First 2 projections are bad when using tomoscan_pso with Grasshopper 3 on 13BM

[MarkRivers]

I am seeing a problem with the first 2 projections being bad under the following conditions:

• Long exposure time (2 seconds)
• Grasshopper3 camera
• tomoscan_pso code
• tomoscan_13bm_pso implements set_trigger_mode

The first 2 projections are garbage. They are overexposed and don't even look like the sample. They ruin the reconstructions.

It appears that when switching from internal trigger mode used for flat fields and dark fields, to external trigger mode used for projections, the first 2 images collected with the external trigger are bad. This is true even though the exposure time is the same as the flat fields.

This is a problem that I had previously fixed in the tomoscan_13bm_mcs class. In that case I always collected 3 dummy images using the external trigger generated by the MCS. In PSO mode generating 3 dummy triggers is difficult, because we would need to have a dummy motion.

In manual testing I have found that the problem seems to be fixed by switching the external trigger source to Software and taking 3 images by pressing the TriggerSoftware button.

I will add the equivalent of this to the set_trigger_mode() function in tomoscan_13bm_pso to see if it fixes the problem.

[decarlof]

did you check if the camera gain is set by chance to auto? if you collect dark at the end this could be the issue

[decarlof]

Ok just ignore it. Is an old issue coming back ...

[MarkRivers]

GainAuto and ExposureAuto are both Off. I had this same issue with my old IDL version, and the MCS version of tomoscan. I am surprised you have not seen it with your Grasshopper3, but maybe you don't use long exposure time and fly scans?

[decarlof]

The exposure times at 2-BM are typically <0.2-0.3s. I am aware of the issue because we had a work around in the old (pre-tomoscan) code at 32-ID

[aps-7bm]

For the white beam tomo I'm doing at 7-BM, the exposure times are usually 10-50 ms, so I've never seen that.