How many times have you been part of the following event? A piece of equipment fails. It is removed from the machine and sent to the shop for repairs. The shop machinist opens the oil drain and black goo pours out. The machinist says, “Yes, the oil was bad. You should’ve seen the bearing; it looked bad too. Looks like your lube mechanic was sleeping on the job again.” You proceed to question the lube mechanic. All of his answers seem to be correct and it appears that he had just changed the oil in the pump, or at least that’s his story. The lube mechanic already got the word that the pump had failed. He knows that high-profile machines always attract attention, and when they fail, he is always the first person to be rounded up in the “usual line of suspects.” He knew you would be coming.

Figure 1

The lube mechanic doesn’t offer any other helpful information, and all that you’ve accomplished is to annoy him. Now let’s take a closer look at the evidence that should have been checked first, the evidence that was thrown away in the metal dumpster. Looking at Figure 1, it’s obvious that the bearing became so hot that it lost its temper and the metal became became soft enough to deform. The subtle colors on the raceway to the left indicate that the surface temperature reached between 450ºF (232ºC) (dark straw) to 600ºF (316ºC) (medium blue). The dark purple color is 500ºF (260ºC). That’s hot. Most bearings will begin to lose their temper at 300ºF (149ºC), and discolor at 350ºF (177ºC) (light straw). The discoloration is important because it shows exactly where the oil was under severe mechanical distress. Now look at the color of the raceway to the right. It has a nice dull gray color to it, signs of good lubrication. There wasn’t a boundary lubrication condition on this raceway; there was a good oil film, and it did not get hot. This observation is perplexing because if the cause of the bearing failure was due to a lack of lubrication, or the wrong lubricant, why then does the raceway on the right look so good?

Figure 2

Take a closer look at the left raceway in Figure 2. The purple and medium blue colors are unmistakable. This bearing became very hot at the fragile interface between the raceway and the ball. Look at the balls in Figure 3 that came out of this bearing. The left ball came from the left raceway and the right ball came from the right raceway. There is a remarkable difference between them, especially for two balls sharing the same lubricant in a static reservoir and less than a couple of inches apart. It is obvious that even with good oil, the film was due to fail because of the high surface temperatures. Which came first in the sequence of events? Did the oil degrade until the oil film could not support the balls and the friction forces generate greater heat? Or did outside forces overpower the oil film that subsequently caused the same sequence of events? If the failure was due to degraded oil, then the raceway on the right should have shown greater discoloration from the distress, and it doesn’t. If the radial and axial loading was normal, then the bearings in Figure 3 should have an equivalent appearance, and they don’t. What occurred here? Nobody usually asks these questions or breaks a bearing apart to see why it failed. Most of the time, the evidence lands in the metal dumpster. All anyone ever sees is the grossly discolored oil.

Figure 3

Fortunately, the question was asked and the answer was found. It was not the oil that caused the bearing failure, it was cavitation of the pump. Cavitation sets up tremendous and unstable thrust loads in the axial direction. The thrust load in this bearing was carried by the left raceway. The right raceway carried the radial load. Although the radial load was not as great as the thrust load, it was still substantial during cavitation because of the tremendous unbalance forces that were generated. The findings were issued to management and were reviewed with the lube mechanic. The mechanic got in a well-deserved “I told you so.”

One part of performing a failure analysis investigation consists of determining whether the equipment was operating outside of its design parameters. It is important to find corroborating evidence during the investigation. A well-lubricated bearing will usually protect a lot of the corroborating evidence that would normally be destroyed. The amazing part is the amount of heat that the oil can withstand in the final hours of a failing machine. It is a losing battle for the oil, but it is an amazing battle. If the time is taken to recover the bearings, there may be a surprise in store. The forensic evidence usually reveals the true cause of the failure.

An old paradigm that is hard to overcome is that when bearings come out looking black and disfigured, the problem had to be caused by a lack of lubrication. If you’ve ever wondered why some lubrication mechanics have bad attitudes, it’s because of these armchair quarterback failure analyses. Give them a break and look at the bearings first. You might be surprised. You’ll end up gaining the respect of your lube mechanic who is your best ally in the front line of preventive maintenance.

How Metals Get Their Hardness
To say that bearings can lose their strength at 350°F (177°C) sounds preposterous when iron isn’t molten until it reaches 2800°F (1538°C). However, one reputable bearing manufacturer (not the manufacturer of the bearings in Figures 1 and 2) actually derates its bearing at temperatures starting at 340°F (171°C). At that temperature, the design radial load rating decreases by five percent. At 390°F (199°C), the radial load rating drops by 15 percent. By the time the bearing operating temperature reaches 570°F (299°C), the design radial load has decreased dramatically by 40 percent!

Why does it get softer? Because of a microstructure known as martensite; the more martensite there is within the metal, the harder the material. Manufacturers get the martensite by rapidly quenching the molten metal to a temperature roughly under 700°F (371°C). The benefit of martensite is that it is very hard (typical hardness 324,000 psi); the disadvantage is that it is quite unstable, as well as very brittle.

If martensite is encouraged with a little heat, it will transform into other microstructures. Martensite will transform into other microstructures most quickly between 300°F and 390°F (149°C and 199°C). Some manufacturers will purposely soften, or temper the martensite in this temperature range to get a compromise between strength and toughness. Some of the other microstructures that are formed, such as cementite or bainite (typical hardness 247,000 psi), are softer and more desirable.

Remember that dark blue color on the bearing raceway? At that temperature, 600°F (316°C), the martensite at the surface of the raceway transformed into something softer and lost about 10 to 20 points on the Rockwell C
hardness scale.

The Rockwell hardness test is a method used to determine the hardness of metals by indenting them with a hard steel ball or diamond cone. A light load is applied first, then increased to a specified higher load, measuring the additional depth of penetration.

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