X-Ray Forum for NDT Users
It is currently 09.12.2019, 03:44

All times are UTC + 1 hour

Post new topic Reply to topic  [ 5 posts ] 
Author Message
PostPosted: 02.12.2019, 15:41 
User avatar

Joined: 19.10.2019, 15:47
Posts: 80
ASTM E1165 is intended to measure the size of focal spots of industrial X-ray tubes with an estimated focal spot size larger than 50µm. It was first published in 1987 from ASTM Committee E-7 in Subcommittee E07.01. There was a major update in 2012 when
- the pin hole method was aligned to EN 12543-2 with a similar pinhole method but still film-based
- a new method for digital evaluation was introduced (ILP-Integrated Line Profiles)
- a table for focal spot classes introduced (Table 3)
- annex A1 with an alternate focal spot measurement method for end users is included that end users can measure the unsharpness due to the focal spot with a simple hole penetrameter
- a test report in accordance with ASTM E691 was done to give a Precision and Bias statement.
For taking the image a pinhole camera is required. There are strong requirements to the pinhole ...

... and it's sizes:

Up to a focal spot size of 300µm the smallest pinhole diameter of 10µm is required. From >300µm to 800µm a 30µm pinhole diameter shall be taken and for larger focal spots a 100µm pinhole is sufficient.
The next requirement is the minimum magnification; it is given in table 2 based on the Focus Detector Distance of 1m to

Up to a 2mm focal spot the magnification is 3 and the distance between focal spot and pinhole shall be 25cm. The 1m FDD shall be used if possible. Sometimes the cabinett where is tube is installed does not allow 1m distance between focal spot and detector; a footnote opens the door to smaller distances:
When using a technique that entails the use of enlargement factors and a 1 m focal spot to detector distance (FDD = m+n) is not possible, the distance between the focal spot and the pinhole (m) shall be adjusted to suit the actual focal spot to detector distance (FDD) used
(for example, if a 600 mm FDD is used, m shall be 150 mm for 3:1 enlargement, 300 mm for 1:1 enlargement, and the like).

For the digital image there are some more requirements:
(a) coverage of focal spot in digital image: more than 20 pixel
(b) SNR in digital image: 50
(c) pixel value of focal spot maximum: >30% and <90% of max pixel value
(d) grey value resolution: 12 Bit and more (>4095 shades of grey)
With all this requirements we get some interesting results from the physics

The unsharpness due to the pinhole is
Ug = P(1+n/m) definitions of P, m and n in tables above
An example for a 250µm focal spot size using the 10µm pinhole: The unsharpness is 40µm and based on spatial resolution 20µm (Ug/2) which is 8% only. This is absolutly in the tolerance of the focal spot fabrication. But with a 50µm focal spot this influence is already 40% :thumbdown:
The example with 300+µm with the pinhole of 30µm shows an error of already 20% due to the unsharpness of the pinhole - this is just in a class of the tube fabrication.
Therefore - if you accept a error of 20% you should not use this methods for focal spot sizes below 100µm and from 0.8mm up to 1mm focal spot you should use the 30µm pinhole.

The second issue is the amount of pixel inside the focal spot - more than 20 pixel. With the 3:1 magnification you would need a detector with a pixel size of 7.5µm :O which is to my knowledge not availabe for a normal amount of :money: . A solution would be to increase the magnification. E.g., if you have a high resolution detector with a 20µm pixel size you would end up with a magnification of 8 and FDD would be (8+1)*25cm = 2.25m. As 20µm is the best you can get, the table above is based on 20µm pixel size of the detector. Following the footnote in Table 2 of E1165 and use instead of m=25cm only 15cm could shorten the length to 1.35m. As described above for a 50µm focal spot the method is due to the error of the pinhole size not suitable, with 100µm focal spot size the magnification has to be four and the lenght of the mechanics would be 75cm (15cm + 4*15cm) which is absolutly okay.
(Hint: EN12543-2 is intended for focal spots larger than 200µm ;) )

