NI 9220 delivers wrong measurements

[imanSaberi]

We are using a cRIO-9074 with a NI 9220 to measure analogue signales and a NI 9474 to trigger deigital outputs.

On our NI 9220 are two analogue signals connected. One +10 to -10 V sinus signal with 2 minutes cycle duration and one +10 to -10 V sinus signal with 10 Hz frequency. After several hours of measurement the 10 Hz signal is measured incorrectly as +2 and -4 V instead of +10 to -10 V. We already crosschecked it via oszilloscope, the signal remains +10 to -10 V. A reset of the cRIO via reset button won't help, only after unplugging power and reconnecting, the signal will be measured correctly again. The other channel which is measureng the 2 minutes cylcle sinus will always show the right results.

Can someone relate to this issues?

Thank you very much.

[HeinzG.]

Hi imanSaberi,

I suspect your device is defective. I would call in at your local branch and return the device.

Kind regards

Heinz

[AndyClegg]

Obvious question - how fast are you sampling ? A +/-10V sinewave will at some point be 2V and -4V.

How are you reading the AIs - in LV RT in scan interface mode with while loops or timed loops, or are you working on the FPGA ?

Sine only the 10Hz signal is wrong, have you tried swapping the physical input pins the two signals are connected to - do you stil get the problem ?

When you connect the oscilloscope and see the correct +/-10V, do you still have the signal connected into the 9220 and does it still show the wrong magnitude, or does the act of connecting the oscilloscope fix the measurement in the cRIO ?

[imanSaberi]

It appears, that a two high input signal (-20 V to +20 V with 2 min period, from which we for the sake of higher resolution intentionally only measured the -10 V to + 10 V part) of the first channel has charged internal capacities and led to false measurement of the second channel. Changing the input signal of the first channel from -10 V to +10 V (by changing the machine output) solved the problem for now. To avoid crosstalk between the channels of an analog input device with multiplexer like the NI 9220, unused channels between used channels should be grounded to discharge capacities.

We had help from our nearest NI distributer.

[AndyClegg]

That particular source of the problem should only really be present if you are sampling the signals very fast. Your sine waves are at quite slow frequencies, so I can't imagine you need to sample particularly fast - but was why I asked the question. Glad it is resolved.