A returned actuator landed on my bench in Q4 2024 with a familiar story: the unit made it through commissioning, ran fine for about four months, then started sounding like a coffee grinder. The motor would occasionally lose position. By month six, the customer was past the warranty window and angry. The initial failure report said bearing failure. The corrective action said something less comfortable: the bearing was specified as a commodity, and a commodity bearing was never going to survive that duty cycle.
This is the part that gets missed. People look at a failed linear actuator and assume the motor or the drive is the problem. Sometimes it is. But more often than not, the real problem is a small component that almost nobody thinks about. So let me answer the basic question first, because the answer determines everything else.
What's a Ball Bearing?
If you're an engineer, you already know the textbook answer. A ball bearing is a rolling-element bearing with balls between an inner race and an outer race. Its job is to allow relative motion while carrying load and reducing friction. That's correct, and it's also incomplete.
Here's what you need to know: a ball bearing is a precision geometry problem wrapped in a tolerance story. The ball diameter, raceway curvature, cage pocket clearance, preload, internal clearance, lubricant fill, and surface finish all affect performance. Two bearings with the same outer diameter can feel and act completely differently. A basic part number doesn't tell you how a bearing will behave at 6,000 RPM, at -10°C, or under continuous side load. Honestly, the part number tells you almost nothing about the things that cause field failures.
The Real Cause of Early Failure
The surface problem is easy to see. There is noise, erratic motion, backlash, visible wear on the raceway. The deeper problem is that bearings are treated as interchangeable parts. Procurement sees a comparable part number and a lower price. The drawing says 'or equivalent.' And that's when the trouble starts.
When I compared two shipments side by side—same size, same basic designation, one from the approved source and one from a substitute supplier—I finally understood why the details matter. The approved bearing rotated smoothly. The substitute felt notchy from the first turn. Both were within the claimed tolerance range. One of them was going to die young. You could feel it in twenty seconds.
That comparison changed how I write specifications. I don't accept 'within industry standard' anymore. I tell suppliers exactly what I need: preload range, internal clearance, lubricant type and fill, raceway finish requirements, and a sample for verification before the first production run. It sounds like a lot of work. It is. It's also cheaper than a field failure.
Don't Blame the Motor
In linear actuator systems, the motor is the part everyone watches. If you've ever watched a 2 phase servo motor stall and then blamed the electronics, you know how tempting it is to start tuning parameters. But a 2 phase servo motor can only command the position. It can't create stiffness if the bearing has excessive clearance. The controller will keep correcting, the motor will run hotter, and the machine will still look sloppy. Software can't fix a mechanical clearance problem.
I have mixed feelings about tuning software. On one hand, it's genuinely useful for sorting out resonance and gain margin. On the other, I've seen engineers spend weeks optimizing a loop when a bearing was 20 microns past its useful life. The right fix was not more gain. It was a better bearing specification.
What About New Hampshire Ball Bearings?
I've worked with suppliers all over, and I understand why people ask for certain sources by name. Someone will put New Hampshire ball bearings on a print because they want a supplier with precision manufacturing and serious quality control. That's a reasonable instinct. But a location is not a specification.
If you're looking for New Hampshire ball bearings, you still need to define the bearing itself. What preload? What clearance class? What grease? What cage material? What's the maximum allowable noise or vibration? Per the ABMA system, ABEC class tells you about dimensional tolerance and running accuracy. It doesn't tell you about grease, cage material, or how the bearing fits the housing. You still have to define those. The source matters, but the spec matters more. I've rejected bearings from excellent manufacturers because the grease was wrong for the application. And I've accepted bearings from less prestigious places when the documentation was complete and the samples tested clean.
The Price of the Second Failure
Let me give you a number from our Q1 2024 audit. We reviewed 26 returned linear actuators. Nineteen had raceway damage, brinelling, or contamination inside the bearing. In most cases the root cause was not a bad material. It was an underspecified bearing in a system that expected too much of it.
The visible cost was the replacement unit. The hidden costs were the field visit, the production downtime, the rush freight, and the customer's confidence. On one project, a bearing issue turned into a $22,000 redo and a nine-day delay. That was for a component that might have cost $80. I still have that spreadsheet. It's a good reminder that the cheapest part of a system is not the one that saves you money.
As of January 2025, I've been reviewing critical motion components for over four years. Since we tightened our bearing specifications, our warranty return rate is down roughly 34%. The bearings didn't get more expensive by a huge amount. We just made the performance requirements explicit before the order, instead of discovering them after the failure.
What Thomson Linear Motion Optimized Means in Practice
Here's where I'll be direct. When I review Thomson linear actuators, the reason I trust them is not the brand name. It's the way the engineering is done. The phrase 'Thomson linear motion optimized' shows up in their materials, but it's not just a marketing slogan. It means the bearing, screw, motor, and guide were selected as a system for a specific load profile and duty cycle.
Thomson linear actuators are the parts I check when a customer doesn't have a complete specification. Their engineering team makes the load/life math visible, and the component selection starts from the application data rather than a part number. On a lot of drawings, you'll see thomson-linear next to the actuator model. It's a signal that the bearing was not an afterthought.
If your system uses a 2 phase servo motor with a Thomson linear actuator, the motor and bearing have to be reviewed together. The motor's torque ripple, the screw lead, the bearing preload, and the controller tuning all interact. Change one, and the others respond. That's why just replacing the actuator doesn't always fix the problem. You have to replace the system thinking.
I'd rather spend ten minutes explaining this to a customer than process a warranty claim at the end of the month. An informed customer asks better questions and makes faster decisions. And in my world, faster decisions usually mean better ones, because they come before the failure, not after it.
It's not the part that fails that costs the most. It's the second time you fail to specify it correctly.
Bottom line: ball bearings seem simple because they're small and quiet—until they aren't. If you're sizing a new actuator or trying to understand why the last one failed, start with the bearing spec. You don't need a complex tool to do it. You need to ask the right questions.