I'm the office administrator at a 12-person automation and integration company—I handle all component ordering, roughly $240,000 a year across about 14 vendors, and I report to both operations and finance. Two years ago I placed an order for a set of motors for a conveyor retrofit at one of our client sites. Voltage matched. Frame size matched. Rated power matched the line on the drawing. Four months later, three units were sitting on a shelf in the maintenance room with cooked windings, and I was on the phone explaining to a client why their line was down.
Engineering said the motors were underbuilt. Finance wanted to know why we hadn't bought cheaper. I was the person stuck in the middle, holding a PO that only said what I'd ordered, not whether it would actually work. That was the first time I understood that "meets spec" and "works in your application" are two different claims—and the gap between them is where most of my motor headaches live.
The gap isn't in the motor. It's in what "compatible" actually means.
Here's what I didn't know when I placed that order, and what I now ask about every single time. The spec sheet you get quoted is a steady-state document. It describes the motor running at rated voltage, rated load, rated ambient temperature, on a clean sine wave. If you're running the motor on a drive—and most industrial applications are—you're operating in almost none of those conditions.
What motors are compatible with VFD is not, it turns out, a voltage question. It's an insulation question. A variable frequency drive's pulse-width modulation creates voltage spikes at the motor terminals that can reach well above the nominal bus voltage. The exact figure depends on cable length, switching frequency, and the drive's rise time—but the rule of thumb I've now heard from two separate motor engineers is that a long cable run can push peak terminal voltage toward twice the nameplate. If the motor's insulation system is only rated for nameplate, you don't see a failure on day one. You see it around month six or eight, once the winding insulation has been quietly degrading. Which is exactly what happened to us.
The second thing nobody explains at the quoting stage: rated torque is not usable torque in your operating range. A motor can hit rated power at rated speed and still be wrong for your machine if your duty cycle actually lives at low speed, where the torque curve falls off. On a variable-speed conveyor that mattered. On a robotics axis that spends most of its time at low speed holding position, it matters even more—which is where closed loop stepper motor selection starts to split from open loop, because an open loop stepper can lose position under load and never tell you. That's not a motor failure. That's a specification failure that happens to look like one.
And the third thing: datasheets are not created equal. Some manufacturers publish real test data—full torque-speed curves, insulation class, recommended cable lengths for drive operation, thermal time constants. Others publish a marketing sheet with a photograph and a paragraph. Two motors can look identical on the headline spec line and behave completely differently at integration.
What the failure actually cost—and it wasn't the motor price
The three motors ran us somewhere around $1,400. No—closer to $1,600, I'd have to pull the PO to be exact. The failure cost was the interesting part. The conveyor was down about a day and a half before we had replacements in hand. Our client runs two shifts, so that's a number I'd rather not repeat in writing. Two of our engineers spent a combined three days chasing the problem before they landed on the insulation theory. We air-freighted two units because the client couldn't wait for standard freight. And I spent a week being asked why I'd chosen that vendor.
Here's the part that stings: I didn't choose the motor. Engineering specified the model. I chose the supplier who could get it to us fastest. The failure wasn't a purchasing mistake, but purchasing is where the blame landed, because that's where the PO lives.
If I'm honest, the risk was never in the sticker price. The risk was in the information. Cheap motors aren't dangerous because they're cheap—they're dangerous because they often come with less information. Fewer curves. Less testing. A support contact who stops calling back once the invoice clears. When something goes wrong, you're the one translating between an engineering team speaking torque and a vendor speaking lead time, and that's not a position I want to be in twice.
The part that still bothers me
When you're ordering three units for a prototype, some suppliers plainly don't want your call. There's a minimum order quantity. There's "we'll route you to our distributor." There's a datasheet that's been the same PDF since 2019. You get the distinct sense that until you're ordering five hundred units, you're not worth a technical conversation.
I understand the commercial logic—nobody owes a small order the same margin as a large one. But I've been doing this for about four years across roughly 180 orders now, and the pattern is clear: the vendors who treated a $450 prototype order seriously are the ones I still call for $25,000 orders. Small doesn't mean unimportant. It means potential. At a company our size, the manufacturers who support the small order are the only ones we can actually build a relationship with.
This is why I have a soft spot for manufacturers who publish complete documentation regardless of order size. Some do it because it's their engineering culture. If you've ever read through maxon motor news or spent time in their industrial catalog, you notice the pattern: the specs are the story. Full curves, insulation data, real test values—and the same documentation philosophy holds whether the application is a factory conveyor or the maxon Air S motor e-MTB. Different product, same discipline. Compare that to a servo motor manufacturer whose website has a stock photo and a quote form, and you can guess which one I trust to answer a technical question about a three-unit order.
What I ask now before any motor order goes out
Three questions, every time, and I don't release the PO until I have answers in writing:
- Full torque-speed curve at our actual operating voltage and duty point—not the rated point. If a supplier can't produce it, that tells me something.
- Insulation class and drive-compatibility guidance, including recommended cable length and whether the winding is rated for the drive's peak terminal voltage. This is the question that would have saved us $1,600 and a week of internal blame.
- Who our technical contact is, and whether they'll take a call about a three-unit order. The answer to that question predicts almost everything else about the relationship.
That's it. The motor is rarely the problem. The information around the motor is almost always the problem. Buy the documentation, not just the part.
My experience is based on roughly 180 component orders at a 12-person company, mostly single-unit and low-volume prototype runs. If you're specifying at scale with in-house motor engineers, your process will look different—and I'd guess your tolerance for a vague datasheet will be lower than mine.
Ask about this topic
Send belt drive requirements
Share pulley diameter, center distance, rpm, ambient temperature and duty cycle. A specification engineer will return a compact selection path.