Occasional shift of pixel values across the entire frame

Discussion in 'STF Series CCD Cameras' started by Elizabeth, May 29, 2015.

  1. Elizabeth

    Elizabeth Standard User

    Joined:
    Mar 11, 2015
    Messages:
    33
    Location:
    Chantilly, VA
    We are using model STF 8300M to capture the image of luminescent biochemical slide (4.5" x 3.25"). Basically, we are measuring the brightness of spots on a slide. We are taking one dark frame and a light frame each for 60 seconds. The resulting image is the pixel-wise difference between the two frames.

    Here is the problem. Once in a while, we see a shift across the entire resulting image while repeating the reading process of the same slide. We believe that the shift happened on the light frame and it was a drop by 100-120 counts.

    The DriverChecker app returns the following versions: sbigudrv.dll 4. 82 Build 1, sbig64.sys 2.41.0.1338.

    Also, I couldn't find information on SBIG support site about which driver version is the latest. Yes, there's info about the DriverChecker and in its release document there're comment about the sbigudrv changes, but it is not consistent. The version that we are running, for instance, was not mentioned anywhere. The DriverChecker and driver seem to be bundled. Is that so? If I was downloading the latest DriverChecker (version 5 I believe) what version of driver would that come with?

    Thank you,
    Elizabeth
     
  2. Doug

    Doug Staff Member

    Joined:
    Sep 25, 2014
    Messages:
    9,932
    Could you please provide examples of "good" and "bad" images in raw format (SBIG or FITS).
     
  3. Elizabeth

    Elizabeth Standard User

    Joined:
    Mar 11, 2015
    Messages:
    33
    Location:
    Chantilly, VA
    The final file format is a *.tiff. Would that be ok? I am running an extensive test right now, repeating dark and light frames for many cycles.
     
  4. Doug

    Doug Staff Member

    Joined:
    Sep 25, 2014
    Messages:
    9,932
    It would have to be a 16-bit TIFF. We would greatly prefer FITS so we have all the information on the exposure parameters.
     
  5. Elizabeth

    Elizabeth Standard User

    Joined:
    Mar 11, 2015
    Messages:
    33
    Location:
    Chantilly, VA
  6. Adam Robichaud

    Adam Robichaud Lead Developer Staff Member

    Joined:
    Sep 29, 2014
    Messages:
    1,012
    Location:
    Ottawa
    Hi Elizabeth,

    Downloading the most recent version of the SBIG Driver Checker 64 from sbig.com will always yield the most recent SBIGUDrv version — we update the executable and the drivers simultaneously. The current version of SBIGUDrv.dll is 4.88 Build 2, and the current firmware version for the STF-8300 is: stf-8300usb.bf2 - version 2.47, and stf-8300ga.bf2 - version 2.18. We haven't updated the SBIG64.sys since we obtained the driver from them (and your version is up to date).

    There are a number of possible causes for intensity drops, but in order to solve the problem it would be nice to capture the effect in an SBIG log. I would advise updating your SBIG Driver Checker (and drivers) to the most recent version, and see if that addresses the issue. Also check to see if your firmware matches the most recent release in SBIG Driver Checker 64's "Firmware" tab. In the mean time, I'll see if we can reproduce the issue in house.

    Cheers,
    -Adam
     
  7. Elizabeth

    Elizabeth Standard User

    Joined:
    Mar 11, 2015
    Messages:
    33
    Location:
    Chantilly, VA
    Thank you Adam, that's exactly what I did. I updated the driver and firmware to the latest version and I am running my test again as we speak. After 100 images taken no problem, however I find it concerning that my values are at different level than two days ago. The same target, same temperature, same integration time, same region-of-interest that gave me ~1270, it now reads ~1160 only. So far the readings are consistent though.

    I am adding to our library the capability of saving different file formats. Do you want the compressed FIT files or uncompressed?

    Elizabeth
     
  8. Adam Robichaud

    Adam Robichaud Lead Developer Staff Member

    Joined:
    Sep 29, 2014
    Messages:
    1,012
    Location:
    Ottawa
    If the reading is consistently lower, that may point to a hardware issue of some sort, though I'm not sure what it would be. The 110 ADU drop being present in both your dark frame and your light frame, and consistently across the board now, hints towards something more consistent.
    As for the FITS files: uncompressed please.
     
