AI big difference in ranges / multi-channel ancquisition

[Georgesvk]

Hello,

I am stuggling to get my system to work well. I am using analog inputs at NI PXI-6363 card, I measure 16 AI differentially.

AI0 is set to -0.1 V to 0.1 V range, I need about 0.1mV precision on this channel (it is 100mV output pressure transducer)

others are set to 0 to 10 V range, I measure voltages in this range.

When I configure task on all channels I end up with interferences between channels. I read 2.3V on AI6 and just after that I measure 5mV on AI0 whereas real signal on AI0 is -0.5mV. This happens on acquisition rate 100Hz, and AI convert rate 90kHz (default value).

To eliminate this intererence I tried to decrease AI convert rate (using timing property), but I had to go all way to the 2kHz. Then it was ok, whereas still not that stable as if I would measure only AI0. I can never get that stable data, even if I decreased AI convert rate to 60Hz.

I need data with defined specific time (100Hz sampling and be able to increase it).

I know it is an interference from previous channel because next channel voltage reacts on change of previous channel voltage.

Issue occures when I acquire these channels one after another. 

Do you please have any idea how to fix this problem?

Is there a way to put time gap between AI conversions? (timestamp difference of  channels had always the same value 11us no matter of AI convert rate property setting).

Thank you for your ideas and answers.

[will.i.am10 · Accepted solution]

Hello,

It is likely that you are experiencing “ghosting” between signals. Follow the link below for common solutions to this problem. The easiest solution may be to arrange your signals so that you minimize the voltage swings between channels.

http://digital.ni.com/public.nsf/allkb/73CB0FB296814E2286256FFD00028DDF?OpenDocument

When you changed the convert clock rate did the timestamp difference change between channels? When you slow this clock rate down it should increase the time between samples.

Thanks,

Will

[Georgesvk]

Hi,

Thank you, indeed, ghosting was a problem. I shorted one channel between switching ranges and it worked well then.

That 11us difference is because how DAQmx treats the multichannel acqusition. It chooses maximal rate and adds 10us padding because of settling time.

http://digital.ni.com/public.nsf/allkb/42484E84DA98053686256D32006E0494

I don't quite understand why I can't increase it, but it works now so I don't have to.

Thank you.