Troubleshooting Tips: Avoid Bearing Failures
| By: Kristin Lewotsky, Contributing Editor
What are some common causes of bearing problems?
Contamination, too much grease, grease incompatibility, misapplication.
—David Steen, Product Manager, AC motors, Baldor Electric Co.
Lubrication. With rotary bearings, you can pack them and usually they're good for life, unless it’s a really heavy force type application. With linear bearings, on the other hand, it can be much more difficult because they kind of distribute their grease. You either have to have the right kind of bearing that holds the grease or use the right kind of grease—if you use the wrong viscosity and the pressure is too high on the ball races, it'll break the grease down and basically turn it to black soup. That's a telltale sign that you’re exceeding the pressure that that grease can take. What's odd about it is that when you look at it the bearing, it looks like it got really hot but there's very little temperature change at all. It's just that the pressure breaks down the lubricant and changes its characteristics.
—Jim Monnich, Engineering Manager, Parker Hannifin
Watch the minimum temperature rating of the grease—cold operation can change the grease into something more like wax if the wrong grease is being used.
— Donald Labriola, President, QuickSilver Controls Inc.
Overloading, impact loads, and binding from misalignment.
—Daniel Lutz, Lead Engineer, item North America
What are some common mistakes engineers make working with bearings?
Lubing a no-lube slide bearing—the grease only serves to catch dust and grit and will cause a premature maintenance issue.
I've been in a lot of design reviews in which we were coming in to replace something that someone had designed and the first thing we see is they’ve got their bearings over constrained. The simplest example is a shaft with two deep-groove ball bearings, one on the front and one on the back. You can do the free body diagram—put a load on the end of the shaft and then based on the distance of that load from the first bearing and the space in between the two bearings you can calculate what the reaction loads are. That’s a statically determinate equation. If you were to put a third bearing into that system you would necessarily be over constraining it. To put that third bearing on, you’d have to be bending the shaft in some way and the shaft stiffness would come into play.
Different bearings serve different purposes in terms of the loads they can take. A deep-groove ball bearing is really only intended to take a radial load. You can counteract a moment load using two deep-groove ball bearings spaced some distance apart, or you can use a crossed roller bearing, which is really intended take a moment load. The problem is that a lot of people use a crossed roller bearing and then at the end of the shaft they put another deep-groove ball bearing. That will again over constrain the system and a lot of times you'll burn out the weaker of the two bearings, which is typically the deep-groove bearing.
—John Rogers, Director of Engineered Solutions,Yaskawa America Inc.
A large number of bearing problems are a result of misapplication. Understanding the application and environmental issues can greatly increase the bearing life and ultimately the life of the machine.
You have to machine the linear bearing mounting surfaces properly so that you don't have a bearing that is overloaded due to uneven mating surfaces. The grade of bearing is another major thing to keep in mind. There are bearing height tolerances and if you use a low tolerance bearing, the assembly may bind and overload the bearings so you'll have high friction and premature failure.
How do you isolate bearings as a cause of problems?
Because bearings are isolated inside the motor, there will typically be some outward effect that indicates there is a bearing problem, such as heat, noise, or vibration. Many larger motors utilize monitoring devices such as RTDs or vibration monitors to determine bearing failures prior to a catastrophic failure occurring.
Radial bearings tend to make more audible noise than linear bearings do, usually because you're rotating at a pretty good speed. Linear bearings are much quieter. If you're running a linear stage back and forth fairly quickly, you might hear something odd—sometimes you might get a warning with noise or vibration—but typically that's not the case. Typically it's a failure that comes out of the blue. Basically the force on the bearing track suddenly goes way up and the balls either stop recirculating or they turn into flats or something like that.
Noise and increased friction.
If you record the speed data and then do an FFT and look at the frequency content, you can see frequencies associated with the rolling elements for that bearing. Let's say you’re running at a constant speed on the shaft. Given the bearing size, you know the rollers are going to be rotating at a given frequency. If you hear something, you can put an accelerometer on the system and look for a peak [in the spectrum]. If you have an issue with, say, an inner or outer race, that'll occur once per revolution of the shaft and a peak at another frequency, will appear if you have a problem with a ball.
