What Causes Thrust Bearing Failure? Notes From the Person Who Orders the Replacements

On a Tuesday afternoon last March, I got an email that's become depressingly familiar: "Maintenance needs another 51210. The indexing table is down."

Not the first time that year. Not the second. The same 51210 thrust bearing, failing on the same machine, for the third time in fourteen months.

I've processed that part number so many times I could order it blindfolded. Which is exactly the problem.

Before I go further, let me explain where I sit. I'm an administrative purchasing manager for a mid-sized manufacturing plant—about 400 employees across three buildings. I manage MRO procurement, roughly $400,000 in annual spending across 8 vendors. Bearings are a big slice of that. When a maintenance tech says "we need a bearing," I'm the one who sources it, buys it, and later explains to finance why we blew the maintenance budget.

I'm not an engineer. I've never rebuilt a gearbox or set up a lathe. But I took over this purchasing role in 2020, and since then I've processed somewhere between 70 and 90 bearing orders a year. I've watched the same failure patterns repeat often enough to draw some honest conclusions.

So, what causes thrust bearing failure? After five years of watching it happen, here's my answer.

The surface problem: "The bearing keeps failing"

When a bearing fails, the default explanation is almost always the bearing itself. "Bad part." "Cheap junk." "Wrong brand." I've heard it all, and early on, I repeated some of it. It's the easiest story to tell: a component broke, so the component must have been bad.

Then I ran the numbers. We had one machine burn through six identical thrust bearings in fourteen months. Different production batches. Different purchase orders. If the parts themselves were defective, we'd be looking at lottery-level bad luck. There had to be another explanation.

The bearing wasn't the root cause. It was the failure point.

When you hear hoofbeats, you think horses, not zebras—but you also check the ground. We never did.

The question we should have asked from day one

Every new buyer learns to ask two questions: which part do you need, and how fast can you ship it? Those are logistics questions. They're not diagnostic questions.

The question that actually matters: what did the failed bearing look like?

For a long time, nobody in our plant could answer that. The failed parts were already in the scrap bin. "It was rough when we turned it by hand," one tech told me. That's not a diagnosis—that's a euphemism for "we didn't really look." I don't blame the tech, by the way. When the machine is down, you replace the part. No one has time to investigate on the spot.

But if you don't investigate, you don't learn. And if you don't learn, you reorder the same part, install it, and wait for the next failure.

I eventually put together a one-page failure log and told the team to photograph every failed bearing. It took a few months of nudging, but once the photos started accumulating, a pattern emerged. And the pattern pointed somewhere unexpected.

The deeper causes behind thrust bearing failure

I'll explain what I've seen in plain language. I'm not a tribologist, but I've talked to people who are—and I've read NSK's catalog literature carefully enough to understand what the rated life numbers assume.

Misalignment and edge loading

A thrust bearing is designed to carry axial load evenly across its rolling elements. In theory. In practice, the bearing sits between a shaft and a housing, and if either one isn't square to the other, the load concentrates on one edge of the raceway. A few rollers end up carrying nearly the entire load.

This was exactly our situation on the indexing table. A previous rebuild had left the housing face slightly out of square—maybe two or three thousandths of an inch. The 51210 has more than enough rated capacity for that application in perfect alignment. With the misalignment, it was edge-loaded on every cycle. The raceway flaked like the bearing was overloaded by a factor of two. (Should mention: that machine had run fine for twelve years before the rebuild. The rebuild introduced the problem, not the bearing.)

Here's the frustrating part: the fix wasn't a heavy-duty bearing. It was a shim and a machining pass on the housing face. Total cost around $180. We'd spent over $2,000 on replacement bearings and downtime before someone thought to check the mounting surfaces.

That's the insight I want to hand to anyone repeatedly replacing thrust bearings: if the same bearing keeps failing, measure the alignment before you order another one. The bearing might be the victim, not the criminal.

Lubrication starvation

Thrust bearings need the right lubricant, in the right amount, delivered continuously. Simple in theory. In practice, I've seen failures caused by:

• A grease gun filled with the wrong grease after the tube of the right one was empty
• An oil level checked while the machine was stopped, which reads higher than it does when the machine is running
• A blocked feed line that hadn't been cleaned since the machine was installed

None of those appear on the part spec. All of them appear as "failed bearing" on a work order. And all are cheaper to prevent than to recover from.

