I ordered the wrong INA thrust bearings — then a ball bearing hunter showed me why

February 2024. I was standing on the plant floor, arms crossed, watching our newly rebuilt positioning station cycle through its first test run. I felt pretty good about myself for about 30 seconds.

Then the actuator started moving. And by "moving" I mean crawling. Like, painfully, slowly. About one-tenth of the speed our production schedule required. The operator next to me turned and asked, "Is it supposed to do that?"

That was the moment I remembered a question I'd brushed off two weeks earlier: how fast can a linear actuator move? My confident, wrong answer had been "fast enough." It wasn't.

Before I go further, a quick intro. I'm a maintenance lead at a mid-sized packaging plant. I've been handling mechanical service and replacement part selection for about seven years, and I've personally made—and documented—a respectable collection of mistakes totaling roughly $14,000 in wasted budget over that time. February added about $3,200 to the tally.

Here's how it happened.

The project: a simple upgrade with a hidden trap

We were upgrading a positioning station on one of our older conveyor lines. The line used standard ANSI #40 roller chain (per B29.1) on the main drive, and we were converting the station from a manual clamp to an automated push-pull mechanism. That meant adding a linear actuator. Not complex in theory: mount it, wire it, set the stroke, done.

I sized the actuator based on thrust load and stroke length. The actuator needed to push product boxes across a short transfer rail, then retract. The load was moderate—maybe 120 pounds including friction losses. I picked a 24V DC actuator with a screw drive, checked the force rating, and moved on.

For the pivots and mounting points, I specified INA bearings. That was the easy decision. Our plant has used INA roller bearings and needle bearings for years, and they hold up well in high-cycle applications. German engineering, dependable performance—that part of the decision never crossed my mind as a risk.

What I didn't do—and this is the part that still stings when I retell this story—is check the type of bearing carefully. I ordered standard radial ball bearings for a pivot that carried a significant axial load in one direction. And that, my friends, is exactly how you end up standing on a cold plant floor in February, watching your actuator fail its first performance test.

The turning point: a slow crawl and a quiet smile

The actuator wasn't the problem—it moved fine when we bench-tested it without a load. But in the real assembly, the INA bearings I'd chosen were binding. They weren't designed to take axial thrust, so the pivot was dragging. The motor was straining, the screw was fighting the friction, and the speed dropped to a crawl.

Let me pause here and say something that feels obvious in hindsight: bearings are load-direction-specific. Radial loads go to radial bearings. Axial thrust loads go to INA thrust bearings—that part of the selection process is straight out of Schaeffler's published engineering guidance. It sounds simple when you say it out loud. But when you're staring at a spec sheet, juggling a deadline, and the bearing box says "INA" on it, it's easy to assume you're covered. You're not.

"You're asking a radial bearing to do an axial job. It's like loading boxes into a truck made for passengers. It'll go, but not well."

I had our supplier on the phone, trying to walk through the problem, and they were patient with me. They asked for load direction, cycle rate, duty cycle. I had answers for none of those, which was the second clue that I'd skipped a step. That's when I caved and asked our senior technician—a guy we call the ball bearing hunter—to come take a look at the assembly drawing.

He earned that nickname for a reason. If a bearing exists that fits an odd application, he knows where to find it. He has grease-stained notebooks going back 25 years, and he treats bearing catalogs with the same reverence some people reserve for religious texts. He looked at my drawing in silence for maybe a full minute. Then he pointed at the pivot joint.

"That's your problem."

"The bearing?"

"The bearing selection." He pulled up a cross-section drawing of an INA thrust bearing on his phone and showed me the rolling element path, how the design handles axial load in one direction. "You're asking a radial bearing to do an axial job," he said. "It's like loading boxes into a truck made for passengers. It'll go, but not well."

I wanted to push back. Nobody enjoys being wrong in front of a guy who's seen everything. But he was right. I'd done the basic speed calculation for the actuator—screw lead times motor RPM—and I'd thought about cycle time, stroke, and sensor mounting. I'd completely skipped the part where the bearing carries the load direction.

The fix: standard parts, zero drama

We swapped the radial bearings at the pivot for proper INA thrust bearings rated for that axial load. The part itself was a standard catalog item. I could have ordered it in the first place if I'd simply confirmed the load direction.

Here's the part that still bothers me: the price difference between the wrong bearing and the right one was about $18 per bearing. Eighteen dollars. That tiny difference was the line between "crawls" and "works correctly."

The difference in performance was immediate. The actuator cycled at rated speed, the boxes pushed cleanly across the transfer rail, and the station passed performance testing on the first re-run.

But I never got these back:

• Three days of troubleshooting the original design
• About $890 in parts that couldn't be fully restocked
• Roughly 40 labor hours across three people
• A few layers of professional pride

Total damage for this one: around $1,200 in parts, a one-week delay on the line upgrade, and the quiet, knowing smile the ball bearing hunter has given me every time I walk past his workbench since. In terms of lessons learned, though, it was worth every penny.

What I tell people now

I should be clear about my lane. I'm not a bearing application engineer. I can't speak to load ratings, lubrication schedules, or service factors the way a Schaeffler application specialist can. What I can tell you, from a maintenance and production perspective, is this: your competence gets judged by whether the machine runs on day one.

Nobody remembers the 37 small things you got right. They remember the actuator that moved slower than a slug during a scheduled production run. That's the quality perception side of engineering—the work you deliver is the only evidence your team and your customers have of how careful you really are.

The INA bearings we installed in February have been running five shifts a week since then without an issue. The INA thrust bearings in particular look like they'll outlive the machine they're in. The brand was never the problem—my confidence was.

And yes, I eventually looked up the exact question I dodged: how fast can a linear actuator move? Short answer: it depends entirely on screw lead, motor speed, and load. Longer answer: if you ask your supplier for the speed-torque curve and read it before you place the order, you avoid becoming the cautionary tale I was.

One more thing, and this is the real reason I wrote this up. Everyone told me to verify bearing specifications before ordering. I only believed it after ignoring it and eating a $3,200 mistake. So here's my advice to you, from someone who learned the hard way: check the load direction, check the speed rating, check the duty cycle. It takes five minutes and saves a week of pain.

This worked for us in a fairly specific context: a 40 roller chain conveyor line, moderate loads, low cycle rate. If your application looks different—high speed, shock loads, heavy lateral forces—please talk to someone who actually specializes in this. Or do what I do now: find yourself a ball bearing hunter before you place the order, and let them check your work.

I've been maintaining our team's bearing selection checklist ever since. We've caught eight potential misorders in the past ten months using it, including one I almost placed for a $6,000 INA linear guide with the wrong mounting orientation. Some lessons stick.