slow image acquisition after upgrade to 1.6
Added by Anonymous over 15 years ago
We have a problem after upgrading from 1.5 to 1.6: after the upgrade, acquisition of 1k, binning 4, images in the MSI-Edge (1.6) has slowed significantly w.r.t. to the old 1.3 application, to the point where manual focus correction, for instance, is barely possible (and painful!), since images only come in every 15-20 seconds or so. This behaviour is not consistent, the delay between 1k images ranges between 3-5 seconds and 30 or more within a single session, with no apparent pattern.
Upgrade on the (Linux) Leginon machine was Leginon, pyScope, pyami and sinedon (all downloaded from SVN), as described in the manual in 'Upgrade from 1.5 to 1.6'; on the Windows microscope machine, we upgraded Leginon, pyScope, pyami and sinedon as well as adding the comarray package. Camera is a Tietz FC415 (which shows up as a Tietz SCX, strangely, in the pyScope/test.py for instance). Has something changed in the camera communication to cause this erratic behaviour, or did we miss something in the installation?
Any suggestions? Thanks
Thorsten
Replies (5)
Re: slow image acquisition after upgrade to 1.6 - Added by Jim Pulokas over 15 years ago
Just to note, the new comarray package should only be necessary for the Gatan cameras (or anything that uses tecnaiccd.dll). I think the problem is with choosing the right Tietz pyScope class to use. There is now a new class specifically for the FC415. Try it out in python:
from pyScope import tietz t = tietz.TietzFC415() image = t.getImage()
If that works without errors, try forcing that as your class in pyScope/instruments.cfg
Otherwise, you may want to try each of the Tietz classes in instruments.cfg until one of them works with Leginon. Or try the above test directly in python with each class. Choices are:
tietz.TietzPXL
tietz.TietzPVCam
tietz.TietzSCX
tietz.TietzFC415
I can now see that the test.py script in pyScope is outdated and does not include this new TietzFC415 class. This may be they one you want.
Re: slow image acquisition after upgrade to 1.6 - Added by Anonymous over 15 years ago
Hi, sorry for my late reply. We tried all other Tietz classes during installation in order to get the camera to work (note that this is a single-port readout FC415). The TietzSCX class is the only one that works at all (i.e. where the camera shows up in the Presets node). Using TietzSCX under v. 1.5, the camera works fine; it's only after the upgrade to 1.6 that we encountered these slow readouts.
For the time being, we have downgraded to 1.5, which seems to work. I have not tried removing comarray again. Do you expect that to interfere with normal camera operation?
Thanks!
Justus
Re: slow image acquisition after upgrade to 1.6 - Added by Jim Pulokas over 15 years ago
I am working on a complete test script for Tietz cameras so help resolve these issues. I will let you know when that is available. One other idea we have is that there could be multiple versions of camc4.exe installed on the system. Anchi recalls that after FEI service did a Tecnai software update, there is now a camc3.exe included in the tecnai files. This new one may be in conflict with the original camc4.exe installed with the Tietz software. The camc4.exe also have a config file called camc.ini, which may cause problems if it is reconfigured. Have you recently had an update of either your Tecnai software or your Tietz software? In any case, you may want to search your drive for all versions of camc4.exe and camc.ini to see if there are multiple versions.
Jim
Re: slow image acquisition after upgrade to 1.6 - Added by Anonymous over 15 years ago
Unfortunately, I have no idea what software updates (if any) were applied to the machine; we will have to check when Thorsten returns from his vacation....
Thanks for your help!
Justus
Re: slow image acquisition after upgrade to 1.6 - Added by Jim Pulokas over 15 years ago
Finally finished the new test script, and a few other fixes to make the tietz.py module more reliable. If you can use the latest pyScope from the svn trunk, you will get this new test script called tietztest.py. Run the script and it should run a test on each of the possible tietz classes and tell you which are successful. I have tested it on our two Tietz cameras, and now interested to see if it runs on any of the other cameras out there.