July
2009 ASNT Digital Imaging VII , Mashantucket, CT
Optimized Digital Radiography Using
Co‐60, Ir‐192 and Se‐75 Sources to Achieve
Weld‐Quality Images
Steve Jelfs – Sentinel™/QSA Global, Inc.
Introduction
Digital radiography is often used in conjunction with X‐ray sources, and is frequently carried out in a
dedicated facility or testing lab. This is partly due to the size and power requirements of the X‐ray tube,
and also because of the size of the digital scanners themselves.
In field applications such as pipeline inspection, plant inspection, and fabrication shops, gamma
radiation sources have been used in conjunction with Computed Radiography (CR) for informational
radiography. However, there is increasingly a requirement for weld‐quality imaging using computed
radiography and gamma isotopes.
Thus a robust and portable computed radiography solution is required ‐ one that has been characterized
and proven with different isotopes so that it is capable of achieving weld‐quality imaging in a variety of
inspection applications.
This paper describes such a system – one that has been optimized for achieving the weld‐quality images
that are required to satisfy commonly‐used inspection criteria such as ASME Section V and API‐1104. It
concentrates on the methodology and results that lead to the scanner optimizations, and on the
radiographic considerations essential for achieving weld‐quality images using gamma isotopes.
Credentials
As the world’s largest radiographic source and equipment manufacturer, providing isotopes, cameras,
accessories and services worldwide, Sentinel™ is in a unique position. We have internal access to the
isotopes, labs, and radiography experts necessary to characterize and validate techniques and
equipment.
This allowed us to test the Computed Radiography (CR) system with a variety of isotopes and
applications in order to optimize the scanner settings for gamma radiography, and validate that it is
indeed possible to achieve weld‐quality images that easily meet the ASME requirements.
System Description
The Sentinel™Vision system is comprised of three elements: a high‐
resolution portable scanner (Figure A), highly effective imaging
software, and a computer/display package in a variety of
configurations. The elements of this system have been carefully
selected to provide extremely high performance, while offering the
robust simplicity necessary for demanding on‐site inspections.
Figure A
Sentinel™Vision scanner
July 2009 ASNT Digital Imaging VII , Mashantucket, CT
Variables
There are many variables associated with image quality in CR applications, and these can broadly be
divided into three groups:
Those which are directly under the control of the
radiographer, such as exposure time, type of film or
Phosphor Storage Plate (PSP), source‐to‐film Distance
(SFD), combinations of lead screens, etc. Following
established best practices in setting up the inspection
geometry allows the best image quality. Figure B
shows a typical inspection geometry for a contact
shot on a 6” pipe.
Some are characteristics of the equipment being
used, and thus can be controlled by the
manufacturer. Examples of these include hardware
capabilities such as scanner pixel resolution, laser
power and spot size, and software tools such as filters
Figure B
which can be applied to the captured image.
6” pipe contact geometry
And there are physical characteristics and the laws of
physics, which always have to be taken into account. These include the type of source, activity of the
source, focal size of the source, the inverse square law and other immutable characteristics of the
universe. This category also includes noise‐generating Compton Scatter, which must be controlled in
order to reduce noise in the final PSP images.
Characterization
Since this investigation sought to deliver the best quality possible, the scanner was calibrated with each
of the plate types to be used prior to taking and evaluating any images. This was necessary because
imaging plates of different types vary in thickness and flexibility, and consequently react slightly
differently to the feed mechanism of the scanner. The calibration removes any image distortion caused
by the mechanical variations. However, this step only takes about 2 minutes, and only needs to be
undertaken once for each plate type, unless the scanner hardware changes.
Digital imaging plates from major manufacturers were used for this investigation. High performance and
normal duty plates were included, and also one plate which fell somewhere between the two in
performance. Initially the exposure time was calculated to give a dose to the PSP equivalent to D4 or
MX125 film, but it quickly became evident that the imaging plates required considerably less dose than
conventional film. While the different manufacturers’ plates varied, the optimal dose was generally
found to be in the 800mR to 1.2R range.
July 2009 ASNT Digital Imaging VII , Mashantucket, CT
Applications tested included contact
shots on 3” Schedule 40 and 6” XS
pipe, stand‐off shots on 2” Schedule 80
boiler tube, and various stainless steel
plates up to 4” in thickness. In all cases
the Source to Film Distance (SFD) was
calculated to meet the requirements
for geometric unsharpness per ASME
Section V. Figure C shows an example
of the unsharpness calculation for 3”
pipe.
