Ceramic Hybrid Bearing Failure Analysis

[ccaissie]

We have had a failure of a ceramic hybrid angular contact bearing, and are seeking the root cause of this.

We are using a lower pair of 6220 deep groove radial bearings which are intact, and an upper pair of 7218 bearings with a preload gap of 11 microns, which was properly torqued and locked to apply about 2000N of preload. The weight of the shaft in this vertical application is about 400kg. Speed 68 rpm.

The lower 7218 bearing of the upper pair is intact. The uppermost bearing has suffered a wavelike pattern of wear, easily seen and detected by touch. there has been speculation that it is of electrical origin, but this is not likely.

We have tested this mixer in house using steel bearings, and then the ceramic bearings and have done periodic vibrational analysis of the geartrain and motor and mixer base. All is normal after weeks of operation. We simulated the vertical load, but cannot test with the actual shaft in place.

The mixer runs with the tank product, a polymer, pressing up and through the bearing assembly and out a flush port...essentially reverse flushing.

I know it seems weird to run these bearings with water or product flushing through, but we have many mixers with these ceramic bearings, and with product flushing through them.

It is suggested that we had lost the preload due to excessive inital wear when the customer ran a week long test with unfiltered flush water that was contaminated by rust and scale.

There is also a suggestion that the long heavy shaft has created resonancies that have overwhelmed the preload and caused this wear pattern related to unusual movement of the balls.

I attach a report with photographs for your review.

Any suggestions?

 

[Tmoose]

Was the angular contact pair supposed to be assembled X (face-to-race) or O (back-to-back)?
Were they universally ground, or match ground?

 

[ccaissie]

The bearings are a ground matched set, Back to Back with the preload accomplished by compression of the 11 micron gap in the assembly. We have pretty much eliminated concerns whether this preload arrangement is correct.

Something evidently over-rode the preload force. We're thinking it might be some unpredicted forces generated in the shaft.

But, I'm opening the query up to some wild ideas at this time.

CSC

 

[unclesyd]

I have been looking for a report where we had a similar appearance on failures in steel bearings on large vertical centered agitators.
Recalling from memory.

We finally attributed the damage to the bearings was initiated by an oscillating axial forces in the shaft generated by the mixer blades. This was a bottom supported mixer shaft and the problem started when modifications were made to the supports and lowering of the side wall baffles. The actual vertical osculation was small but it had a tremendous force as the mixer was driven by 75 HP motor.

We had other failures of bearing on large mixers due to failure of the bottom bearing which caused the shaft and impeller to go in orbit. If the orbit was circular we would get indication of noise prior to failure of a bearing. If the orbit was eccentric or had any osculation it was a short happy life for the bearing.

 

[thgirw]

What about the surfaces in contact with the shaft? If there is a problem with pre-load usually there are some fretting evidences on shaft/bearing contact area.

Cheers

 

[dinjin]

It is suggested that we had lost the preload due to excessive inital wear when the customer ran a week long test with unfiltered flush water that was contaminated by rust and scale.

I like that analysis.

 

[ccaissie]

Dinjin:

We saw no signs of unusual contact or pressure or wear on the shaft. The 440C (or european equivalent)bearing steel seems to behave well against the 316SS shaft.

We're running tests now and all seems well, though a new sound, a squeak, has appeared, so we're getting a vib analysis reading on that now.

Thanks,
CSC

 

[ccaissie]

thgirw:

I addressed the post to dinjin by mistake. See above.

 

[ccaissie]

dinjin:

My first impression also was that the customer ruined our mixer by flushing it with dirt. But we need to be more analytical than that.

CSC

 

[caoimhin1]

Very interesting damage in the photo. You see some strange things on bearings, but not seen that before. Looks like static vibration damage, but as if the bearing had been indexed around while its happened.
Cerobear's analysis is good, this wouldn't have happened if the bearing had been preloaded. Adding springs will help.
Does the process fluid need to pass thru the bearing? Could you use a sealed precision angular contact?

 

[ccaissie]

Thanks, Caoimhin1.

The mixer was shipped so that there was no vibration and rotation which could have "indexed" a pattern. The number of indents and ball count are showing a realtionship.

Flush-through is our standard design, but this case makes us want to think again about that.

CSC

 

[unclesyd]

Your last post keep rolling over in the windmills and as I recall we had failure of baring on a very large suspended agitator in a 100,000 gal stirred reactor. The agitator had a 12" shaft and was driven by a 2000 HP motor. I can't recall the specific bearing and am thinking it was a Timken roller bearing on the output side of the gearbox.
The reactor is in a chain of 5 in series and by taking one or two out of the loop it gives production a very high turndown ratio. The trouble is that in the process flow still has to go through the unreacting reactor and in this case the 30' long shaft with two set of blades just sits there. It was during a long spell of business slump that this reactor was taken off line several times. While off line the flow from a 20" line was impinging on the blades causing the whole thing to wobble back and forth with no rotation of the bearing. Th process people turn off the vibration alarms during these outages as they keep going. It was over a peroid of several months that there was very little run time on this bearing and when they came up to speed they couldn't reset the alarm, so it was jumpered to be able to run a few days until the gearbox could be changed out. The outlet side bearing had failed and the recalling a pattern was embossed in the race way and was almost the same as seen in your failure.

I was out of the loop on most of the boring stuff at this time but it was revealed that they had a very similar failure a year earlier, but the bearing was destroyed and the failure was attributed to water in the oil getting to the bearing while the agitator was not turning.