...

Possible breakthrough:  if I power cycle the camera and run without reconfiguring then L1Accepts show up in pgpreader at the correct rate.  Feels like our configuration is breaking the camera?  Compared the power-on defaults (https://www.slac.stanford.edu/exp/npa/manuals/opal1000man.pdf) to our config.  Don't see an obvious difference in UART settings except we have VR=1 (vertical remap) and OVL=1 (overlay frame counter and integration time on early pixel data)?.  Other differences (BL(black level), DPE (defect pixel correction) OFS (output offset)) feel like they don't matter.  Tried setting VR=0 and OVL=0 but it didn't help.  Set them back to 1.

Horrible workaround:  if I configure the camera, then powercycle, it appears to run.  

The XPM link went bad again, and recovered with TxPhyReset.

Questions for Larry/Ben:

• does clearreadout trigger blowoff and fiforeset? (and reset the latched xpmoverflow)
• why would TxPhyReset help the xpm link?

Not Understood Failure Modes

...