Dave suggested if HWSY induced glitches in HWSX again maybe we can do some cable swap test to see if the effect is reversed. However both cameras ran fine simultaneously after a computer restart so the test didn't happen.

I did a little follow up to see if there's any pattern to these glitches. Both HWSX and HWSY codes used to be able to run together for >10 days prior to Ubuntu14. These glitches happened on random days (not every maintenance Tuesday for instance), and not necessary in order (sometimes HWSX stopped running first, sometimes they both stopped at the same time). This time I only left HWSX running to see if the code would run for much longer time. If there's still an issue this might suggest that the problem is not with the PCI-e card.

The plot attached show 10min trend of HWSX and HWSY spherical power over hundred some days including the time prior to Ubuntu 14 (installed Nov 4, 2016). Flat line indicates that the HWS code stopped writing data (due to camera glitch). Sorry for the messed up time axes. Dataviewer doesn't always plot nicely when I ask for too much data. The hand written time on data after Ubuntu 14 was acquired with another second-trend plot (not shown).