Hi John,
Ø
EPICS_CA_MAX_ARRAY_BYTES is set to 100MB on both the server and client sides.
That is actually not the best practice. My understanding is that once the CA server needs buffers larger than 16KB it actually allocates buffers of size EPICS_CA_MAX_ARRAY_BYTES. Thus you should
not make it that much larger than it needs to be,
As far as I know there is nothing that ADPilatus can do to cause camserver to write a bad TIFF file. If the TIFF file written by camserver is bad then I would look on the camserver side. ADPilatus is just
sending simple ASCII commands over a socket to camserver.
On my system I see this when I start camserver on my Pilatus 100K system.
PILATUS 100K, S/N 1-0229, BNL
...Code release: tvx-7.3.13-121212
1 PSI PMC GigaSTaR card(s) initialized
PMC GigaSTaR kernel driver version: gs_drv-2.16
PMC GigaSTaR device no. 0, Version 0x1041a424
So I appear to be running tvx-7.3.13-121212, which is almost identical to your version tvx-7.3.13.121212b. Yours just has a b at the end.
My Pilatus 100K is running fine with the latest ADCore and ADPilatus.
Mark
From: John Dobbins <john.dobbins at cornell.edu>
Sent: Friday, October 2, 2020 10:06 AM
To: Mark Rivers <rivers at cars.uchicago.edu>
Cc: Arthur R. Woll <arthurwoll at cornell.edu>; tech-talk at aps.anl.gov
Subject: Re: ADPilatus
EPICS_CA_MAX_ARRAY_BYTES is set to 100MB on both the server and client sides.
The image I provided a link to was a screen shot of an ImageJ display (areaDetector plugin), but the tiff file saved in alignment mode looks the same when opened with ImageJ or
Matlab.
I will try the NDFileJPEG plug-in when there is an opportunity.
Hi John,
That looks like it could be a problem with EPICS_CA_MAX_ARRAY_BYTES. How did you produce that .jpg image? Was it with an EPICS CA client?
Please use the NDFileJPEG file plugin to save an image. Does that have the same problem? If not then I suspect it is EPICS_CA_MAX_ARRAY_BYTES.
The version of TVX/camserver should not matter for the NDArray data. The communication of the data between camserver and the IOC is via TIFF or CBF files, which should be completely independent of the version of TVX/camserver.
Mark
We have a number of DECTRIS PILATUS detectors for which I recently added an up-to date areaDetector installation.
Previously these detectors were using areaDetector R1-9-1, EPICS Base R3.14.11
I added areaDetector R3-9, EPICS Base R3.15.8
These are the three detector and their tvx versions - note that the 100K tvx version is different.
PIL 100K tvx-7.3.13.121212b
PIL 200K tvx tvxe-7.4.01-130212
PIL 300K tvx tvxe-7.4.01-130212
The 200K and 300K are working fine with the new software, however the 100K is not.
The image from the PIL 100K looks like this:
The first 26 rows appear fine, i.e. pixels typically show tens of counts, then in row 27 pixels with very large numbers and pixels with zero start appearing.
Neither the IOC nor camserver is reporting errors.
The PILATUS 100K continues to work as expected with the old areaDetector installation.
Any ideas on how to proceed?
Research Support Specialist
Cornell High Energy Synchrotron Source