I like to think I am pretty familiar with identifying rolling element bearing defect frequencies. I have experience with identifying faulted bearings due to emergence of BPFO, BPFI, and BSF. I have never, however, witnessed the emergence of Fundamental Train Freq (cage frequency) in a bearing. I believe I have seen FTF as a sideband of Ball Spin Freq, but never just FTF by itself...until now (I think?).

We have a belt-drive squirrel cage fan supported by double-row tapered roller bearings. I don't currently know the bearing defect freqs (I asked for that help in a previous posting "Help w/ Dodge Brg Defect Freqs?"). Reviewing data, however, I see the emergence of a peak at 0.45 X shaft speed. It has one smaller 2X harmonic. At first I thought it was belt freq...but after calculation, it doesn't appear to be. It is accompanied by a gradual increase (over 3-4 months) of HFD. I'm defaulting to the notion that it may be an FTF peak.

I'm looking for experience that explains the type of bearing degradation that would be evidenced by emergence of FTF by itself...not as a sideband of another defect freq...but simply by itself in a velocity spectra?

Thanks in advance,
George

George
At that frequency it could be fudemental train frequency. I have picked this up on a number of our bearings and have been shown to be correct. Could you post your spectra?

Though I have found FTF normally as sidebands of BSF and also seen at 1 x rpm sidebands,I have seen only FTF harmonics few times.In CSI peakvue I have seen FTF harmonics very early

Pl find the spectrum with FTF on belt driven fan(fan rpm is 2550 and motor rpm is 1480).this is not belt frequency.

I must add that this is a new bearing fitted 3 days back(SKF 2311 K).This fan has a reputation of sudden cage(steel type) failure with little or no indication .Even sometimes there wont be much change in velocity/ acceleration and SKF envoloped accleration.I am lucky if I find a bearing problem in this fan.I have tried to replace this self alighnng ball bearing(very little adaptability to axial load) with spherical roller bearing ,but for high rpm.

Spectrum with FTF

FTF.doc (34 Kb, 38 downloads)

George,

Assuming your Dodge has an SKF insert in it, a 22215, the cage frequency would be 0.435 (rest 3.726, 8.28, and 10.74). The Dodge 22215ESK has 0.4309. Both would be close.
I have seen cage defect by itself in Spike Energy, and watched it get big and then dissapear. When the bearing was changed, the cage was broken between rollers. Customer wasn't excited, I was.

Dave

George, I have only seen this 1 time in 10 years of vibration work. It was on a 1500 HP vertical pump motor. This pumped around 30,000 gallons per minute of river water to our plant. I noticed the peak at cage frequency in June and knew that we would really need to have the pump in July and August. It was very costly to pull the pump and send out for repair. I really studied over that decision. I did not want to make that call, but I went with my gut feeling. I went to the repair shop when the motor was torn down. The cage on the top bearing was cracked all the way through on 1 side.

George,
I have experienced these kind of defects many times in my career. Cage defects showing up in the velocity spectrum get a higher level of serverity in my book. These defects are not usually associated with a high level of impacting but when one of these little cage bars brakes and allows the bearing roller to turn side ways it is not a good thing. Most times FTF can be easily detected in the velocity or peakvue waveform. I find FTF is often found when a defective roller is present because the rollers travel through the load zone at the same speed as the cage. (usually around .4 orders) When the bearing is removed and inspected physical evidence may or may not obvious. Cage damage is sometimes very hard to see, look for worn cage bars, sharp edges etc. If the rollers easily fall out of cage this is a good indication of excessive cage wear. Cage wear is a good indication of ineffective lubrication.

These plots are from a reject refiner motor that failed about two hours after this data was taken. The spectrum shows many multiple orders of the cage fault frequency at 0.38x RPM. The PeakVue spectrum also shows the cage frequency.

cage_fault_frequencies.doc (126 Kb, 38 downloads) Cage failure