Analog voltage signal noise NI-9204 vs NI-9239

[rockyd]

More of a curiosity question here. I'm measuring a +/- 10 VDC signal from a load cell amplifier that is paired with a +/- 10k lbf load cell. I originally had the analog input connected to a NI-9204 module in differential input mode. I was getting quite a bit of measurement noise, so I tried connecting it to a NI-9239 module I have kicking around. The NI-9239 provides a much cleaner signal. My question is what specific feature does the NI-9239 have that provides the superior signal conditioning? Could I achieve the same signal conditioning with a basic RC filter?

My load cell is approx 12 feet away from my control cabinet and runs near some AC lines. I've used shielded cabling and kept the cables as far away from noise sources as I can. The 3 attached graphs are all sampled at 1 kHz and the screenshots are taken from within NI-MAX. I tried the 9204 in RSE and Differential mode with little to no improvement noticed. I then swapped the wires in my cabinet over to the NI-9239 module and I get a very different looking signal. All other conditions were identical between the 3 graphs. 

[santo_13]

Both are designed for different applications.

NI 9239 is a DSA card designed to acquire high resolution (24-bit sigma delta) audio or vibration signals which are typically sine waves with dedicated ADC per channel with additional filters.

NI 9204 is general purpose AI to capture normal resolution (16-bit SAR) signals and only one ADC multiplexed across all the channels.

If you notice a lot of noise but accurate signal, go with averaging for a smoother signal.

[rockyd]

When you say "if you notice a lot of noise but accurate signal, go with averaging for a smoother signal", would you acquire the data and run it through one of the low-pass filter VI's? Any examples you would recommend? 

Ultimately the waveform data I am interested is in the less than 10Hz range. 

[mcduff]

The 9239 is 24bit and implements a digital anti-aliasing (AA) filter, the 9204 has no AA filter. I am guessing you are seeing high frequency noise that aliases down on the 9204. As another test, put a square wave in both DAQs, one will show Gibbs phenomena (correct), the other will not.