appr oxi mat el y at r oom
t emper at ur e. A s a r esul t , t he f r equency devi a-
t i on A f / f i s easi l y f ound to be l ess t han
0. 7ppmi n t he t emper at ur e r ange of -30 to t85"C.
Speci f i cal l y, t he f r equency devi at i on f or sampl e
No. 1 i s t 30/ - 150ppb i n t he f or egoi ng t emper at ur e
r ange. Thi s is because t he pr esent GT cut
r esonat or has a val ue of t he second or der
t emper at ur e coef f i ci ent P+=O appr oxi mat el y and
al so a ver y smal l absol ut e val ue of t he t hi rd
or der t emper at ur e coef f i ci ent 7 + such as I O- ''
to Thi s val ue i s one hundr edt h smal l er
t han t hat of an AT cut quar t z cr yst al r esonat or
whi ch i s wi del y used. Wher e t emper a t 11 r e
n MEASURED
E 0 51 Sample No.1
@. .
__n---.
...............
\ 0.51 No. 2
................. -.- ............
- 0. 51* , . I . I . , , I . ,
-60 -40 - 20 0 20 40 60 80 100
TEMPERATURE ( "c 1
Fig. 3 Frequency temperature behavior of f i ve
S-TCXOS.
measur ement s wer e per f or med at t he excur si on of
l ow t emper at ur e ( - 30 %) t o hi gh t emper at ur e
(t85T).
~- 4 . 3 ~~ Thermal Hysteresis Characteristics
Fr equency hyst er esi s ver sus t emper at ur e i s
next shown. T he t her mal hyst er esi s i s exami ned
f or f our t emper at ur e excur si ons of (a> to (d) as
i s descr i bed bel ow, and al so measur ed f or
f r eqoency- st abl e t i me of f i ve mi nut es ver sus t em-
per at ur e change of 3% and f or shor t er f r equency-
st abl e t i me of t wo mi nut es.
(a) I 25% -*+90%--45%++25%
(b)
CC) - 45%-+90%--45"C
(d) +9OoC--45"C-+90%
+25%+- 45%-+ 90%- +25C
F i gur e 4 r epr esent s t he t hermal hyst er esi s
char act er i st i cs of f i ve S-TCXOs measur ed at an
aver age t emper at ur e speed of 0. G"C/ mi nut e (3%/
5mi nut es) f or each t emper at ur e excur si on.
F or t he excursi on(a), t he sampl es showeach t he
t her mal hyst er esi s of 0. 13 to O.2ppm at maxi mum
at t hc excur si on of r i set emper at ur e( +25 "C-
-I- 90%) t o f a 1 1 - t emper at ur e ( +9 O Y - -t- 25C 1.
I I owcvcr , t he t hermal hyst er esi s is har dl y ob-
ser ved f or t he excur si on of f al l - t emper at ur e
437
0.5-
- 0
2 -0.5.
No. 4
-
I
0.5-
a O V = - ~ - L-'-.::::::::-$::
-0.5.
1.0
E -0.5
c4 -60 -40 -20 0 20 40 60 80 100
*.- tt.*c*s*-'
.......
-
-.- =:::::.'::::::,..- . -
--.-=?..---
.................
-*
0.5 N O . 2
- .............
........... -3.
-0. : : +-
n 0 :
-0.5
No. 3
......................
0.5 e-.-.*>
-0.5.-
-
........................
0 - .......... ..-.---::-
.... -
-......- @?=q--e. L.-::::::1.-
2_2 - ---
. . . . . .
& ~ -o.5 o - : p - . . " - <..---*- ............. -- ..*.. --.*._
- - ....... :::t.
g -1.0
w
.......
-1.0
-60 -40 -20 0 20 40 60 80 100
TEMPERATURE ( 'c )
LL -60 -40 -20 0 20 40 60 80 100
TEMPERATURE ( 'c )
(d)+ 90% - 45C + +90%
E 0.5. Sample No.1
a 0
-0.5
\ -1.0-
- -.. W
Q-l
z
Lr. -1.0
TEMPERATURE ( ' C
Fig. 4 Thermal hysteresis characteristics (3"C/5minutes).
