Dear community,
I am currently trying to acquire acceleration signals from PCB Piezotronics (model 352C23) plugged to a national instruments PXIe-4492 card. The signal I get looks OK but there is a DC offset witch is allmost 10% of the maximum value of my signal (Fig figure).
One possibility would be to subtract the DC offset but this offset possesses a small random frequency as shown in the figure and I cannot filter it since its frequency correspond more or less to the one I’m trying to acquire. When I’m doing integration of this signal, even passing through the frequency domain, results are completely wrong. The other things to mentioned are:
1) I’m using the AC coupling built-in option of the PXIe-4492 (see Fig) so I should not get DC offset (eve if I cut the excitation sources and wait for 10s, the DC remains).
2) The setup is built on a stabilized table so it is isolated from “outside” vibrations.
3) I put aluminum foil around all the BNC connections but no change.
4) Inverting physically 2 accelerometers at the connection of the card level (for example leading to the green and red line of the Fig) change nothing in the DC level of each channel (the green is going to remain on the top at the same level and the red on the bottom). This means that the DC is not coming from the accelerometers but from the card, in agreement with comment 2 and 3.
5) Change the current provided by the card from the default 4 mA to the maxium value allowed by the sensor (20 mA). No difference in the curves.
I plugged direclty the accelerometers to the card but is there anything to do with the ground?
If anyone get idea on how to solve this problem, directly in LabView or physically, they are welcome.
Best,
Florian
