The Breakdown That Started It All

The line stopped at 2:47 PM on a Tuesday. Our conveyor had been running three shifts a day for seven years, and this was the first unscheduled stop that wasn't caused by a sensor or a jam. The maintenance lead put a stethoscope to the gearbox and gave me the look. “Worm gear's gone, or the bearing. Either way, we need a replacement.”

I’ve been tracking our parts spend since 2018—around $180,000 cumulative, maybe $190,000, I’d have to check the system. So I know the drill: pull the asset record, find the old quote, get two or three vendor offers, build a quick TCO comparison. For a replacement like this, the total cost of ownership is more than the invoice price. It’s the downtime, the installation labor, the maintenance calls, and the odds of doing it again too soon.

The asset record said the old unit was a Winsmith speed reducer 917—20:1 ratio, 1 HP input, mounted to a standard NEMA AC motor. The original PO from 2019 was $1,850. Replacement quote for the same model came back at $2,410. Then the distributor said they could do a Winsmith speed reducer 920 for $2,230. That’s $180 less for the “same thing,” supposedly.

The $180 That Almost Cost Us a Lot More

It’s tempting to think you can just compare brand, ratio, and horsepower. The 920 had the same ratio and the same motor frame. It even fit the same bolt pattern. So why would I pay $180 extra for the 917?

I almost didn’t. But our procurement policy now requires a spec check against the motor’s service factor before approving any replacement gearbox. That policy came from a painful mistake in 2023, when we ordered a rush reducer without checking the input speed. We didn’t have a formal approval chain for rush orders back then, and the 'approved' unit ran hot for six months before failing. I built the checklist after that. This time, the checklist did its job.

I asked the distributor to send the thermal ratings for both models. They did, and the difference was surprising. The 917 was rated for 1.4 HP of continuous thermal capacity in our ratio. The 920 was rated at 1.05 HP. We have a 1 HP AC motor with a 1.15 service factor. Different story.

Per NEMA MG 1, a motor's service factor is the overload capability it can handle continuously. For most general-purpose AC motors, it's 1.15. That means the reducer has to survive 1.15 HP, not just 1 HP. The 920 was cutting that margin dangerously close.

Why does this matter? Because heat is what kills worm gear reducers. A worm reducer generates heat at the contact point. If the housing can't shed that heat, the oil cooks, seals shrink, and the unit fails—often at the worst possible moment. The 920 might have been fine for a light-duty application running two hours a day. For our 24/7 line, it was a gamble.

So we ordered the 917. It cost us $180 more. The alternative would have been explaining to my manager why the same line went down again in six months to save $180. I dodged a bullet. Period.

But Wait—What About Servos and Steppers?

While we were waiting for the new reducer, one of the automation techs asked, “If we’re replacing the gearbox anyway, why not use a servo or a stepper and get rid of the machinery?” It’s a fair question, and it’s one I’ve heard more and more in 2025.

First, what’s a stepper motor? In simple terms, it’s a brushless motor that moves in discrete steps. Feed it digital pulses, and it rotates by a fixed angle per pulse, no feedback needed for basic positioning. That’s why steppers live in 3D printers and CNC routers. But for continuous rotation at 60 RPM under a heavy industrial load, a stepper is the wrong tool. It runs hot, draws current even when holding, and can lose steps if the load spikes. AC motors don’t care about a spiky load—they just keep turning.

I also pulled up the SG90 servo motor datasheet to make the comparison tangible. The SG90 is the classic hobby servo found in Arduino kits. Its datasheet lists stall torque at about 2.4 kg-cm, roughly 33 oz-in. Our conveyor needed about 120 lb-in at the output shaft. Do the math: you’d need nearly 60 SG90 servos to match one 917. It’s a different universe. (Not that I didn’t appreciate the tech’s creativity—surprise, surprise, the guy likes building robots on weekends.)

This isn’t a knock on servos or steppers. They excel at positioning, registration, and any task that needs acceleration control. For a simple conveyor running one direction at constant speed, a gear reducer paired with an AC motor is the honest, low-maintenance, cost-effective choice.

What I’d Do Differently (And What I’d Keep)

If I remember correctly, the 917’s catalog page has a foot so small that you need a magnifying glass to read the thermal ratings. That’s not a criticism of Winsmith—it’s a reminder that the detail that saves you money is often in the fine print. The 917 was the right call for this application. But that doesn’t mean the 920 is bad.

Here’s the honest limitation: if your duty cycle is under 50%, or your load is well below the reducer’s rated thermal capacity, the 920 will save you $180 and probably run for years. For intermittent operation—say, a washdown station that cycles every 10 minutes—the 920 is a legitimate option. For continuous use, I’d pick the 917 every time. And if you’re dealing with a servo or stepper question, remember the SG90 datasheet comparison. Different jobs, different tools.

Looking back, I wish I had formalized the spec-check process sooner. The old 917 lasted nine years. The previous engineer had specified it correctly. My only near-miss was letting the price gap hijack my judgment. A $180 difference is nothing compared to a three-day plant shutdown. I built the TCO spreadsheet, but the real lesson is simpler: verify the numbers that matter, then let the price fall where it may.

The better buy isn’t always the lower quote. It’s the one you can forget about for the next ten years.