We are currently transitioning from a development to a production environment. The software and hardware has been fully validated and verified to work on both our development platform and our first production unit. One of our tests uses a NI-5112 PXI digitizer to capture a position switch pulse. This pulse is 28VDC and approximately 200ms in duration. Now for the problem.
During system verification after S/W load of our second production unit we encountered false triggering during the previously mentioned test. Basically the test arms the scope and waits in a while loop for the pulse to occur via the LV scope Status.vi. Once proper status is verified a fetch of the data is performed using the LV scope Fetch.vi followed by an abort of the scope via Abort.vi to keep the scope from retriggering. Due to the while loop the fetch is continuously repeated but returns the same data over and over again as has been previously verified during testing.
On this particular unit however each fetch returns not only the nominal waveform but various "noise" spikes. These spikes show up in various locations throughout the waveform and have had magnitudes of up to -19VDC. Because of the while loop these "spikes" appear to our program as a pulse and are evaluated instead of the actual pulse thereby causing an error in our test. Throughout it all though the original pulse is present and stationary within the waveform.
After complete reverification of our S/W configuration we started investigating for H/W failures. We started by replacing the 5112 with a spare unit but the problem remained (although the pulses were of a lessor magnitude). After some other intermediate steps we replaced all of the cards in our PXI chassis with a set from our development unit (which does not exhibit the phenomenon). We then retried the test and had nominal results. We then replaced each of the original cards one-by-one back into the unit testing for the failure each time. The failure returned when we replaced the NI-5112 digitizer. It seems that both the original digitizer and the spare digitizer (purchased at the same time) exhibit the problem but the older digitizer in the development unit does not.
We have been able to limit the problem by eliminating the multiple fetches by storing the first fetch to memory and reusing it instead of constantly refetching however occasionally the "noise" spike will be present in the first fetch causing a test failure to occur.
Is there possibly a difference in firmware levels between these digitizers that would explain the differences we are seeing? Did we just get a bad batch of cards? Anyone else ever experience something like this? I've looked through the site but can't find a similar problem noted.
We are in the process of moving the card to a different location in the PXI chassis to see if there is any difference noted.
