AI voltage channel read in vi does not match actual voltage

[test-lab-joe]

Hello,

I am running LavVIEW 8.6, using PCI-6225 and SCB-68. I have a vi that runs a simple system and takes some data. I am progressively modifying the vi to increase its functionality. I seem to have introduced an error that I can't find. Unfortunately, I am unable to post the vi or screen shots at this time.

I use DAQ Assistant to gather 4 AI voltages. I recently added the fourth and shuffled it to the top of the order since its the only low voltage signal. Now the last channel read drifts and asymptotes to a consistent value 2.35 V. The measured voltage at the source and at the SCB-68 is ~.21 V. When I monitor the channel using Measurement and Automation Explorer it also reads ~.21 V. It seems the vi is doing something to the signal. Any ideas?

Thank you.

[crossrulz]

Did you try to test it inside of the DAQ Assistant configuration window?

[test-lab-joe]

Yes. It also drifted to the same asymptote there.

[test-lab-joe]

Initial functioning vi successfully monitored 3 channels (ai69-71). The all read correctly, regardless of order in DAQ Assistant.

I created a new vi that just reads the 4 channels (ai68-ai71) with DAQ Assistant. I played with the order (24 permutations possible), running the Express Task to evaluate the results. In some orders (8 of 24), channel ai71 asymptotes to 2.35 V. In some orders (4 of 24), both channel ai70 and ai71 asymptote to 2.35 V. In other orders (12 of 24) all channels behave correctly.

Symptoms occur at both tested sampling frequencies (1kHz and 10kHz).

I thought the device cycled through channels sequentially, so I tried to order sequentially also and group signals with similar magnitudes.

Could this be a settling time issue between channels? Could it be because I'm wired (or mis-wired) to use Differential Terminal Configuration (e.g. pins 1 and 35 for ai71)?

[_Bryan]

What is your sample rate?  Slow it down and see if it goes back to the correct voltage.

Also group similar voltages together and read the channels in order.  If you sample fast enough and there is a big enough difference between channels you can run into issues with the time each channel needs to settle to get a good reading.

This might apply.  But without seeing any code it is just a guess.

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

[Bob_Schor]

My advice would be Never Use DAQ Assistant (except for quick "Look how easy this is" demos to Students) -- you have no idea what is going on "under the hood", and cannot easily "fiddle" to make your own adjustments.

What happens if you simply open MAX and run some Test Panel "tests" on your board?  Can you get it to function properly?  (I'm betting the answer is "Yes", and if the answer is "No", I'd consider sending the board to NI to test).

A wonderful White Paper to wean you from the DAQ Assistant is this one.

Bob Schor

[johnsold]

Could be ghosting.  What is the source impedance connected to the "misbehaving" cahnnel?

Lynn

[test-lab-joe]

Thank you for pointing out the white paper. Will work toward using the basic building blocks for greater control. MAX showed appropriate behavior for one channel at a time.

[test-lab-joe]

Impedances were mismatched (2 orders of magnitude) but the "order" for successful reading seems random. I might be misunderstanding what "order" actually does in DAQ Assistant, especially relative to the physical channel assignment.