( +25"C- - 45"C) t o r i se- t emper at ur e ( - 45 "C- exampl e f or t he wi der t emper at ur e excur si on
+25"C). Next , f or t he excursi on(b), t he ( - 45"C- +90 %+- 45 "C) of r i se- t emper at ur e t o
t her mal hyst er esi s i s har dl y obser ved, si mi l ar to f al l - t emper at ur e. Each sampl e shows t he somewhat
t he excursi on(a), f or t he excur si on of f al l - l ar ge t her mal hyst er esi s of 0. 25 t o 0. 38ppm at
t emper at ur e( +25"C- - 45"C) t o r i se- t emuer at ur e maxi mum as compar ed wi t h t he af or ement i oned
( - 45"C++25"C) , but t he sampl es show each t he
t her mal hyst er esi s of 0. 15 t o 0. 3l ppmat maxi mum
f or t he excur si on of r i se- t emper at ur e( +25 "C-
+90C) t o f a1 1- t emper at ur e( +90"~- +25"C) .
Fr omt he r esul t s of (a) and (b), t he t her mal
. hyst er esi s i s f ound t o occur somewhat f or t he
excur si on of r i se- t emper at ur e t o f al l -
t emper at ur e. I t is, t her ef or e, pr edi ct ed t hat
somewhat l ar ge t her mal hyst er esi s my occur i n a
wi der t emper at ur e r ange of r i se- t emper at ur e t o
f a 1 1 - t emper at ur e. T h e excursi on( c) i s an
438
t emper at ur e excur si on of (a) and (b). T o t he
cont r ar y, i t i s pr edi ct ed f r om (a) and (b) t hat
t he t her mal hyst er esi s har dl y occur s even i n a
wi der t emper at ur e r ange f or t he excur si on of
f sl 1- t emper at ur e t o r i se- t emper at ur e.
T h e excur si on (d) i s an exampl e f or t he wi der
t emper at ur e excur si on( t 9O"C- - 45C- + 90C) of
f a1 I - t emper at ur e t o r i se- t emper at ur e.
T h e sampl es show each t he t her mal hyst er esi s of
0.01 t o 0. l ppm at maxi mum and t he hyst er esi s
becomes ext r emel y smal l as pr edi ct ed. Thus ,
n
0. 5-
W CL -ops
0.5-
Q.+ 0
4 -0.5
"0 0. 5-
- 0
~ -o*5
w
0. 5-
E
- 0. 5-
f r equency hyst er esi s ver sus t emper at ur e of t he
S-TCXOs empl oyi ng new shape GT cut quar t z
cr yst al r esonat or s i s somewhat dependent on t he
t emper at ur e excur si on, but t he excel l ence of t he
S-TCXOs i s due t o t he f act t hat t he f r equency
devi at i on i s l ess t han O.7ppm i n a wi de
t emper at ur e r ange of - 30 to t85"C i n f r equency
t emper at ur e behavi or and 0. 38ppm at maxi mum at
t he t emper at ur e excur si on of -45 to t90 and -45C
f or t he t her mal hyst er esi s char act er i st i cs. I n
addi t i on, t he r eason why causes t her mal
hyst er esi s SI i ght l y dependent on t he t emper at ur e
excur si on i s not cl ar i f i ed, but i t is concei vabl e
t hat t he pedest al wher eon a quar t z cr yst al
r esonat or i s mount ed, i nf l uences t he r esonat or .
c--
n Sample No.1
E 0. 5- ......... -.--------
-31
Q. 0 22-.-.-.-"
_ _ %?.--.:::::z==. 7
- ---&..-------
..... tzzz=s?:-G-.- - 0 -..--..
__.__.------- a .
Sample No.1
k=-
- - 0. 5-
No. 2
7 0.5. c_
No. 2
. -..--...
