Random AO channels swapping on PCI-6115

[iveilleux]

Hi,

I am using a PCI-6115 to do simultaneous AO-DO and AI "finite samples" acquisitions using the onboard clock. Using NI-DAQmx, I observe random signal swaps between AO-0 and AO-1 when I stop/start the task (not happening if I clear/recreate the task every iteration)
I am using NI-DAQmx (8.7.1f3). My PCI-6115 board is revision F (register problem) but I installed the EEPROM update utility.

Any idea what is wrong here?

Thanks,
Israel

[Brooks_C]

Hello Israel,

Can you explain what you mean by "signal swaps?"  It might also be helpful if you can post the VI you're using.  I can't think of a circumstance where the output intended for ai1 would suddenly switch to ao0 and vice versa in the hardware.  My suspicion is that you're unintentionally switching the values you want to write to these channels somewhere in software.  I would use the probe tool in LabVIEW to verify the values you're writing to your outputs.  You can access this tool by going to View>> Tools Pallet...

The middle tool in the 3rd row (the p with an arrow through it) is the probe tool.  With this tool you can add probes to the block diagram to see the values on wires during VI execution.

I hope this helps, and if you're still having trouble after you check the values in software please feel free to post back with more information.

Cheers,

[Tom_W_[DE]]

Hi Israel-

I assume you're talking about the EEPROM utility contained in this KB.  If so, can you please execute this command from the command prompt:

"C:\Program Files\National Instruments\NI-DAQ\TradDAQ6115\TradDAQ6115.exe check"

and post the result?  This will help us determine whether the EEPROM utility is updating your device correctly.

Thanks-

[iveilleux]

I have attached a simplified VI of my application.

Connecting AO-0 and AO-1 to AI-0 and AI-1 respectively, I observe random channel swaps in the Channel 0 and Channel 1 graphs for some values of the 'size' control (eg. 204, 212, 220,...). The problem arises from the AO (not the AI) as observed on an oscilloscope.

Can somebody reproduce this behavior?

Thanks again,
Israel


P.S. Running 'TradDAQ6115.exe check', I get the IOPCR = 0x153 as expected

[Brooks_C]

Hello Israel,

I have obtained a PCI-6115 and I have been able to reproduce the issue so I've have notified R&D.  Since it doesn't look like this can be fixed immediately I'd like to focus on helping you find a work around.

In my testing I've found that the signal switching occurs if you generate arrays of size = 8*x + 4 (where x is an integer), so if you can avoid this in your program you shouldn't experience this issue.  Also, if you need to have a size equal to a multiple of eight plus 4, then you can also add one to the "samples per channel" input on the DAQmx Timing VI so that it is set to (Rate^2) + 1 and not just Rate^2.  This second option should only have the effect of outputting a single extra point--I'd imagine it would be unnoticeable for most applications.

If neither of these work arounds is suitable for your application please reply back with more details about your application and current issue with changing one of these sizes.

Best Regards,