Reply with quote  
PostPosted: 02.12.2019, 17:07 
User avatar

Joined: 19.10.2019, 15:47
Posts: 80
Set-up before image capture using a digital detector
First you have to assure the conditions (b) to (d) above and additionally
(e) Calibration of pixel geometry with a precision of 2µm or 1% of pixelsize.
As it is difficult to put an object with known size of this precision into the beam (should be inside the focal spot camera :O ) , you should spend some effort on exact geometry of you system; most important is the distance from the focal spot to the surface of the pinhole. Most tubes have a colored dot at the position of the focal spot; verify the geometry with high precision and if you see a deviation from the 25cm (or 15cm in the compact version) take for calibration of the effective pixel size of the system the measured value.
E.g.: With a 20µm pixel size of the detector, m=15cm and n=45cm (mag=3:1) the effective pixel size of the system would be 20µm/3=6.67µm. If you measure instead of 15cm 14.5cm only, the magnificaion would be 3.103 and effective pixel size of the system would be reduced to 20µm/3.103=6.44µm what would lead to a difference of 3.3% in the result.

It is also very important that you position the focal spot camera in a way that the central axis of the beam is aligned directly to the pinhole. Note 2 of E1165 recommends to use a special collimator to ensure conformance even with +/-1° alignment tolerance. You shall have the maximum intensity in the center of your detector (if the detector is centered in the camera); small deviations should be corrected, otherwise an elliptical shape of the focal spot may occur.

The voltage of the tube shall be 75% of maximum but not more than 200kV. If the intensity on the detector is too high you may use prefilter with a homogeneous material like copper or brass. To avoid to much signal from the lower energies I recommend to use a 0.5mm brass filter in front of the detector for voltages up to 150kV and 1mm for higher voltages. Additionally the detectors are more sensitive to lower energies and most of the high energy quants will pass the detector without producing any signal and the filter will correct that a little bit.

If you have more than 40 pixel on the focal spot you are allowed to remove outliners by a 3x3 Median filter ...

... but due my experience there is not a big difference as the method of integrated line profiles is quite robust.

Reply with quote  
PostPosted: 02.12.2019, 19:28 
User avatar

Joined: 19.10.2019, 15:47
Posts: 80
Image capture and evaluation of the images
Some focal spot cameras are equipped with a fixture for photostimulable luminescence storage phosphor imaging plates (commonly called IP’s) for the dental systems which mainly have the same size as the older films. The only way to controll the required PV is the exposure time - a few exposures for testing are required.
Using DDAs you see the image on the screen just after expousre. The reuqired PVs can also be set by the exposure time, e.g. a smaller pinhole requires a longer exposure time. To test it, an image with 1s exposure time can be done; the 90% of maximium PV of the detector is divided by the maximum PV of the 1s exposure and the intergration time is multiplied with this value. If the SNR does not reach the requires SNR=50, several image may be taken with same conditions and accumulated in a computer (hint: four accumulated images double the SNR value).
Be sure the offset value is substracted all the time and sometimes it is required to do a gain calibration. For details refer to and the topic about .

For evaluation a special software tool within the line profile function of the imaging software is required.

Two line profiles with about three times the width of the focal spot are drawn and the integrated signal in both directions is used for evaluation. E1165 offers to download the iSee software in a demo version for free from the BAM server but the link is no more valid; now the software can be downloaded with full functionality directly from - the only restriction is that the images can not be stored.
Before giving a "hands-on" with the ISee! software, the procedure shall be explained.
First calibrate the pixel size. Draw a line profile with about three times the width of the focal spot across the horizontal direction pinhole focal spot image:

Integrated (accumulate) the line profile signal along the profile:

with the result of an "S-curve". Set marker in the steep area – at 16% and 84% following the Klasens method:

The distance from 16% to 84% is 388µm; it has to be extrapolated to 100% by the factor of 1.47 Hint: 100%/(100-16-16)%=1.47:

The width of the focal spot is 571µm and the size in the pinhole image:

matches the calculated value quite well.

Now the same procedure has to be done for the vertical direction:

and here the calculated size is 546µm:

which matches also the pinhole image quite well:

The final result of the evaluation of this pinhole image give a size of 571µmx546µm:

That is the theory of the measurement method.

