Hi picojohn; thanks for listening. Most failures were due in the past to the computer locking up waiting for the trigger. We learned to arm the trigger only shortly before the event (less than 2 or 3 minutes); this seemed to help. Getting a newer computer also seemed to help. (Boy did we have problems with a laptop running Windows 98!) Last year we only had one failure, but unfortunately in the confusion we didn't find out what caused it. I don't think Picoscope had locked up, because I have a completely blank trace I must have triggered in some manner after the failure to trigger. (Actually as I write this I wonder if it triggered and somehow collected a blank trace.)
We use a single trigger on voltage rise from a designated one of three transducers and save each trace manually after it's captured from -0.1 second to +0.9 second. We do this several times a day, no faster than a half hour between events, during maybe 10 or 15 days a year. The signal goes past the trip point in several milliseconds and stays high for almost the whole 0.9 seconds, so it's not that the event was only a blip too fast to be seen by the relatively slow ADC-11.
I'm baffled; if this failure happens again I will immediately perform the manual air blast injection to the transducer (which I now do as a dry run before every pressure event; this tests the entire signal train) to see if it triggers on that. If it doesn't, it would seem that the signal didn't reach the ADC-11 or PS has locked up. Windows Task Manager (after crtl-alt-delete) could tell me if PS had locked up, because when this has occurred in the past, 2 Picoscope entries were seen in the TM, and CPU usage was usually 100%.
No obvious cause seems apparent here; I was fishing for anecdotes or past experience from the forum that might lead me to software conflicts, bugs, tips, etc. Luckily we're more reliable today than we've ever been, but not quite where we want to be yet.