When sampling multiple analog input voltage channels in a single DAQ task, using the PCI-6221 (in RSE mode), my voltage readings are not accurate. As I change sample rate up and down my voltage readings move up and down as much as 2 volts. Is there a workaround/fix for this? I assume what must be happening is that the PCI-6221 has only one ADC and multiplexes the inputs for multi-channels causing inaccuracies in the voltage readings. Is this correct?
There a few things that could be causing your problems.
1. As you guessed, there is a chance you might see some "ghost" channels because there is only on ADC on the DAQ board. Normally, though, this discrepancy will NOT be 2 volts or anything nearly that large.
The voltage readings are as large as 2 volts, because I forgot to mention that I am using a custom scale with my configuration, which obviously amplifies the problem. The scale has a slope of approximately 5 (calibrated to a specific resistive voltage divider) and Y intercept of 0. I'll check the links you posted to see if I can track down a solution.
In the attachment is a Measurement And Automation Explorer test I used to verify my problem. When applying 10 VDC to an input and setting the sample rate to the max of 62500, I read 7.73 VDC. Changing the sample rate to 100, I read 10.00 VDC. Accuracy appears to roll off dramatically at about 2000 samples/s. Does NI offer a DAQ card that would be capable of sampling 4-8 channels simultaneously at >50 ks/s without this settling problem? It looks like I may have chosen the wrong card for this application, but I'm sure the price of the right card will also be prohibitive.