Crosstalk of NI6014 DAQ card

[jmgong]

I programed the NI6014 to generate a waveform through ports DAC0out and AOGND. 1.2K ohm load and -4V~+4V amplitude was adopted. At the same time, the differential input ports of channel 0 is directly shorted together, i.e. ACH0 and ACH8 are connected directly. All the connections were based on SH6868 cable assembly, and there is no other connection.
Ideally, when I changed the output vlotage of the DA output, the input signal should not change. However, I did get a waveform at the analog input channel! The waveform is similar as the signal of DA output, and the amplitude is about 1mV. After I removed the load, the crosstalk disppeared. I guess it is due to the crosstalk between the DA output and AD input, but the amplitude is too large!

[JoshuaP]

Hey jmgong,

The reason you were probably seeing this cross talk is because your analog input was set for a differential measurement and you were reading a floating load (not grounded). This is not a recommended setup. With your current setup NI recommends that you attach a 100kOhm resistor from ACH8 to GND. This is to allow bias currents to escape the completely floating system.

The other solution would be to use RSE mode instead of differential and connect the ACH0 directly to AIGND.

Here is a knowledgebase that explains the differences.
http://digital.ni.com/public.nsf/websearch/B9BCDFD960C06B9186256A37007490CD?OpenDocument
http://zone.ni.com/devzone/conceptd.nsf/webmain/01F147E156A1BE15862568650057DF15?opendocument

I hope this helps out.

Joshu
a P.
Application Engineering
National Instruments

[jmgong]

Sorry, I tried this method, but it did not work--the crosstalk kept the same when I used the bias resistor of 10K, 50K, 100K respectively.

[jmgong]

i did more experiments, and found the crosstalk depends on the output current of DA channel. When the output current is of order 1mA, the crosstalk is of order 1mV; and for current of order 0.1mA, the crosstalk 0.1mV. The results are almost the same whether I added the bias resistor or not.

[JoshuaP]

Hey jmgong,

Have you tried RSE mode and then short the analog input channel to AIGND. Then test to see if you still see the cross-talk. I am just trying to see how bad the cross-talk is.

Just to double check did you try connecting the bias resistor between ACH0 to AIGND and from ACH8 to AIGND at the same time. You might even want to put a 100K resistor between the channels to simulate an actual sensor?

-Josh