First, we could feel pounding at the motor during regular data collection. Since this is the pump that supplies the recycle paper stock to 10 PM and 10 PM is the biggest user of recycle stock it is a critical motor. Our data indicated problems at running speed with other running speed harmonics running closely at the same frequencies. I was worried that the motor was going bad internally and that it had bad bearings too even though there weren’t any bearing frequencies present. I wrote the WO it to run the motor uncoupled and see if we could determine the problem.

After the guard was removed we also found a gear style coupling with a jack shaft. This brought out another possible problem in that the gears in the coupling could be wearing out. The coupling was cleaned of the grease and the teeth inspected and they showed minor wear. A work order was written to change this coupling over to an Atra-Flex.

We also found the motor end bonnet over the fan had broken bolts and was vibrating hard. We repaired this and ran the motor uncoupled. As it started it was very noisy as in a very high pitched whine. After it ran for a while it settled down but still felt rough. We called electrical supervisor so he could hear it and we were trying to decide if we should change it now or run it until next shutdown because it was getting to be too late in the day to start a motor change.

Before shutting down the motor and tagging it out we wanted to put some grease in the bearings. The Oiler arrived to grease the bearings and as he greased the shaft end bearing something banged and the high pitch we heard at startup started again only loud enough that we moved away from the motor! We waited a minute to see what was going to happen and then moved in to collect vibration data to see if we could identify problem. My last overall on this motor was .1 overall inch velocity driving the pump. Now running uncoupled I had peaks approaching .8 and an overall at .9 inches velocity! The peaks appeared to be in the bearing range and after downloading the data I identified an inner race frequency.

There was no doubt that we had to change the motor now. A new motor was ordered and was brought to the job site. In confirming the bearing number in the old motor we found that the shaft end bearing was a straight roller bearing made for a side pull. It should be a roller bearing for a motor driving an in line machine. A motor that has a side pull like a belt drive should use the straight roller bearing. Per Ken Hamilton at Reed Electric, the problem is that the straight roller bearings have extra clearance to accommodate the side pull and when coupled inline to a motor there is not enough side force to turn the rollers. This causes flat spots on the rollers. After examining the bearing in the old motor it was obvious that this is what was going on. The new motor has a warning tag pasted on the side informing that this motor has straight roller bearings and it needs a side pull.

This also brought up another question of the amount of grease and frequency of greasing for the different style of bearings in shaft end bearings. The condition of the old bearings confirmed that the rollers were not turning and the grease was rock hard.

I tend to agree with Mr. Hamilton and doubt that a change in lubrication practices will do much to help.

I know Toshiba and Siemens both have re-greasing interval charts for different ambient environments and hours of operation

As you say, it sounds like misapplication of a "cylindrical roller bearing". When it does not have sufficient preload (side load in the case of CRB), then it can skid.

It also sounds like a piece of that rock hard grease got pushed into the bearing during greasing, and was the immediate cause of the final failure.

Is there a relationship between the grease being hard and the skidding? Does skidding cause heating which baked the grease? Or some other grease problem (mixing greases, high D*N, age)

I have seen many applications where 'roller bearings' were used in direct applications for many years without incident, although it is not good practice.

A couple years ago we had 1000hp motor in a direct (inline) application that drove a large centerhung ID fan. When it was being rebuilt in the shop it was realized that there was a 'roller bearing' outfitted on the drive end. When the plant engineer was notified that the shop wanted to convert the roller bearing to a ball bearing he contacted the manufacturer to verify the procedure. The manufacturers comments were that the motor should not be converted to a ball bearing. Instead we should intentionally misalign the motor .003" in order to 'load' the bearing radially.

Seemed pretty silly to me, but it may have explained why I've seen so many motors in these types of applications that have ran so well for so long. Maybe they've all been misaligned!!!