06-19-2017 03:23 AM
Hi,
I currently have a CDAQ (NI 9132) and 3 cards connected collecting data via labview.
However the data seems to have show quite a bit of noise (see attached). The chart is from excel as the application is pretty basic and logs data to a tdms file only
My question - is there something that can be done within the program to eliminate the noise? Why is the noise there? Is it because of wiring issues? The equipment we are connected to is pretty old (25+years) so a few extra counter/relays have been added to obtain the 4-20mA signals. Could it be due to that?
Any help of links to articles is much appreciated.
Solved! Go to Solution.
06-19-2017 05:50 AM - edited 06-19-2017 05:51 AM
Spikes can occure due to many reasons... and analysing the reasons will lead to more insigths on your setup!
I suggest a median filter with a length of 11, (left and rigth rank = 5), but try different lengths
apply it to all chanels and the group delay doesn't harm you
06-19-2017 09:32 AM
Additionally, I would recommend physically conditioning the signal as much as possible before reading it, to avoid transients from destroying your instruments.
06-19-2017 09:46 AM
4-20mA if wired correctly should not be badly effected by a noisy environment and the averaging recommended by others will smooth things out a lot.... but...
The magnitude of those spikes could be a sing that CMR of your input device is being exceeded. In that case, you need to pay attention to the grounds since just one channel going out of the CMR range will/can drive the signal conditioning out of its operating range.
Disconnecting one channel at a time may tell you something as well.
Ben
06-22-2017 02:16 PM
Is that a time domain signal or has it been FFT'd?