2026-11-17 · Ingrid Nystrom
We Blamed the Cheap Motors for 8 Months. Then We Read the VFD Nameplate.
A procurement manager walks through how a $630 motor substitution turned into $1,850 in downtime, why the failures weren't really the motors' fault, and what our team now does differently when specifying Leeson 1.5 HP motors, reading Leeson model numbers, sizing a VFD for a 5 HP conveyor, and comparing electric actuators to pneumatics.
6:40 a.m., Line 3, and a smell
February 13, 2024. I was walking in from the parking lot with a coffee I never got to drink. Line 3 was down, and the maintenance lead was already standing at the conveyor with the guard off, holding a motor that smelled like burnt varnish.
It was a 1.5 HP TEFC unit on the incline section. It had been in service for 19 weeks. The label on it was not a brand I had approved, and I knew exactly how it got there, because I was the one who signed the purchase order.
That morning cost us about 4.5 hours of production. What followed cost us considerably more, but the expensive part wasn't the motor. It was the eight months I spent believing the wrong explanation.
How I got there: the Q3 2023 audit
I manage the MRO budget for a 240-person contract packaging operation — roughly $1.2M a year, and I've been tracking it line by line for six years. In Q3 2023, our plant manager asked for a 7% cut in maintenance spend without cutting headcount. So I did what procurement people do: I pulled every invoice from 2022 and 2023 and started ranking line items by frequency times unit price.
Electric motors landed near the top. Not because they were expensive individually — because we were buying them sloppily. Out of 23 motors on the floor, I found nine different catalog numbers doing what should have been maybe four distinct jobs. Two of the incline conveyors had different frames for identical duty. One spare sitting in the crib had a C-face flange we had no application for.
Most of the installed base was Leeson, which told me the original OEM had specified them and then nobody had ever revisited it. My honest first reaction was that this was exactly the kind of quiet default that costs a mid-size plant money — one brand, one distributor, no competitive tension, no reason for anyone to sharpen a pencil.
So I did the exercise: I pulled quotes on the three highest-turnover sizes. For the 1.5 HP units, the Leeson equivalent came in around $415 per unit through our distributor. A comparable import was quoted at $310. Six motors a year, roughly, at the time. That's about $630 annual difference on paper.
I approved the substitution for the non-critical conveyors. I want to be specific about that decision, because it wasn't stupid. It was made on the information I had, which was unit price and a spec sheet that matched on the numbers I knew to check: 1.5 HP, 1,725 RPM, 56C frame, TEFC, 1.15 service factor, 230/460V.
The part where it fell apart
Between October 2023 and February 2024, we installed six of them. Three failed inside five months. Not catastrophically — nuisance trips first, then winding failures on two of them, one bearing growl on the third. Labor and downtime on those three events came to about $1,850, which is roughly triple what I thought I was saving.
Here's where I want to be careful, because the easy article would end right here: cheap motors bad, buy the brand name, the end. That's not what happened, and if I'd stopped there I would have kept having the same failure every eight months with more expensive parts.
The twist: we were sizing drives by horsepower
In March 2024 I finally sat down with our drives integrator for two hours instead of trading emails. I brought the failed motors and the trip logs. He asked one question I didn't have a good answer for: what's the drive's output current rating at our input voltage and phase?
I said "it's a 5 HP drive." He said, "that's not what I asked."
That was the whole ballgame. Those conveyors run on 240V single-phase input feeding 230V three-phase motors. The drive nameplate said 5 HP — but that rating was for three-phase input. On single-phase input, the same unit was rated around 3 HP. We had been running 5 HP motors, with a full-load current of 15.2A at 230V (I know that number cold now — it's NEC Table 430.250), off drives that were thermally undersized for the duty. Add a 140-foot motor lead run with no load reactor, and you have a recipe for nuisance overcurrent trips and winding stress that has nothing to do with where the motor was made.
The assumption was that the motors were failing because they were cheap. The reality was that the drives were undersized and the lead length was long, and the cheaper motors simply had less thermal margin to absorb it. The causation ran the other way from what I'd assumed.
Could the Leeson units have survived it? Probably a couple of them would have lasted longer. I'm not 100% sure they'd have lasted either — we never ran that experiment and I'm not going to pretend we did. What I am sure of is that if I had spent the extra $630 and changed nothing else, I would have concluded the whole thing was a fluke, learned nothing, and been back in the same meeting in 2025.