Reply with quote  
PostPosted: 08.12.2019, 15:12 
User avatar

Joined: 19.10.2019, 15:47
Posts: 80
Here is the hands-on version with ISee!:
First thing to do is the calibration of the pixel size. This can be done with Image - Input Pixelsize and SRb... and in the bos the system pixel size shall be put. For the example this is 8µm or 0.008mm. A detector with 24µm pixel size was used at magnification of 3:1. Put in for SRb the same value.
With Mode - Profiler (or just F3) the line profile function will come up. Draw a horizontal line profile with a length of threetimes the size of the focal spot.

The line profile of a single line is shown in the graph.
Now some adaptions to the task have to be done. First increase the line width (1) to a value of about threetime the height of the focal spot - here 99 pixel (hint: use odd numbers for a central line). Then Detrend the line (2) and apply the integration function (3).
Finally switch from manual measurement to 16%-84% edge unsharpness (4). This will use the Klassens Method for evaluation as required for E1165 evaluation.

Obvious this was the horizontal size but the vertical size is required also. Draw a new line profile perpendicular to the one before. You have to adjust the width - here to 199 pixel. If you activate OSD you will see the results in the image.

For your report a screenshot could be done (CNTL Print). Additionally ISee offers to list the ROIs (and a profile is an ROI also).
Open the list of ROIs (third icon from the right) and now the results are shown here.

You should enter a description to document the direction of measurement. (Hint: This evaluation was done with a second image I did with same parameters - a small deviation is always possible but will not change the focal spot class.)
Last thing to do is to match the results with the focal spot classes in Table 3 of E1165:

and create a report as shown in Table 4 of E1165.
Measured width (X): 0.69mm
Measured length (Y): 0.3mm
Reported width (X): 0.8mm
Reported length (Y): 0.32mm
Focal Spot Class: FS 8

If you operate an YXLON Image 3500 system this functionality is available, too. The pictures in Fig. 2 of E1165 were done with an YXLON system and the steps you have to follow are the same as with ISee.
If you would like to get a graph as shown in Fig.2d you should start ImageJ with the focal spot image, klick Plugins - 3D - Interactive 3D Surface Plot and you will get a diagram like this:

In the left 3D diagram of the small focal spot you can see the side wings beside the central spot quite well. The right diagram of the large focal spot shows the two edges of the filament with higher intensity - if you rotate the graph to the best direction.

Reply with quote  
PostPosted: 08.12.2019, 22:20 
User avatar

Joined: 19.10.2019, 15:47
Posts: 80
Difference to the evaluation of EN12543-2 (version from 2011)
From more than 200 focal spots the pinhole images were available and evaluated with both method - the 10% threshold of EN2543-2 and the ILP method of ASTM E1165 (and may be EN12543-2 2020 ;) )
In the result the focal spot sizes are slightly smaller with ILP method than with EN12543-2:

For the Precision and BIAS statement in 9. ASTM requires an evaluation following E691 for repeatability. In E1165 chapter 9.1.1 the parameter of a round robin test are given for a standard tube with a focal spot size of about 550µm - similar to the tube in the beginning of this thread. The integrated line profile method was tested with more than 100 pin hole images (focal spot length) using 5 different CRs, different positions and evaluation at three different Labs and the consistency is given:

Images taken with MXR160HP11 at COMET in Flamatt, Switzerland; Evaluation at
YXLON CRx Software Evaluation with Image 3500 and IP x at YXLON, Hamburg
COMET CRx Software Evaluation with Matlab and IP x at COMET, Flamatt (CH)
BAM CRx Software Evaluation with ISEE visual and IP x at BAM, Berlin

Reply with quote  
Display posts from previous:  Sort by  
Post new topic Reply to topic  [ 5 posts ] 

All times are UTC + 1 hour

Who is online

No registered users

Similar topics

ASTM E1165 Measurement of Focal Spots by User Method
Forum: Standards for Focal Spot Measurement for NDT
Author: Klaus
Replies: 2

You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum

Search for:
Jump to:  

Hosted by iphpbb3.com

Impressum | Datenschutz