Contamination and careless storage

Bearings arrive from the factory sealed in moisture-proof packaging. They don't stay sealed when someone leaves an opened box on a shelf in the maintenance room. Dust and humidity get in, and the bearing runs with fine abrasives in its raceways. It fails fast, and it looks cheap, even if it's a premium NSK bearing.

We now keep bearings in a clean cabinet away from the washdown station. Sounds basic, but it's made a real difference in how long certain replacements last.

Preload problems

Some thrust-bearing applications need precise preload. Too little lets the rollers skid. Too much builds heat and cooks the grease. Both look like premature failure. Neither is the bearing's fault.

If your installers set preload "by feel," you're gambling with rated life. I do not say that to insult anyone—I say it because I've seen the same installer do a perfect job on Tuesday and a rushed one on Friday.

What this really costs

Let me make the business case, because that's my side of the job.

One unexpected thrust bearing failure at our plant runs roughly:

• $200–900 for the replacement bearing, depending on type and size
• $3,500+ per hour of machine downtime on a production line
• $150–400 in expedited shipping, because nobody plans for a failure
• $600–1,500 in after-hours labor, with overtime and call-out rates

Add the invisible costs—or rather, the costs that don't show up on a purchase order: rescheduled production, a late customer order, an awkward conversation with your boss. I've had that conversation enough times to know how much it damages the credibility of an entire team.

On the worst machine, we spent more than $30,000 on bearing-related issues in a year and a half. That's a third of my annual parts budget for the whole plant, consumed by one repeating problem.

The supplier factor

Here's a hard truth about the supply side. If your distributor only asks "which part, and when do you need it?" you've got a logistics vendor—useful for moving boxes, not for solving the problem that's moving your boxes.

In our 2024 vendor consolidation project, we made failure-analysis capability a selection criterion. The distributor we work with didn't just process the order for the replacement 51210. Their applications engineer asked, "why did this one fail?" That one question redirected us to the alignment issue, and the recurring failures stopped.

An informed customer asks better questions and makes faster decisions. I'd rather spend ten minutes explaining a failure mode than place my fourth emergency order for the same part.

What actually helped us

These won't fit every facility, but they're cheap, and they address the root cause rather than the symptom.

1. Photograph every failed bearing before disposal

Ten seconds with a phone creates evidence. If the same failure appearance shows up twice, you have a pattern. A pattern is the beginning of a diagnosis.

2. Ask your supplier what the failed part means

Most reputable bearing distributors have engineers who can read spalling, smearing, and discoloration like a detective reads tread marks. Ask them. If they can't help you, question whether they're the right partner.

3. Verify the specification, not just the part number

We've caught more than one instance where the wrong bearing class was installed—right dimensions, lower load rating. No surprise when it fails early. Confirm the spec against the OEM's requirement before you reorder.

4. Write down the failures

My spreadsheet has a row for each incident: date, machine, part number, what the failure looked like, who replaced it, and what we checked. It's not fancy, but it's been worth more than any single engineering recommendation I've received. We didn't have this process at first, and it cost us months of avoidable failures.

5. Don't replace the same part twice without asking why

If you're about to order the identical bearing for the identical machine, take a breath. Spend that extra day investigating. The next failure you prevent will pay for a lot of extra days.

The bottom line

Honestly, I'm not sure why the flaking on a thrust bearing forms the exact crescent shapes it does. It's one of those details I've accepted without fully understanding. If a lubrication engineer reads this and wants to explain it in plain English, I'd genuinely welcome the education.

But I do know this: after five years of buying NSK deep groove ball bearings for the motors, sealed spherical roller bearings for the dusty equipment, Dodge pillow block bearings for the conveyors, even the occasional crossed roller bearing for a precision indexer—the most persistent lesson is that the bearing is rarely the actual problem.

When a thrust bearing keeps failing, look at the machine. Look at the installation. Look at the lubrication, the alignment, the storage, the preload. Look at everything around the part that creates the conditions for failure.

That's where the cause lives. Finding it—and buying the right part, and installing it properly—is the whole game. At least, that's been my experience after 5 years of ordering the replacements.