NI 9219 Noise Floor Specs

[prof29]

I am using an NI 9219 module in cDAQ 9171 USB chassis to measure very low-voltage signals(under 50 uV) from a strain-gauge setup. I have a half-bridge semiconductor/foil strain gauge with high-resolution sampling timing mode and I am seeing a consistent noise level of ±60 µV fluctuations. I expected the noise level of the semiconductor gauge to be lower however since this is consistent across the different gauges, I am suspecting if this is the intrinsic noise level of the NIDAQ itself? I have reviewed the datasheet and calibration but am unsure how to interpret the noise level if the measurement is being taken at room temperature(under 25 C) and if there is no load(0 V) on the strain gauge? Any help in figuring out the noise level or how to reduce it would be much appreciated! foil_gauge_no_strainsemiconductor_no_strain

[Henrik_Volkers]

The noise level of the 9219 is specified in ppm FS  of Vrms, (3.8 ppm @ 500mV in your case) 

 (shortcut at the 9219 input) and at that level adding an actual cable and the noise of the brige resistors...

so some general hints:

- Know your enemy, have a look at the power spectrum of your signal. That give indications to hum or other periodic noise sources

- The better signal processing has longer integration time 😄

 so if possible use a lower sample rate (with sigmadelta converter that is longer integration time)

or use 'lock-in' approach , repeat the measurements and to a weigthed mean to the exitation (if possible)

 lowpass filter help ..

 I would give a second/third order SG filter a try , 5 to 20 sidepoints (an improved SG filter also include a weigthning window in the filter cernel, but haven't seen that in LV yet)

If you post the CSVs of your two pictures, maybe I or other migth find the time to give more hints