At 2:47 on a Tuesday in early February, the maintenance manager walked into my office carrying a bearing wrapped in a shop towel. He didn't say hello.

'Thrust bearing,' he said. 'Mixer No. 3. We need one by Friday.'

I wrote down the part number, and as soon as the door closed I typed the phrase that took over the rest of my week: what causes thrust bearing failure.

A buyer with a parts-room vocabulary

I should be clear about who is writing this. I am not an engineer. I am the office administrator for a 180-person manufacturing plant, which in practice means I handle maintenance and repair purchasing—roughly $190,000 a year across nine vendors. I translate between production, maintenance, and finance.

Since I took over purchasing in 2021, I have signed for a lot of genuine SKF bearings. I can order a radial ball bearing from a catalog without calling anybody. I have bought ball bearing wheels for maintenance carts more times than I can count. I know SKF spindle bearings usually come as matched sets with precision-grade suffixes on the boxes. I know SKF needle bearings show up when someone needs radial load capacity in a small space. But the word thrust stopped me cold, because I had assumed, wrongly, that a thrust bearing was just a stronger version of a radial ball bearing. It is not.

What a thrust bearing actually does

The failed part was an SKF single-direction thrust ball bearing. In the catalog, it looks like a flat washer with balls running between two rings. It does one job: it carries axial load along the shaft, usually the downward weight of a vertical mixer shaft. It is not designed to handle side-to-side forces—the loads engineers call radial load. SKF's bearing basics section at skf.com describes a thrust ball bearing as a bearing intended for axial loads, not for radial support. Once I finally read that carefully, a lot of things made sense.

Here is the mental image that stuck with me. Press your palms together with a coin flat between them. The coin handles that axial force easily. Now slide one hand sideways while keeping the pressure on. The coin tries to tilt and slip out. A thrust ball bearing behaves the same way. It tolerates force through the shaft, but it does not tolerate a shaft that wobbles.

When a shaft starts moving sideways by even a few hundredths of a millimeter, the balls in a thrust bearing no longer roll in their intended track. They edge-load, skid, and generate heat. Soon the race starts spalling. The damage looks like a bad bearing. Often, the bearing was only the messenger.

The quote I almost sent back

Our regular supplier, an authorized SKF distributor, responded with a quote. It wasn't the lowest number I received. An online vendor had the same part number, a lower price, and faster delivery. But the distributor's quote included one line the others didn't: 'Before we ship this, I want to look at the failed part and the machine.'

Finance asked why we were paying for that. My answer, in slightly less professional language, was that guessing a bearing by its part number is only half the job.

The engineer didn't argue about the brand or the steel. He pulled the coupling guard, put a dial indicator against the shaft, and rotated it slowly. The needle moved about 0.1 mm. Then he checked the lower guide bearing—the radial ball bearing that keeps the mixer shaft centered. It had noticeable play.

The diagnosis was not 'the bearing failed because of too much thrust.' The diagnosis was that the thrust bearing had been forced to do a radial job as well. The lower guide bearing wore out, the shaft began to move sideways, and the flat thrust race received load on its edge instead of evenly across its surface. Replacing only the thrust bearing would have been the wrong fix. It would have failed again, and we would have spent another weekend explaining why.

So what causes thrust bearing failure?

In my experience, the practical answers fit into four buckets:

  1. Radial load or misalignment. A worn radial support bearing, a bent shaft, an out-of-alignment coupling, or a housing that lets a vertical shaft move sideways. This was our failure, and it's the one that seems most often forgotten.
  2. Lubrication problems or contamination. Wrong grease, not enough grease, or water and dust getting into the housing. Grease ages, so the regreasing interval matters too.
  3. Mounting mistakes. The shaft washer and housing washer are not always interchangeable. A rough mounting seat, a burr, or a housing bore that isn't square creates uneven load.
  4. Actual overload. The downward force or shock load on the bearing exceeds what the bearing was selected to carry. This can be steady load or sudden impact loads from the process.

There is an official way to talk about this too. ISO 15243:2017 classifies rolling bearing damage into six families: fatigue, wear, corrosion, electrical erosion, plastic deformation, and fracture/cracking. Our failed bearing showed fatigue and plastic deformation. But knowing which damage family you are in is only half the answer. The more useful question is what started the damage: load, alignment, lubrication, contamination, installation, or something else in the system.

The replacement went in Thursday afternoon, after we had replaced the worn radial guide bearing and squared the coupling. The mixer has run quietly since. Boring is good in MRO.

What I check now before I order

For the bearings I order regularly, the same load-direction logic applies. Most SKF spindle bearings for machine tools are precision angular contact ball bearings, and because they handle radial and axial loads together, they are usually supplied and mounted as matched sets. SKF needle bearings provide radial load capacity in a compact cross-section but do not replace a thrust bearing. A radial ball bearing is the workhorse for mostly radial loads. A thrust ball bearing handles axial loads. If someone offers you a cheap equivalent and cannot tell you which direction of load it was designed for, that is a red flag.

Here is something vendors don't always tell you: a replacement part number tells you what you need, not what killed the original. Anybody can sell you the same code. The harder, more useful service is telling you why the old bearing died, even when the reason is inconvenient.

Looking back, I should have asked one question before I made any phone calls: 'Is there anything else supporting this shaft?' If the answer had been 'only the thrust bearing,' I would have known we were already late.

The bottom line for anyone who lands here after searching what causes thrust bearing failure: check the radial support first. Ask what else is touching the shaft. And before you choose between quotes, ask what service is included in the price. Trust me on this one—the supplier who says 'let me look at the failed part first' is not trying to slow you down. They are trying to save you from doing the same job twice.