  9. Elizabeth

    Elizabeth Standard User

    Joined:
    Mar 11, 2015
    Messages:
    33
    Location:
    Chantilly, VA
    This is a plot of a 100x100 region of interest's ADU average. The images were taken with 1x1 binning for 60 seconds at -5 celsius and data are absolute values. So no dark subtraction is involved. The two blue lines are results using the older driver 4.81 and the green lines are results using driver 4.88 and fw 2.47.
    We have two concerns:
    1. the occasional drops with the older driver; is this anomaly truly gone with the new driver?
    2. the global shift with the new driver.

    upload_2015-6-5_14-19-43.png

    Can I please have the release notes of the incremental driver and firmware versions?

    Thank you,
    Elizabeth
     
  10. Adam Robichaud

    Adam Robichaud Lead Developer Staff Member

    Joined:
    Sep 29, 2014
    Messages:
    1,012
    Location:
    Ottawa
    It's difficult to pin down the exact cause of the issue, and whether or not our changes definitely fixed it, but there were a number of changes in SBIGUDrv that were targeted directly at stability and reliability at both the API level, and the USB communication level, in the months leading up to, and following, our acquisition of the driver. I've done some testing with version 4.82 against our current version 4.88 build 3, but it doesn't really seem to have the effect you're observing. Could you zip your copy of sbigudrv.dll (found in C:\Windows\system), so I can do some further testing? In the meantime, here's my nightly build of SBIGUDrv (aforementioned 4.88 build 3).

    The SBIG Universal Driver has a changelist included in SBIG Driver Checker 64. Just click the "ReadMe" button at the bottom of the window, and scroll down to the changelist. It is possible a version or two was missed during the transition between SBIG and Diffraction, but for the most part the list is meant to be exhaustive.

    We weren't given any incremental changelists for any of the SBIG cameras' firmware.
     

    Attached Files:

  11. Elizabeth

    Elizabeth Standard User

    Joined:
    Mar 11, 2015
    Messages:
    33
    Location:
    Chantilly, VA
    Here I am sending my version of driver, the one that gave stable values. I can send you later the previous version of the driver too if you want to, I just need to get it off of another system which is running my test at this point.
    I have attached the log file too which captured the last part of the test run. I think the log file grew too some max storage and I lost the beginning part. Again, this is the log of the run that I was happy with.
     

    Attached Files:

  12. Adam Robichaud

    Adam Robichaud Lead Developer Staff Member

    Joined:
    Sep 29, 2014
    Messages:
    1,012
    Location:
    Ottawa
    The old version of the driver (the one that isn't stable) would be much appreciated. There's no rush.

    I just thought of something though: I suspect it may be an issue with the STF's gain and offset being updated to non-standard values in the old driver. Can you (using the old and new driver) take a bias frame and a dark frame (with the same camera under the same conditions), and send them to me? I'd like to see what their distributions look like.

    Cheers,
    -Adam
     
  13. Elizabeth

    Elizabeth Standard User

    Joined:
    Mar 11, 2015
    Messages:
    33
    Location:
    Chantilly, VA
    If I want to go back and use the old driver is it enough to overwrite the new with the old "windows\system\sbigudrv.dll"?

    Thanks,
    Elizabeth
     
  14. Adam Robichaud

    Adam Robichaud Lead Developer Staff Member

    Joined:
    Sep 29, 2014
    Messages:
    1,012
    Location:
    Ottawa
    That's it! -A
     
  15. Elizabeth

    Elizabeth Standard User

    Joined:
    Mar 11, 2015
    Messages:
    33
    Location:
    Chantilly, VA
    Please access my OneDrive folder. http://1drv.ms/1Fv9BjE

    The files I prepared for you are DarkDrv481 and BiasDrv481, images using driver version 4.81 and the corresponding ones done with driver ver 4.88. You will also find the copy of the driver that we used to have on our systems. We did a sweeping driver update to ver 4.88.

    I assume the bias frame is a dark frames with 0 integration time.

    Thanks,
    Elizabeth
     
  16. Adam Robichaud

    Adam Robichaud Lead Developer Staff Member

    Joined:
    Sep 29, 2014
    Messages:
    1,012
    Location:
    Ottawa
    I know what the difference is from! We recently disabled "Auto-Bias Level Correction" in the SBIG Universal Driver, because it was causing instabilities. That is likely the cause of the drop.
     
  17. Elizabeth

    Elizabeth Standard User

    Joined:
    Mar 11, 2015
    Messages:
    33
    Location:
    Chantilly, VA
    In our application we are calling the GrabSetup with autobias enabled. How will this play out with the driver that you are saying this feature is disabled?
    Is the application control disregarded by the new version of the driver?

    Thank you.
    Elizabeth
     

Share This Page