What we actually changed
Four things, in the order we did them.
One: we started sizing VFDs by current, not by nameplate horsepower. The rule is simple and I now have it taped inside the panel shop cabinet. Take the motor's full-load amps at your actual voltage — 15.2A for 5 HP at 230V three-phase, 7.6A at 460V — then look up the drive's heavy-duty output current at your input voltage and input phase. If the drive doesn't publish a single-phase rating at your horsepower, that's not a 5 HP application, that's a question for the vendor. In our case, the honest answer was a drive frame one size up, plus a load reactor on the long runs. Most of these headaches are preventable if you spec the current instead of the number on the box.
Two: we built a cross-reference sheet instead of trusting model numbers from memory. This is the part that has saved me the most time. A Leeson catalog number encodes frame, enclosure, mounting, and a handful of feature suffixes, and two numbers that look nearly identical can differ in something that makes them non-interchangeable — a C-face versus a rigid base is the classic one, and I have personally ordered the wrong one. I don't trust myself to decode the numbering scheme off the top of my head and I'd advise you not to either. What we did instead: one spreadsheet, one row per installed motor, with columns for catalog number, frame, enclosure, mounting, voltage, phase, RPM, service factor, and — critically — a column for "what this motor actually does." When we need a replacement, we match on the function, not the string. Leeson is part of Regal Rexnord now, so a chunk of the older catalog references route through their literature site; our distributor keeps a current cross-reference and I make them send it in writing every time we reorder.
Three: we standardized the 1.5 HP workhorse. Sixteen of our 23 motors are 1.5 HP. We now hold two variants in the crib and nothing else: three-phase TEFC 56C, and a single-phase version for the two conveyors that are on straight across-the-line starting. Everything else is a special order with a two-week lead time and that's fine, because we now know which two are the emergencies.
Four: we stopped comparing servo control and stepper control as if they were the same purchase. I made this mistake too. When the labeler's servo axis failed in 2023, I asked whether we could save $900 by going to a closed-loop stepper. On paper the holding torque looked fine. In practice the torque fell off exactly where we needed it — at speed, under a reversing load — and the operator spent two weeks chasing phantom position faults. The lesson wasn't "servos are better," it was that we had sized by holding torque instead of by the torque-speed curve and the load-to-motor inertia ratio. If you're going to compare them, compare them at your actual index speed, with your actual load inertia, and ask what happens on an e-stop. If you ask me, that comparison is worth an hour of an integrator's time and almost nobody does it.
The actuator line item nobody budgeted for
While I was in there, I re-ran our rod-style industrial electric actuators against the pneumatic cylinders they replaced in 2021. The original business case was clean: no compressed air prep, programmable positions, quieter. All true. What the original case missed was the rebuild interval on the seals and the rod bellows — roughly annual on the high-cycle stations in our washdown area. Two of the five units have since needed it. That's not an argument against electric actuators; the positioning repeatability has genuinely saved us scrap on the labeling stations, and I'd do it again. It's an argument for putting the rebuild interval on the TCO spreadsheet before you sign, because the maintenance line is where the savings quietly evaporate.
What I got wrong about the whole exercise
The 7% cut happened. We got there — mostly through consolidating nine motor part numbers down to six and through not paying rush freight on things we should have had in the crib. What I got wrong was the mental model. I went into that audit thinking of motors as commodities, where the only real variable is price. They're not commodities. They're components in a system, and the system includes a drive, a cable run, a load profile, and a maintenance tech who has 20 minutes to figure out why a conveyor stopped.
Where I've landed: the lowest quote is rarely the most expensive decision, but it's rarely the cheapest either. The number that matters is the total — the unit, the commissioning time, the failure labor, the spare you have to stock because the lead time is four weeks, and the two hours of an engineer's attention the whole thing quietly consumes.
One boundary on all of this. This worked for us because we're a mid-size plant with fairly predictable duty cycles and a maintenance team that logs failures. If you're running seasonal demand with three shifts and no historian data, or if you're dealing with a much larger installed base where standardization has its own changeover cost, the calculus is going to look different. I can only tell you what the invoice and the trip log told me.
And I still keep the failed motor. It's on a shelf in my office with a sticky note on it. The note doesn't say the brand. It says "15.2A."