A microphone and a PC-based FFT analyzer can help locate which bearings are involved, given some knowledge of the bearings, by displaying the frequencies composing the noise.
If it's a linear bearing, we check the alignment on the rails and make sure they aren't out too far. The condition of the grease is another telltale.
What should engineers keep in mind when specifying or designing in bearings?
Some chopper drive designs with motors can induce sufficient current in the bearings to cause pitting and life issues. The life is dependent on load and temperature and number of rotations—keeping motors cooler can significantly extend the life of the bearings.
If the application is well understood, the most economical bearing that will get the job done and provide the longest operating life is desired. Thermal instability is one example of a difficult environment. Use heat stabilized bearings and high temp grease.
A bearing should always be preloaded so that you’ve ensured the balls are rotating and not sliding. If you don’t preload a bearing, there’s a potential for the bearings to be relatively noisy from instability and even worse, the balls slide in the raceway groove which will prematurely wear the bearing.
—Alex Garcia, applications engineer, New Hampshire Ball Bearings
You want to be able to know deterministically what your loading is on the bearing. Do the calculations—that's the quickest path to success.
Pay attention to operating temperature. We had an issue with a linear guide that had a ball bearing separator chain between the balls on the linear guide system. The customer had a robotic system that was operating at around -10°C and never had a problem; then he went down to -20°C and after a short period of time the ends of the bearings blew off. The chain inside the bearing had become embrittled by the cold temperatures and failed. As soon as it did, the spacer chain material jammed the ball recirculation, causing the bearing to be destroyed.
What is the biggest mistake engineers make trying to specify or integrate bearings?
Using off the shelf motors for special applications. Not getting application issues addressed in advance. Continuing to replace bearings without isolating the root cause.
In the miniature world, one of the things we run into quite a bit is people overpress a bearing, which means that the shaft is too big for the bore or the housing is too small for the OD of the bearing, so that you’re actually squeezing the balls before you put any load on them.
Care should be taken when considering radial play. There should be a some play in the bearing after it is mounted so that you’re not squeezing the balls. The appropriate way to eliminate the radial play after it is mounted is to put an appropriate preload amount on the bearing to ensure the proper function of a bearing system.
I think probably the biggest thing is that they size it wrong. Typically, it's undersized—we see people take a ball screw and undercut the end of the ball screw and put a small little bearing on it. You have to size your bearings appropriately to the load.
The old school way is to look at it and say that bearing looks big enough, and for the most part you could be right, but I think sometimes people overdesign it and put too large a bearing and that can hurt your budget.
What’s the biggest mistake they tend to make in terms of isolating faults?
Assuming all vibration is bearing related.
The problem a lot of engineers have is that they present symptoms; they don’t present the actual failure mode. They’ll say the bearing is noisy but why did it become noisy? Was it over pressed? Was it too much preload?
What’s your bearing horror story?
Encountering the same failure on a large group of installed motors.
One of the issues we had was with a support radial bearing. There was slippage between the inside diameter of the bearing race and the inner diameter of the shaft. This was a very-high-speed, high reversal application—it was reversing at least two or three times a second. Within three or four months of operation, the whole shaft would actually be reduced in size by 0.01 to 0.02 in. There would be lots of rust and contamination and of course we were scratching our heads because we had used good solid practices as far as putting those two pieces together.
What was occurring was that as the shaft would rotate one direction everything would be fine, but when it would reverse at high speed, although the two shafts were pressed together, there would be a very small slip, on the order of hundredths of a degree. Over time, this small slip would produce wear between the two surfaces and generate contamination and would reduce the fit between the two parts, so with each reversal, the two parts would slip a little further. This process would escalate into a runaway situation until the shaft was so reduced in size that it would make noise in the application. The bearing was fine—the shaft was just eaten away.
The solution was twofold—one was to use a tighter press so it couldn’t slip but a problem that you have with that is if you go little bit too far you can actually deform the ID of bearing and cause a premature bearing failure. We put a bearing thread locker on it. The problem with bearing lockers and pressed shafts, though, is that when you put glue on and you press the shaft in, all the glue just gets pushed off. What we did was make a small cut around the shaft of about 0.010 in. where the glue stayed put. Once we did, that the problem was totally eradicated.