-.-:: .._._
Q -0.5 - - --+-
No. 3
No. 3 0. 5-
[Link]-=- - 0 ----2-==--: '.*.S.=,..---- -
- 0. 5 * -
VI---..-C-
- 1
No. 4
-
No. 4 0. 5-
--..-:- .------
0 - - - e=::--=::=- E o - 0. 5 - LL --
>
F i gur e 5 r epr esent s t he t her mal hyst er esi s
char act er i st i cs of f i ve S-TCXOs measur ed at an
aver age speed of I . 5"C/ mi nute( 3 "C/Z mi nut es) f or
each t emper at ur e excur si on. Namel y, t hi s is t he
case t hat t he aver age speed i s f ast er t han t hat
of t he exampl es i n Fig. 4. T he appear ance wher e
t her mal hyst er esi s occur s f or t he excur si ons of
(a) to (d) shows t he same t endency as t hat of
Fi g. 4. Fur t her , f or each excur si on of (a), (b),
Cc> and (d), each sampl e has t he t her mal
hyst er esi s of 0. 13 to 0. 18ppm, 0. 12 to 0. Zl ppm,
0. 25 to 0. 38ppmand 0. 02 to 0. OSppmat maxi mum
r espect i vel y. Thes e r esul t s ar e by no means
i nf er i or to t hose i n Fi g. 4. Thi s i mpl i es t hat
t her mal hyst er esi s i s not dependent upon a
0
No. 3 U 0. 5-
....................
............................
..... s o - 0. 5.
E 0. 5- >
- 0 ............... a.
-0.5. P
No. 4 0. 5-
0
5 - 0. 5.
0.5-
a 0
E -0-5
LL -1.o-.
....
"" :.......
...._ 2 - \
- -- - * .=.,*,,>*
....
439
No.3
-5-
No. 4
__,- - .....
-u.-==--
No. 5
/---
- -I..--_______.
, . , . , . , . , . , , ,
Lead Wi re
V
Supporti ng Por t ions
Fi g . 6 A novel GT resonat or wi t h no cerami c pedest al .
Y . .
(d) -t 9OT + - 45T + +9OT
==
2 0.2
's 0.1
E O
g-0.1
z-0.2
cr
-60 -40 -20 0 20 40 60 80 100
TEMPERATURE (T
Fi g . 7 Comparison of t he l argeness of t hermal
hy s t er es i s f o r pedest al use and no pedest al .
t emper at ur e speed when t he speed i s so t aken t hat
t he t emper at ur e i n an oven get s st abl e.
A cer ami c pedest al wher eon a new shape
GT c u t r esonat or i s mount ed, may i nf l uence t he
t her mal hyst er esi s, t her ef or e, i n or der t o
exami ne i t s i nf l uence, a novel GT r esonat or
wi t h no cer ami c pedest al is desi gned and i t s
hyst er esi s char act er i s t i cs ar e exami ned.
F i gur e 6 r epr esent a novel GT r esonat or
of 2. l hI Hz and R1=93 Q wi t h no cer ami c pedest al .
The r esonat or compr i ses a vi br at i onal por t i on
and t wo suppor t i ng por t i ons, t he suppor t i ng
por t i ons ar e al so connect ed t o t he sur r oundi ng
f r ame, and t he r esonat or i s mount ed on t wo l ead
wi r es.
440
- - 0.44
4- I -
0.07
0.31
No.11
.__.._ -----.--
% 0.50
Efl 0.25
' 1 : o
Ir, -0.25
No.13
- - - - + 0.07
I I I
-0.25
0 2 4 6 8 10 12 14 16
MONTH
Fi g. 8 Agi ng c h ar ac t er i s t i c s .
