I have a client in Brazil that is convinced that he is getting electrical noise in his acceleration signal. He has run his cable from his accelerometers into a cable tray with what he believes is some cables carrying electric current. Not ideal.

He has been unable to send me a wave form as they do not collect wave forms at this site. (I know please don’t chastise me I have had the discussion with them about the importance of wave forms in vibration analysis) In any case I have looked at his spectrums and I see no peaks at line frequency (60Hz). I would expect to see at the very least a large peak at Line freq if he was truly experiencing cross talk with the electric cables.

Does anyone have a white paper or technical paper on electrical noise in the acceleration signal. I have tried searching on Google and have not turned anything up.

Thank s ahead of time fo the help.

Gerdau_Morgan_Block.doc (112 Kb, 17 downloads)

My first thought would be to evaluate what's in the signal that isn't coming from the machine being monitored. Have him take data with the machine shut down and see what peaks remain. Anything still there is background vibration or electrical interference.

If the machine can't be shut down for some reason, a comparison between the signal from the accelerometers and data taken with a portable data collector could be made.

Running signal cables in a tray with power cables is a bad idea, but I've really only seen a problem a few times when the signal cables were run with high-current supply cables to the machine. You're right that you'd expect to see line frequency signal if there was crosstalk. If the signal cables are run with cables from a VFD, all bets are off.

Jon
Spintelligent Labs

Wow, really good point about the VFD, I had not taken that into consideration.

The hard part about trouble shooting is that the cables are integral with the accelerometers installed inside of the gear box. The process is a continuous process 24 7 and they shut down only for failures and forced maintenance.

I also agree that running data cables and power cables together is not the ideal situation.

Thanks John

If the machine is on a VFD, changing the speed and verifying that all the peaks track will be helpful in seeing anything that's not related to the machine.

Even measurements taken on the gearbox case should provide some verification that the peaks seen in the spectra are real. The levels may change, but anything that's real should show up on the casing.

Jon
Spintelligent Labs