Ground reference noise during welding

[jonhelpmejon]

I am measuring proximity probes (electrically isolated signals) using a SCXI system with a 1202 module. There is welding going on around the areas where the probes are installed. Seeing the probes signals are electrically isolated, I need to reference them to ground. However, the electrical noise created from the welding is unacceptable. Is there any way to get around this problem or recommendations to reduce the noise. Thanks

[Sean_C.]

Hello Jon,

Thank you for contacting National Instruments.

One thing you can try is to move the 1102 (I assume it is an 1102, as we do not make a 1202) out of the first slot. The first slot in the SCXI chassis is the noisiest because of its proximity to the power supply.

Placing ferrite rings around your sensor leads also helps to reduce noise. Additionally, you can oversample your sensors and take a running average of the last 10 points or run your signal through a software filter.

Of course, the best way to reduce this noise would be to eliminate the source, the welding. But if this is not possible, then the above methods will help clean up your signal.

The following Tutorials offer more details about noise reduction:
http://zone.ni.com/devzone/conceptd.nsf/webmain/99b4323fab04ffd686256811005025f2?OpenDocument
http://zone.ni.com/devzone/conceptd.nsf/webmain/01f147e156a1be15862568650057df15?OpenDocument

Let me know if you have any further questions.

Sincerely,

Sean C.
Applications Engineer
National Instruments

[jonhelpmejon]

Thanks for the quick response Sean. Yes, the module is an 1102 my typo error. Unfortunately we cannot stop the welding because what we are measuring is a function of the welding. The biggest problem I think is that we are referencing ground which is the same ground that the welding machine is using. Is there a way to not have to reference ground. Would it help to put the daq system on battery power and not reference ground at all. The readings that we get from a portable fluke are very accurate without noise during the welding.

[Buechsenschuetz]

I think the only solution is to avoid a common ground of the acquisition system and the welding devices. Once you have noise on the ground line it cannot be filtered.

Another solution would be to use galvanic isolation between the probes and the acquisition system, and to use a separate power supply for the probes. It is possible to isolate analogue signals but it is not easy, especially if you have very low output signals (in the mV range) from the probes.