First sign of electrical fluting are BPFO or outer raceway with harmonics. These are most noticeable in the demodulated (ENV) spectrum first and later in the FFT between 100K and 140K CPM. I'm beginning to see cage frequency side banding around the fundamental BPFO and its harmonics. I suspicion that this results from cage damage due to electrical discharging originally and elevates in amplitude due to damage progression on the BPFO raceway. Thoughts on this please! Has anyone else had success looking at fluting in this manner?

You're on the money with your diagnosis. I have had similar success. Could you post a spectrum?

I found this electrical fluting on bearing from a 1250hp motor on 4/20/06 I'll try an post a picture of it

Nice billboard...I mean picture. Thanks.

Tommy, please post the spectrum and waveform on this bearing.
Thanks, Scott

Would cage frequency sidebanding around the fundamental BPFO be caused by cage damage?

I would be more suspicious of ball or roller damage, which will be modulated by the cage frequency.

It's still a good indication of fluting.


regards,
Ian

I will attempt to post the spectral data.

First is an Enveloping spectrum Fmax of 150K, second is an FFT Fmax 153K. BPFO markers are dropped. Accompanying these are a jpeg close up of the fluting. Hope I have them attached and they are satisfactory. The cage or rolling elements showed no faults save a little wear. So I assume that once you have discoved fluting you would trend the bearings with the elevating of sidebanding as a mark for change. This being only if measures such as installing of grounding kits to short circuit the fluting through the bearing. Would this be an acceptable method of tracking to failure? Thoughts please.

Regards, JB and thanks for all input!

Spectrums didn't make it I'll try again

FFT Spectrum

Scott,

Here is the spectrum and waveform of the billboard (sorry about that) size picture. I hope that this is not that big also.

Tommy

Tommy
What does Peakvue look like? Can you post it?

Dan

Who needs Peakvue? Bearing fault in the velocity spectrum, 26 g's peak in the time waveform at bearing fault. Your Peakvue will simply demodulate the time waveform and create a spectrum out of it and give you a pk-pk overall of 52g's.

JB, and others,

I would like to know what preventative measures you will be taking to prevent the fluting in the future. We recently had a customer with the same issue, and the 6 - 9 month delivery for ceramic bearings was unacceptable, as well as the time and expense to pull the motor and ceramic coat the shafts. After much research, we finally installed some carbon grahite brushes on the outside of the endbells. Very easy to install, but still have no results back yet as we only installed them last week. Other ideas and experiences would be great.

Dan,

The only Peakvue I have is in the horizontal direction and is not near as high as the vertical direction that was posted earlier, but here it is.

I have heard that there are some special nonconductive greases that can be used, but I do not have any specifics. Anyone out there heard of that?

To Brian specifically and to all the rest,
The key to preventing EDM (electric discharge damage) in the bearings of equipment on a VFD is very simple. High frequency grounding. The cable between the motor and the drive MUST include a 360 degree braided or copper flim shield. This MUST be connected to the drive ground on one end and to a good motor frame ground lug (preferably inside the motor terminal box) on the other end. The high frequency transients (5 MHz) generated in the motor windings need to return to the drive ground bus. If a low impedance ground is not provided, then they will find their own path back, and that is usually through the bearings.
NOTE - 8 gage stranded copper wire was tested and found to have too high an impedance for 5 MHz.
Also, the less paths to ground for this high frequency (a belt driven motor is the worst), the quicker the failure.
The answer to your question is not to continue to look for ways to identify EDM in the vibration, but to eliminate it!
Non-conducting grease? Doesn't conduct, doesn't lubricate either.
Shaft brushes? Bandaids!
Insulated bearings? Conduction through couplings will ruin the driven equipment. Also, putting high frequency current on the plant ground grid (because that is the path it will take without the proper drive cables) is not a good idea.
I have been troubleshooting these problems for Rockwell Automation for over 5 years.

Ron, Thank You for your clear and complete response. Naturally there are questions.

How does one know the ground cable has 360 degree braided or copper film shield?

Where on the drive does the ground go?

If the ground is correct, how important is the cable length from the drive to the motor?

Can the current go up belts to the driven fan?

Does 10 amps through the ground cable seem about right? (just kidding, not)

Thanks again,

Sorry, I was in beautiful Bar Harbor, Maine for a week. Just got your reply.
The cable part number tells you if it has the braided or copper shield. Most people are finally putting in the right cable, but not terminating it properly. Strip the braiding once the cable is inside the motor terminal connection box and then lug it to one of the bolts that goes through the throat and into the motor frame. On the other end, strip it once the cable is inside the drive cabinet and lub it to the ground bus. It must be terminated on both ends.
Length doesn't matter to these currents, however real long runs on drives can create standing waves that can cause turn to turn failures in the motor winding. The fix for this is a terminator (ballast) at the motor.
No, the current can't go up the belts (are you funnin' me Bub?) They will pass through any and all metal connections trying to find that low impedance path back to the drive ground bus.
I have seen 2 to 3 amps. The rule of thumb is that the shielded cable is sufficient to return the high frequency transients.

I don’t understand the purpose of a low-resistance / low impedance path connecting machine frame ground to vfd ground. Can you explain how / why that would help?