The bearing in Figure 1 came out of a pump that operated approximately 24 hours a day, 7 days a week, for 18 months. At 3600 rpm, that equates to 2,712 million revolutions. Looking at it another way, that equates to more than 6,000 three-hour trips in a car traveling at 75 mph. That kind of use demands a lot from the oil, regardless of the application. Take a closer look at the bearing in Figure 1 and its outer race in Figure 2. The ball cage is in excellent condition; there is absolutely no wear on the cage.

Figure 1

Every bearing manufacturer will tell you that ball cages rarely fail unless there is a lubrication problem. Look at the ball path in the outer race shown in Figure 2 (between the white lines). There is no sign of pitting, spalling or contamination. The ball path has a nice dull gray appearance to it. It is easy to surmise from these two figures that the oil was in good condition and in sufficient quantity.

Figure 3

Figure 3

Look at the bearing in Figures 3 and 4 that had the same amount of service but didn’t fare as well. The raceway in Figure 3 is severely spalled and blackened. What’s more astonishing is the deformation of the rollers shown in Figure 4, especially the roller to the left. The two rollers should be the same size and shape. The roller on the right is closer to its original shape. See how the left-hand roller is barrel-shaped and how the metal has moved over the top edge? Notice the metal flakes on the left side of the left-hand roller, about a third of the way down. This metal was forcibly moved laterally toward the sides of the roller. The oil that came out of the bearing housing was severely discolored. Vibration analysis predicted that this bearing was going to self-destruct in short order. The proactive action to remove the pump from service kept the bearing from self-destructing and damaging the shaft and impeller. When the shop mechanics took one look at the severely deformed bearing they immediately concluded that the cause was from a lack of lubrication. “Surely. . . ” they said,“. . . what else could cause this kind of damage. Did you see the way that bearing looked? It must have gotten extremely hot to deform those rollers. Only a lack of lubrication can cause something like that.”

That was not the case, however. In fact, the only reason that the bearing did not seize up was because there was an adequate supply of oil to the bearing. It is possible that the oil degraded to the point that it could no longer lubricate the bearing, except for one small detail: The severely deteriorated bearing in Figure 3 had something in common with the nearly pristine bearing in Figure 1. Both bearings came from the same pump and shared the same oil reservoir.

There is a detail in bearing failure analysis that is commonly overlooked. When the rolling elements in a bearing deform, such as the roller in Figure 4, it isn’t necessarily because of a lack of lubrication. The fact that the oil is discolored and is probably oxidized still doesn’t mean that the oil was bad. In order for the rolling elements to deform, the bearing surface temperature has to exceed 300°F to 350°F. Those surface temperatures can, and will exceed 600°F when the bearing is overloaded; this was the case with this bearing. Of course the oil is going to degrade under those conditions. What many people fail to recognize is the fact that the weak link in a highly loaded application may be the bearing and not the oil. It has to do with the operating temperature of the bearing and the ability of the oil to efficiently remove the frictional heat load.

Overloading a bearing creates a boundary lubrication condition. Under a boundary lubrication condition, the bearing rolling elements will generate a greater amount of heat and the oil is responsible for removing that heat. With forced lubrication systems, where the oil is continually circulated under pressure, the desired oil temperature is maintained through the use of a heat exchanger. The circulating oil readily removes any additional heat that develops under a boundary lubrication condition. However, with a simple splash-lubrication oil reservoir, as was the case here, the heat that the bearing develops is not removed quickly enough under a boundary lubrication condition. This is because the oil that is splashed away by the rolling elements is replaced by oil that has been sitting next to the bearing in the reservoir and has heated up. There is no way to remove the added heat from the oil other than what is conducted through the reservoir housing. Therefore, the raceway and rolling element temperatures will increase. As the oil temperature rises, the surface temperature of the raceway and rolling elements also increases. Between 300°F and 350°F, the rolling element softens and deforms. As the rolling element deforms, the frictional forces increase and the surface temperature increases. The process feeds upon itself. Without an increase in the cooling capacity of the oil or a change in the oil’s viscosity, the bearing will continue to heat up as it continues to deform to the point of self-destruction.

What is the definition of a lubrication failure? For me it is simply the inability of oil to separate two surfaces and keep them from adhering to each other. Look closely at the deformed rollers. There are no signs of adhesive wear; there are no visible indications of scoring or galling. There are most certainly signs of metal movement. The only way those rollers could deform as much as they did was if there was metal-to-metal contact under some form of lubricating condition. Metal-to-metal contact without scoring is impossible unless there is a lubricant, a good lubricant. As a side note, the spalling on the inner raceway in Figure 3 is an indication of fatigue from loading conditions, and it is not a form of adhesive wear.

A reputable bearing manufacturer states in its engineering handbook:

“Most ball and roller bearings can be operated satisfactorily up to a temperature of 250°F providing bearing lubricant and cage properties are compatible with such temperatures. If bearings are to operate at higher temperatures, they must be subjected to special heat treatment (stabilization) to prevent unacceptable dimensional changes resulting from structural alterations of the bearing steel. Special heat treatment to obtain dimensional stability at elevated temperature tends to cause reduction in material hardness, resulting in decreased load-carrying capacity; that is, basic load rating is reduced. Therefore, when bearings are required to operate at elevated temperatures, it is advisable to consult with us.”

The unacceptable dimensional changes and structural alterations of the bearing steel that the bearing manufacturer is referring to are graphically illustrated in Figure 4. The bearing that deformed came from a pump modified to handle a bearing with a higher load rating. The modification was the result of a failure analysis performed on the pump. The mean-time-between-failure of that pump was six months. The higher-rated bearing lasted 18 months. The pump received an inordinate amount of engineering and inspection time and detail. Therefore, the overall picture was more readily understood for such a high-exposure machine. However, the majority of pumps in an industrial application rarely get such close scrutiny. The longer lasting bearing was fatigued, as evidenced by the spalling shown in Figure 3. It did not fail because of a lack of lubrication. To say that the lubricant was improper is a relative term when considering that the life of the pump was increased by 300 percent.

When deformed bearings are found, it is best to first consider the possibility that the machine is operating outside its design parameters, before blaming the failure on the oil. If the pump had been allowed to fail, the rolling elements would have turned into lumps of indiscernible metal and the evidence would have been lost. The scapegoat of choice would be the lubrication mechanic and he would have been hung - again.

Author Richard Adler has more than 25 years experience within the fields of maintenance and maintenance engineering. He has worked for several companies in the petrochemical, oil refining, specialty chemical and pharmaceutical industries. Photos and articles about actual failure analysis events can be viewed on his Web site: www.RESnapshot.com.

Photos © 2002 R.H. Adler