08-03-2011 02:39 PM
Hello!
We are using the AnalogIn on a PCI-6071 and the .NET DAQmx tools, and we're seeing strange behavior in the timestamps of the returned AnalogWaveformSample's. While acquiring data, every 5-10 seconds the timestamps for a block will suddenly jump backwards in time ~60 seconds, and then when returning the next block they jump forward again by a similar amount.
The settings used are:
sample rate=6000
Sample Quantity Mode = Continuous
samples per channel = 600
I've attached a spreadsheet(csv) that shows what I mean. First column is timestamp in microseconds, second column is dt. Look at the second column, which shows the effect I'm talking about. You can see it on rows 6001 and 6601, and then 16801 and 17401.
We have used the PCIe-6323 and USB-6343 and do not see this. Any ideas?
08-04-2011 03:21 PM
Mojo,
Are you seeing the behavior on the USB and PCIe cards using the identical code?
If you re-run the code, do you see it on the same sample numbers every time?
Regards,
Kyle Mozdzyn
Applications Engineering
National Instruments
08-04-2011 04:19 PM
Are you seeing the behavior on the USB and PCIe cards using the identical code?
I can confirm that we do not see this behavior on the PCIe card using identical code. We didn't see it on the USB using similar code.
If you re-run the code, do you see it on the same sample numbers every time?
No, it's a random occurrence. It might not happen for 30s, and then it will happen five times in 10s. Never on the same sample numbers, except that it always happens on the first sample of a new block of samples.
I've attached a minimal example program that reproduces the behavior for me.
Nathan
08-05-2011 11:20 AM
Nathan,
Are these also happening on the same computer? If not, it could be some really bizarre behavior of the system clock on the machine producing the issue. If the same machine has the 2 different behaviours, Ill need to really dig in and see what may be happening, as this is very, very odd behavior.
Regards,
Kyle Mozdzyn
Applications Engineering
National Instruments
08-05-2011 11:45 AM
It's two different computers. I'll pull the 6071 out and put it in another machine and get back to you.
08-08-2011 11:06 AM
Nathan,
Sounds good, let me know how that goes for you!
Regards,
Kyle Mozdzyn
Applications Engineering
National Instruments
08-09-2011 12:09 PM
Ok, moved card to another PC, did not have the timestamp issue. Looks like the system clock of the first machine is wigging out. Thanks for the help!
Nathan
08-10-2011 02:52 PM
Nathan,
Glad to hear we managed to locate the issue. I ran into similar behavior before (but MUCH more difficult to track down in that case) which caused a CVI program to intermittently crash during a very long test routine at midnight. Turns out the clock would throw a spurrious value right at midnight which would overflow the stack and cause very weird crashes.
Regards,
Kyle Mozdzyn
Applications Engineering
National Instruments