Figure C Sources used included Ir‐192 Plus™,
3” pipe unsharpness calculation Co‐60 and 75Sentinel™ with different
activities and diagonal focal sizes,
75 ntinel™
although most of the shots were taken using Se and Ir‐192 Plus™ since these isotopes are most
commonly used for pipe weld inspections.
Scanner variables were adjusted to gauge their effect on image quality. The pixel resolution (µm/pixel),
photomultiplier voltage (amplification factor), pixel binning, and other factors were adjusted. Exposures
from a variety of applications were then scanned and evaluated at each setting.
Images were evaluated by three experienced radiographers using a 5 Megapixel monochrome monitor.
Inspection criteria were selected by reference to the ASME Section V standard, using both plaque‐type
and wire‐type penetrameters. Where the standard called for the smallest wire of a wire IQI, the next
smallest IQI was also placed on the sample to measure the best quality that could be achieved.
Details of each shot were recorded in a spreadsheet, with analysis by consensus and notes added to
clarify the image quality achieved. Scanner settings were modified in response to these results, and we
were quickly able to determine which made a significant difference, and which had minimal effect.
Once we were confident that we had optimized the scanner settings, we were able to concentrate on
investigating which physical source characteristics and radiographic techniques achieved the best image
quality in combination with particular isotopes and scanner settings.
Results
In general we found it relatively simple to achieve weld‐quality images with isotopes using the CR
system. In many cases we were able to exceed the code requirements for wire‐type IQI’s and plaque‐
type penetrameters.
July 2009 ASNT Digital Imaging VII , Mashantucket, CT
The pixel resolution of the scanner has a significant effect on the ultimate quality of the image, and also
affects the scan time of the PSP through the scanner. In fact it is a trade‐off ‐ a lower resolution such as
100 µm/pixel gives a faster scan time but a slightly lower‐quality image. In areas such as profile
radiography this may well yield adequate images which meet all criteria of the governing codes.
Figure D
2” boiler tube CR image
However, for the highest quality images, we determined that a 50 µm/pixel resolution provides a good
balance between image quality and scan time. Although the scanner is capable of achieving higher
resolutions, the increase in image quality is not typically necessary, since the most stringent image
quality criteria can easily be met at the resolution suggested above. In addition making the scanner
resolution smaller to the point where it exceeds the capability of the imaging plates provides no benefit,
since the plate then becomes the limiting factor in image quality. Figure D is a CR image of 2” boiler
tubes scanned at 50 µm/pixel. It clearly shows the required #6 wire and 2T sensitivity, as well as visible
weld characteristics such as tungsten inclusions and cluster porosities.
Likewise, the gain of the scanner PMT (photomultiplier tube) has a significant effect on both image
quality and shot time. Raising the gain of the PMT allow shorter shot times, since the dose on the plate
does not need to be as high. However the trade‐off is an increase in image noise, which can lead to
unacceptable results.
Non‐scanner variables which affect the image quality include both the plate type used, and the
characteristics of the source.
July 2009 ASNT Digital Imaging VII , Mashantucket, CT
We found that that the lower‐performance plates from any manufacturer are not typically adequate for
weld‐quality imaging. However, both the high performance plates, and the mid‐range plate that we
tested are certainly able to provide results that meet or exceed the relevant codes. When scanned with
optimized scanner settings, we found that a dose of 800mR to 1.2R on the imaging plates delivered
acceptable image quality. This compares favorably, for example, with the dose required to achieve a 2.5
density on D4 or MX125 class film. Using CR technology allows the shot time to be reduced by about
50%, improving efficiency and reducing personnel radiation exposure. If shot times are not critical, then
a lower intensity source can be used, which can potentially reduce both radiation exposure and the size
of any controlled area.
Because of the above, the use of CR technology in combination with SCAR (Small Controlled Area
Radiography) techniques and products is a viable approach to minimizing both radiation exposure and
controlled area, while providing acceptable image quality in a short amount of time.
An additional benefit of smaller activity sources is their reduced focal size, which contributes to lowering
geometrical unsharpness and improving the definition of the image. We found that the diagonal focal
size of the source was critical in achieving the image definition necessary to meet the code
requirements, particularly for contact shots in the 3” and 6” pipe applications that we tested. It is
essential that the source focal size meets the requirement for geometrical unsharpness required by the
relevant codes – if it does not, then no amount of tinkering with the shot time or scanner settings will
result in acceptable images. Using an advanced isotope such as Ir‐192 Plus™, which offers a smaller
diagonal for a given activity, clearly helps in this regard. Figure E shows a 3” 216 wall pipe contact image
shot with a Se‐75 source with a .107” focal size. This image clearly shows 3 wires (#6, #5 and #4) on the
ASTM A‐set wire IQI, and even some of the #3 wire, well in excess of the required #5 ASME V‐2
specification.
