Noise spikes on USB 6259

[jayrandom]

I am trying to measure a 10-0.1V signal using a USB-6259 BNC device.  I have one channel from an already amplified pressure gauge that is being measured diferentially.  There are relatively large (~0.1V), very fast spikes in any data I take, either through labwindows or measurement and automation explorer.  These spikes appear in data from my pressure gauge or from an agilent 33220A function generator.  The spike pattern appears to be somewhat random, although the spikes do vary in both amplitude and spacing.  There does not appear to be any appreciable width to the spikes, however.  Is this a known issue with the USB-6259 or is there something I could do to rectify this problem.

[jayrandom]

This is a sample of my noise from the Agilent 33220A with 100000 Khz acquisition rate and a 1V signal level.  These spikes are only 60mV p-p, but I've seen others with 100+ mV.  Any ideas on how to fix this problem?

Thanks!

[Henrik_Volkers]

Things to try:

analog:

Check your grounding and shielding.

Find the noise source and turn it off   That's not so easy, if even possible . 

Lower the signal path impedance. If the sensor output can drive 10kOhm put a 10K at the DAQ input. 

My first try at high spikes is a RF transformer ring core with 5 to 10 wounds of both signal inputs. See picture.

A 5B Module to get a galvanic isolation.

digital:

LP filter or very effective on spikes: median filter

Message Edited by Henrik Volkers on 06-27-2009 02:07 PM

Message Edited by Henrik Volkers on 06-27-2009 02:10 PM

[jayrandom]

Thank you for the response.  I don't think the problem comes from my signal, as I don't get this noise when using an internal NI-6036 with BNC-2110 with the otherwise exact configuration.  I'll try your suggestions, though, to see if that fixes the problem.

[Henrik_Volkers]

Sad but fact that DAQ inputs are also a known source of spikes.

USB device vs internal .... ground current loops (my guess: by capacitive coupling)!!

Good chance, that the ring will help, separate the powersupply and/or add a ring into the power supply cable too

The transformer ring forces (better supports) that the sum of all currents through it equals, therefore minimizing unwanted current loops. Best would be to open the loop 😉

Your DAQ input is 10 GOhm!! Take a look at the spec of your sensor output and reduce the input impedance if possible.

Message Edited by Henrik Volkers on 06-29-2009 02:45 PM

[stilly32]

What is your input configuration (differential, RSE, NRSE) , and what have you tried? How do you have the switches (floating source vs grounded source) configured? This is generally the first cause of noise. Most calibrators are considered a floating source, so a differential meausrement with the switches set for a floating source is your best bet. If the calibrator is not a floating source, use differential with the switch set for a grounded source.

Hope this helps, 

Andrew S