F i gur e 7 r epr esent s t he t her mal hyst er esi s
f or pedest al use and no pedest al at t he
t emper at ur e excur si ons of Cc) and (d). A s i s
appar ent f r om Fig. 7, t he r esonat or wi t h no
pedest al shows t he hyst er esi s smal l er t han t hat
wi t h t he pedest al . T h e f r equency devi at i on i s
310ppb at maxi mum f or pedest al use, whi l e i t i s
95ppb at maxi mum f or no pedest al . I n case of
t he excur si on (d), t he f r equency devi at i on by t he
hyst er esi s get s smal l , namel y, i t is I OOppb f or
pedest al use and l Oppb f or no pedest al at
maxi mnm. Thus, t he t her mal hyst er esi s f or t he
pr esent S-TCXO was f ound to be ver y dependent on
a means f or mount i ng a GT r esonat or , but t he
r eason i s not known why t he hyst er esi s t akes
pl ace and al so i s dependent on t he t emper at ur e
excri rsi on.
4 . 4 Aging Characteri sti cs
Fi gur e 8 r epr esent s agi ng char act er i st i cs
of S- TCXOs usi ng new shape GT cut quar t z cr yst al
r esonat or s. Agi ng quant i t y i s shown af t er t he
pr e- agi ng of about t wo weeks and si x mont hs.
T h e agi ng quant i t y f or one year and t wo mont hs
i s f r om0. 25 to 0. 44ppm( mean: O. 33ppm) af t er t he
pr e- agi ng f or about t wo weeks and al so t he
quant i t y f or about ei ght mont hs i s f r om0. 01 t o
0. 12ppm( mean: O. 065ppm) af t er t hat f or about si x
mont hs. Li ke t hi s, t he S- TCXOs ar e excel l ent i n
f r equency st abi 1 i ty f or a l ong t er m
Par t i cul ar l y, t he f r equency devi at i on f or ei ght
mont hs i s, as pr evi ousl y ment i oned, 0. 065ppm at
t he mean val ue af t er t he pr e- agi ng f or si x
mont hs, t hus, f r equency devi at i on f or t he S- TCXOs
i s har dl y obser ved. Her eaf t er , a f ur t her l ong
t ermagi ng wi I 1 be exami ned.
5. Conclusions
I n t hi s paper , a st udy was per f or med wi t h a
vi ew t o cl ar i f yi ng t he f r equency t emper at ur e
behavi or , i t s t her mal hyst er esi s and agi ng
char act er i st i cs of S- TCXOs usi ng new shape GT
cut quar t z cr yst al r esonat or s.
Fi r st , a ci r cui t di agr amof a S- TCXO was
shown and by usi ng new shape GT cut quar t z
cr yst al r esonat or s, t he f r equency devi at i on was
f ound t o be l ess t han 0. 7ppm over a wi de
t emper at ur e r ange of - 30 to t 85" C wi t hout any
ext er nal t emper at ur e compensat i on. Especi al l y,
t he S- TCXO has a ext r emel y smal l f r equency
devi at i on of t 30/ - 150ppb at best i n t he f or egoi ng
t emper at ur e range.
S econd, t he f r equency t her mal hyst er esi s
char act er i st i cs ver sus t emper at ur e of t he S- TCXOs
wer e exami ned i n a wi de t emper at ur e r ange of -15
to t90"C. As a r esul t , t he t her mal hyst er esi s
i s har dl y obser ved f or t he excur si on of f al l -
t emper at ur e t o r i se- t emper at ur e, but i t i s
obser ved somewhat f or t he excur si on of r i se-
t emper at ur e to f a1 I - t emper at ur e, however , t he
var i at i on quant i t y i s so smal l as t o be
sat i sf act or y as a r esonat or .
F i nal l y, f r equency devi at i on was exami ned
f or t he agi ng t er m of about one year and t wo
mont hs. T h e agi ng quant i t y i s 0. 33ppmat
t he mean vaI ue af t er t he pr e- agi ng f or about t wo
weeks and 0. 065ppm at t he mean val ue af t er t hat
f or about si x mont hs. Thus , t he S- TCXOs wer e
f ound t o be successf ul l y obt ai ned wi t h ext r emel y
smal 1 agi ng.