Figure E
3” pipe contact with 75Sentinel™
July 2009 ASNT Digital Imaging VII , Mashantucket, CT
We also found that 75Sentinel™ images tended to be smoother and less grainy than those captured using Ir‐
192 Plus™. This is most probably due to the softer gamma‐ray spectrum of 75Sentinel™, which has a mean
photon energy of 215 KeV vs. the Ir‐192 mean photon energy of 357 KeV. However, using 20 mil front
lead screens with Ir‐192 Plus™ yielded satisfactory quality images by ASME code standards, comparable
to 75Sentinel™ with 10 mil screens under the same conditions. Figure F illustrates this point.
Figure F
3” pipe contact with Ir‐192
Although the working range of 75Sentinel™ is typically limited to about 1 1/8” of steel when used with film,
we found it can be used beyond that material thickness range when used with CR. This is due to the fact
that the digital imaging plates require less dose than conventional film, and also the fact that the CR
software has tools and filters that effectively increase the latitude of the
image and allow hidden details to be brought out. Again though, there is
a trade‐off ‐ increased shot times and image noise have to be weighed
against the convenience of using a single isotope across a broader range
of applications. Figure G is a CR image of 3” stainless‐clad carbon steel
shot with 75Sentinel™, and clearly shows the #12 wire ‐ one wire greater
sensitivity than the code requirement.
Cobalt isotopes are also commonly used with thicker materials. Figure H
shows a 4” thick stainless steel plate shot with Co‐60, and again easily
shows the required sensitivity, with the #12 wire and the 2T hole on the
50 plaque penetrameter both being readily visible.
Figure G
3” SS plated carbon steel plate
July 2009 ASNT Digital Imaging VII , Mashantucket, CT
There are two schools of thought regarding image quality. One
approach is to strive for Just Enough Image Quality – images that
meet the required standards, but without exceeding them. Exceeding
the required image quality can lead to unnecessary additional
expense (more expensive imaging plates or longer scanning times, for
example), and even to interpretation problems where flaws which
are too minor to be problematic are clearly imaged and thus become
all too obvious. Thus the concept of “Just Enough image quality” can
be used to describe a balance between image quality, and the cost
and effort involved in getting to that point.
On the other hand, there is another school of thought known as
”Gold Plating the Job”. This rationale involves achieving better‐than‐
required sensitivity, thus ensuring that technicians will turn in shots
that meet code, and possibly reduce reshooting. This technique can
frequently compensate for the many variables in radiographic set‐ups
Figure H that may produce unsuitable images.
4” SS plate shot with Co‐60
Conclusion
Image quality is influenced by a variety of factors. Control of scanner parameters is one major factor
essential to achieving weld quality, but cannot compensate for sub‐optimal sources or faulty
radiographic technique.
The experiments performed by Sentinel™ allowed us to optimize these scanner settings for different
applications, and they are now accessed through a simple drop‐down menu in the Sentinel™Vision
software. So ease‐of‐use is maximized for the user, since it is now merely a matter of selecting the
isotope, desired quality, etc. from the menu.
To support the scanner’s capabilities it is recommended that the following guidelines be adhered to, in
order to achieve the best results:
• Use higher‐performance digital imaging plates, rather than lower‐performance ones.
• Expose the imaging plates to a dose in the 800mR to 1.2R range.
• Use the smallest diagonal source that is available, such as Ir‐192 Plus™
• Pay close attention to calculation of geometrical unsharpness.
• Use 75Sentinel™, rather than Ir‐192 Plus™ where a smoother image is desired.
• When using Ir‐192 Plus™, consider adding 20 mil lead screens on the source side of the imaging
plate.
Computed Radiography is capable of producing viable weld‐quality images using gamma sources,
provided both the scanner settings and the radiographic factors are in harmony with each other. The
work we have done, and the solution we provide, allows the user to repeatedly meet industry standards
with maximum efficiency.
July 2009 ASNT Digital Imaging VII , Mashantucket, CT
Acknowledgements
The author acknowledges the invaluable contributions of David Carter, Dave Crocker, and Bob Kelly. This
paper is all the better for their wisdom, suggestions, and review. Any errors that remain are solely my
responsibility.
I also am grateful for the help and expertise provided by Allpro Imaging and Dürr NDT.
Trademarks used in this paper are the property of their respective owners.
