A multi-device ADC application based on a PXI-1045 chassis has started to fail in a strange way. We have a PXI-6608 in slot 2 and 7 PXI-4498's in slots 3-9. The 4498's sync'ed and I ususally designate the card in slot 3 as the master, the rest as slaves.
I route the sample clock from the master over RTSI to a counter pair on the 6608. I use it to clock a pulse train task with a period of 1000 ticks, and the output of this task gates a buffered event task to timestamp every 1000th sample.
Recently, this counter task has been failing after 10-20 minutes: no buffered events appear. The VI has been working for years and hasn't changed so I suspect a hardware problem. I noticed that if I designate the card in slot 4 to be the master, and the other slaves, everything works fine. And it does seem to be the slot number that matters, and not the actual card, because swapping only the cards does not fix the problem.
Could it be that I have a bad slot in my PXI-1045? Perhaps the RTSI connections are flaky? Can you think of other ways that I could test this idea?
It may be best for you to call National Instruments for some troubleshooting on the phone and perhaps sending in the chassis for repair.
Please go on www.ni.com/ask to create a service request number and then call 866-275-6964.