Quick answer: the component you're ordering is rarely the real risk. The real risk is everything you're connecting it to.
I'm a maintenance lead at a packaging plant, and I've been ordering motors, capacitors, actuators, and controllers for about seven years. I'm not an electrical engineer—I mention that because the mistakes below are the kind an engineer would've caught from across the room. Since 2018, I've personally made and documented four significant ordering mistakes, totaling roughly $4,900 in wasted budget. Not one of them was caused by a bad part. They were all caused by me buying a part that didn't match the system it went into.
If I could give every new maintenance buyer one rule, it would be this: the most expensive spec on a motor or actuator is the one that connects it to everything else you already own.
The $26 capacitor that turned into a $400 emergency call
In September 2022, our air compressor motor started humming instead of starting. The motor—a 1.5 HP Leeson, nothing exotic—had a swollen capacitor on top. I did what half the internet does: I searched for a leeson motor capacitor 80 µf and ordered the first cross-referenced listing that claimed to fit.
It fit. That wasn't the problem. The problem was the number I didn't copy.
The original cap had two ratings on the side: 80 µF and 330 VAC. The replacement I bought was 80 µF but only 250 VAC. They look the same. Same terminals, same diameter, same price range. I installed it on a Wednesday, and the compressor ran fine for two weeks. Then a Friday-night shift called me with a burnt electrical smell and a motor that wouldn't start at all.
The replacement cap had failed, and the start winding had gotten hot enough to make the motor smell like a burned-out clutch. I'm not a motor designer, so I can't give you the full electromagnetic explanation of why voltage rating matters beyond the obvious. What I can tell you from a maintenance perspective is that a start cap sees voltage spikes, and an under-rated cap eventually loses. And when it loses, it can take the motor with it.
The math stung more than the failure: the capacitor cost $26. The emergency electrical call was $350. And the line was down for four hours on a weekend, which our production manager kindly reminded me was worth more than all three of my previous mistakes combined. The leeson-motor wiring diagram that showed the correct 330 VAC rating was online the whole time. I just hadn't opened it.
Lesson: capacitor ratings are not “fits 1.5 HP.” They're microfarads and voltage. If a listing says 80 µF, keep reading until it tells you the VAC.
This is somewhat embarrassing to admit, but the same pattern repeated a few months later with a bigger motor.
Five horsepower is an output, not a bolt pattern
In May 2023, a pump motor on our washdown system failed. Bearing noise, high temperature, the whole story. The motor was a C-face Leeson, 5 HP, 1745 RPM, driving a hydraulic pump through a coupling. I found a motor online with “Leeson 5HP motor” right in the title, 1745 RPM, and a price that beat everybody else. I ordered it before lunch.
The pump couldn't use it.
What I'd missed was frame and mounting. Our pump needed a 184TC motor—T-frame, C-face, with the flanged pilot that locates the pump. The motor I ordered was a foot-mounted 184T. Same NEMA family, but no pilot, no bolt holes for the pump adapter, and a shaft that was close-but-not-close-enough for the coupling hub.
Our lead fabricator said he could make an adapter. He could, and he did. It took a day, it added vibration we could feel in the pipework, and it undermined the whole reason we'd bought a new motor in the first place. We ended up ordering the correct C-face motor, paying a 25% restocking fee on the wrong one, and eating the labor.
Here's what I now check before any motor order, in this order:
- Frame (184TC vs 184T vs 182T)
- Mounting (C-face, foot, or both)
- Voltage and phase (208 vs 230 vs 460—not “same” until you've checked)
- Service factor and duty rating
- Enclosure (TEFC vs ODP matters more than you think in a washdown area)
If you take a photo of the old motor's nameplate and send it to the supplier, you skip all of this. I was too rushed to take a photo. That rush cost more than the motor.
Electric actuator manufacturers quote thrust, not compatibility
At the start of 2024, we were retrofitting a packaging line and replacing a pneumatic cylinder with an electric linear actuator. I compared a few electric actuator manufacturers the way I'd been trained: stroke length, thrust rating, price. We picked a unit with more than enough force, and it arrived on schedule. That part was fine.
Then came the controls.
The actuator ran at 24 VDC and came with limit switches but no position feedback. The manufacturer's linear actuator controller was a forward-off-reverse unit, and it added $180 to the kit. I decided we didn't need it. We had a generic PWM controller on the shelf from a previous project, so I used that instead.
It worked in testing. It failed in production—or rather, it didn't fail, but the actuator did. The generic controller couldn't interpret the actuator's internal limit switch, so when the gate reached the end of its stroke, the controller kept sending power until the limit switch fried. That happened on a Thursday, in the middle of a 14-hour run. The gate jammed, the line stopped, and we learned that “24 VDC” is not a complete controller specification.
An actuator and its controller are one system. If you buy them from different sources, you own the integration risk. That risk is real, even for something as simple as a limit switch. I'm not an automation engineer, so I can't speak to every feedback protocol out there. What I can tell you from the maintenance side is that the actuator manufacturer's own controller would've been $180 up front and about zero dollars in overtime.
The worst part? We'd done the same kind of penny-wise thinking a few months earlier with a stepper motor.
“How fast can a stepper motor turn” was the wrong question
In fall 2024, I asked a supplier the exact question you just googled: how fast can a stepper motor turn? We were trying to replace a servo-driven indexing station with a stepper motor to save money, and I wanted to know if it could hit our required speed.
The answer I got was something like “up to 1,000 RPM depending on the model.” Technically true. Completely misleading.
What nobody tells you until you stare at a torque-speed curve is that stepper torque falls off hard as speed climbs. A motor that holds 400 oz-in at standstill might give you 80 oz-in at 600 RPM. Our application needed a specific torque at a specific speed, not just a speed the motor could theoretically reach spinning nothing.
We ran the prototype and it stalled. Not subtly. It stopped mid-index, the controller faulted, and a half-built carton got shoved down the line. Honestly, I'm not 100% sure why the initial torque calculation looked as good as it did—my best guess is we used the holding torque spec without derating it for speed.
When someone asks how fast a stepper can turn, the follow-up question should always be: how much torque do you need at that speed? If a supplier gives you a max RPM without a torque-speed curve, that's a red flag, not an answer.
The stepper project ultimately cost more than the servo repair would have. We now have a stepper on our bench doing low-speed positioning where it shines, and the servo is back where it belongs.
What I do now, and when I'll still take a shortcut
After the capacitor incident, I started a loose pre-order checklist. After the actuator controller incident, I added to it. After the stepper stall, I added to it again. The current version is short enough that I actually use it:
- Photo of the old part's nameplate or label, sent to the supplier.
- One written line: “this replaces X, installed in Y, controlled by Z.”
- Cross-reference verified against the original spec sheet, not just the seller's listing.
- If a controller is involved, it comes from the same manufacturer as the actuator unless I can prove otherwise.
- For steppers: torque required at operating speed, not holding torque.
I don't want to overstate this. If you're replacing a component with the exact same model number, in the same application, this process takes five minutes. And if it's a $60 fan motor on a non-critical line, honestly, roll the dice. Sometimes cheap and fast is the right total-cost answer.
But the moment you're cross-referencing a motor, a capacitor, an actuator, or a controller—especially across brands—slow down. The listing that says “fits” doesn't know what your motor actually does. The only person who can connect the part to the system is you. I learned that the expensive way so you don't have to.
Discuss this motor note