What Causes Thrust Bearing Failure? A Maintenance Checklist for INA Bearings Repairs

Back in 2017, I was the guy who ordered replacement bearings by reading the old part number and clicking order. That worked until it didn't. A thrust bearing failure on a mixer line ended up costing us around $3,700 in downtime and rework—maybe $3,200, I'd have to check the work order. The replacement bearing was correct, but the shaft shoulder had been machined wrong. The bearing was the symptom, not the problem.

Since then, I've changed how I handle bearing-related repairs. This is the checklist I now use before ordering anything from an INA bearings website or an INA bearings catalog. It's five steps. Step 4 is the one I skipped for years.

If you're dealing with a failed thrust bearing, a noisy ball screw, or wondering whether to replace a ball screw with an electric linear actuator, this is for you. It won't tell you which part to order. It will tell you what to check before you spend money.

Step 1: What causes thrust bearing failure? Inspect the actual bearing

First, answer the question: what causes thrust bearing failure? In my experience, the usual suspects are contamination, misalignment, incorrect lubrication, and overload. But you can't just pick the one that sounds plausible. You have to look at the bearing and the machine around it.

I now check for four signs:

• Spalling on one side of the raceway: thrust load is offset, not perfectly along the shaft axis.
• Brown or blue discoloration: lubricant broke down from heat.
• Pitting and rust: water or coolant got into the housing.
• Brinelling: something was impact-loading the bearing, often while the machine was stopped.

That last one was my downfall. In 2018, I saw galled rollers and assumed contamination. I replaced the bearing, and three weeks later it failed the same way. Turned out the machine was creating axial oscillations at startup because a clutch was engaging too fast. The bearing was just the witness. If you skip this step, you're essentially paying to have the same failure again.

Step 2: Use the INA bearings catalog, not your memory

Once you know the failure mode, you need actual dimensions. The INA bearings catalog isn't a suggestion. It lists boundary dimensions, load ratings, and the little mounting notes that catch people. I used to think I could rely on part numbers. That's how I ordered 30 INA needle bearings in 2019, and half didn't fit. The marking on the old bearing was partially worn. The bearing was a different width than I assumed. The order cost maybe $1,100, and the rush replacement cost more than that.

Now I open the catalog every time. The INA bearings website is a good quick check, and the searchable data is useful, but I keep the catalog PDF saved offline because it has mounting tolerance notes. Those notes have saved me more than once.

Step 3: Measure the housing and shaft before you search

Catalog load ratings don't mean anything if the shoulder against the thrust bearing face isn't perpendicular. I measure shaft diameter, housing bore, and fillet radius. The fillet one is easy to miss. In 2021, I ordered a bearing that matched the part number exactly, but the shaft fillet was too large. It seated against the fillet, not the shoulder. The bearing lasted a week. My foreman asked, 'Did you measure the fillet?' I hadn't.

Measuring isn't glamorous. It takes ten minutes. I've lost more time to one wrong bearing installation than to all the measuring I've done combined.

Step 4: Make ball screw repair a measurement decision, not a gut decision

This is the one I'd always gloss over. Ball screws seem simple: if it's noisy, replace it. But a noisy ball screw doesn't automatically mean the screw is shot. The problem could be in the return tubes, the preloaded nut, worn wipers, or the support bearing. I've handled ball screw repair cases where the fix was just cleaning, new balls, and proper preload. I've also seen screws that were bent beyond saving, and no repair would bring them back.

My rule now: measure before deciding. Put a dial indicator on the screw, check backlash, and inspect the ball tracks. If the track is polished but smooth, you may only need a clean and re-lube. If there's flaking or visible dents, that screw is done. And here's the trap: if the screw is repairable, replacing it is usually more expensive. If the screw is scrap, repairing it is even more expensive. The measurement isn't a preference; it's the branch that tells you which path is actually cheaper.

When someone asks about ball screw repair, I ask to see the inspection report. If they don't have one, that's a red flag. A good repair shop will tell you what they measured, not just what they plan to replace.

Step 5: If an electric linear actuator is in the conversation, compare total life cost

Electric linear actuators get a lot of attention, and sometimes they're the right solution. But I've watched too many teams swap a ball screw plus motor setup for an electric linear actuator because the actuator looks simpler. Simpler isn't always more reliable.

In 2022, we replaced a ball screw assembly on a filling line with a linear actuator. The actuator was about 20% cheaper than the ball screw assembly quote. Six months later it failed. The lead screw was corroded and the plastic guide ring had worn down. The unit wasn't rated for washdown duty, and our line gets sprayed every night. The replacement actuator with proper IP rating and stainless steel parts cost roughly 2.5x what we'd 'saved.' On top of that, the downtime cost more than the parts.

So now I don't ask 'which actuator?' I ask about duty cycle, ingress protection, mounting alignment, and daily cycles. Half those answers aren't in the brochure. If someone quotes an electric linear actuator without asking about washdown, I don't trust the quote.

Notes and common mistakes

Three mistakes I still see on other people's machines:

1. Believing the part number alone. Part numbers get revised. The INA bearings website has supersession data, but I still verify in the catalog. In 2023 I was told a new number superseded an old one, but the new suffix changed the cage material. It mattered.
2. Forgetting to inspect the new part. I found corrosion on sealed bearings that had been sitting in a warehouse. It's rare, but it's the kind of rare that ruins your schedule.
3. Ignoring the lubrication path. Thrust bearing failure is often blamed on the bearing, but the real cause is oil starvation because a pump lost priming, or an oil level was set too low. The bearing isn't the first thing to fail; it's just the first thing you see.

One caveat: this checklist is tuned for our old food processing line—constant load, moderate speed, daily washdowns. If you're running 20,000 rpm spindles, you're dealing with a different world. And honestly, I'm not sure why one of our actuators survived for years while its twin model failed twice. Some of this is experience, some is luck. A checklist just makes the experience part repeatable.

Bottom line: the cheapest repair looks good until the machine stops again. I don't care if a part is 20% cheaper if it fails in a year. That's not moral outrage, it's arithmetic. A $300 bearing that fails in six months costs more than a $450 bearing that lasts four years, because the labor and lost production dwarf the price difference.