As t he comi ng subj ect s, an exami nat i on wi l l
be per f or med on t he r eason of t her mal hyst er esi s
occur ance f or t he excur si on of r i se- t emper at ur e
to f a1 1- t emper at ur e and f ur t her l ong t er magi ng.
References
[ll H. Kawashi ma: An Anal ysi s of Osci l l at i on Fr equency
Char act er i st i cs i n a CMOS Osci l l at i ng Ci r cui t Usi ng a
Coupl i ng Quar t z Cr yst al Resonat or " , Pr oc. t he 44t h
Annual Symp. on Fr equency Cont r ol , pp. 585- 592( 1990) .
121 H. Kawashi ma and K. Sunaga: " Temper at ur e Compensat ed
Cr yst al Osci I l at or mEmpl oyi ng New Shape CT Cut Quar t z
Cr yst al Resonat or , Pr oc. t he 45t h Annual Symp. on
Fr equency Cont t ol , pp. 410-417(1991).
131 tl. Kawashi ma: " Var i at i onal anal ysi s of GT cut qual t z
cr yst al r esonat or s wi t h t he suppor - Li ng por t i ons at t he
ends" , Pr oc. t he 40t h Annual Symp. on Fr equency
Cont r ol , pp. 193-200(1986).
[41 H. Kawashi ma: " A vi br at i on anal ysi s of a GT cut
quwt z cr yst al r esonat or t aki ng account of di ssi pat i on
t er m" Tr ans. I EI CE(A1. J apan, J72-A, 4, pp, 633- 639
( Apr i I iw).
151 H. Kawashi ma: " An Anal ysi s of Fi . equency Temper at ur e
Char act et . i st i cs and El ect ri ca I Equi val ent C i I - cu i t
Par amet Pr s of a NPW Shape CT Cut Quar t z Cr yst al
RPsonator For med hy an Et chi ng Met hod" , I EEE
U I t i mor i i cs Symp. , Pr oc. pp. 465- 470 (1989).
[6] tl. Kawashi ma, 11. Sat o and 0. Ochi ai : " New f i - equency
t empemt ur e char act er i st i cs of mi ni at ur i zed GT cut
qi i artz r esonat or s , Pr oc. t he 3. l th Ai i nual Symp. on
Fr equency Con t 1.0 I , pp. 131 ~ 139(1980).
171 11. Kawashi ma and M. Mat supama : " An Anal ysi s of
El ect r i cal Equi val ent Ci r cui t Par amet er s f or a GT Cut
Quar t z Ci yst al Resonat or wi t h t he Soppor t i ng Por t i ons
at Bot h Ends" I Tr ans. I EI CE(A), J apan. 571-~A, 6,
pp. 1217- 1224( J une 1988).
[81 D. L. f l ammond, C. A. Adams and A. Benj ami nson : " Hyst er esi s
Ef f ect s i n Quar t z Resonat or s' ' , hoc. t he 22nd Annual
Symp. on Fr equency Cont r ol . pp. 55 66 (1968).
191 H. J . For st et - : "Thermal Hyst er esi s of AT t SC- Cut Quar t z
Cr yst al Rpsonat or s" , Proc. t he 36t h Annual Symp. on
Fr equency Cont r ol , pp. 140- 147 (1982).
[I OI R. I. F i I lei. : '' hf easur ement and Anal ysi s of Thei mal
Hyst er esi s i n Resonat or s and TCXO's- , Pr oc. t he 42nd
Annual Symp. on Pi - equency Cont r ol , pp. 380-388 (1988).
[l l l J . A. Kust ei s and J . R. vig: "Thelma1 Hyst er esi s i n Quar t z
Resonat or s- A Revi ew , Proc. t he 44t h Annual Symp. on
Fr equency Cont r ol , pp. 165- 175 (1990).
1121R. 1. Fi I let- : " Ther mal Hyst e!; esi s i n Quar t z Ci - yst al
Resonat or s and Osci I l at or s , Pr oc. t he 44t h Annual
Symp. on Fr equency Cont i . ol . pp. 176-184 (1990).
441