Which Motor Lasts Longest? That's the Wrong Question to Ask

I keep a spreadsheet of my own failures. Sixteen entries, $47,300 total. I open it every quarter, mostly to remind myself that spec sheets lie — not on purpose, but they do.

I've been an applications engineer for drive systems since 2016. That means I'm the person who tells you which motor to put in your cargo bike or your packaging line. And for the early part of my career, I got it wrong often enough to build a very informative spreadsheet.

Most of the entries started with a perfectly reasonable question from a customer.

• “How long does a Bosch ebike motor actually last?”
• “Bosch vs Shimano — which drive unit is more durable?”
• “I need a high torque servo motor. What do I order?”

These sound like different problems. They're not. They're all versions of the same mistake: treating motor lifespan as a fixed property of the brand or the model. As if durability were a spec you look up, the way you look up weight or IP rating.

It's not. The motor itself is rarely what fails or survives. The system around it decides that. I've got 16 entries proving it.

The Number That Gets Everyone in Trouble

Every premature failure in my spreadsheet traces back to a number that looked fine but was read wrong. The most dangerous one is peak torque.

Peak torque is what a motor can do for a few seconds. It's not what a motor can do.

In March 2022, I approved a mid-drive for a small delivery company's cargo bike prototype. The route had a 12% grade, and the bike carried two people plus a loaded box. The torque curve looked fine. I told the customer: “This motor handles your hill.”

What I meant was: it handles the hill if the climb stays under about 40 seconds and the ambient temperature stays under 25°C. The customer heard: this motor handles the hill.

On a warm afternoon, the unit thermal-faulted near the top of a two-minute climb. The motor was fine on paper and cooked in reality, because a peak curve is a sprint. A route is a marathon.

This is the same trap behind most “high torque servo motor” requests. The torque usually isn't the problem — the continuous torque at your actual working speed is. If a datasheet doesn't clearly state a duty rating, that's a red flag. IEC 60034-1 defines duty types S1 through S10; the “S” tells you how long the motor can sustain its rated output. If you're only looking at “peak” without a duration, you're not reading a spec. You're reading a marketing claim.

Why “Bosch vs Shimano” Is the Wrong Question

Brand quality matters. It genuinely does. A well-engineered drive unit with good bearings, proper sealing, and a sensible controller strategy will outlast a cheap unit in the same application. But that phrase — “same application” — is doing a lot of work.

Here's an example from the e-bike side. In the EU, EN 15194 limits e-bike assist to 250 W nominal power and 25 km/h. A Bosch Performance Line CX, which I spec regularly, is rated at 250 W continuous with 85 Nm peak torque. That gap between continuous and peak isn't a design flaw. It's the entire story.

Ride that same unit on a flat commuter route with low assist, and it can run for years without issues. Hammer it up a hilly cargo route at maximum assist all day, and the thermal load is completely different. Same motor. Different lifespan.

That's why I get slightly annoyed when someone asks for “the service life of a Bosch ebike motor.” I have a customer in Munich who emails me about motor Lebensdauer every spring. The honest answer is never a number. It's a set of questions about their route.

I am not saying the brand doesn't matter. I am saying it matters less than your duty cycle, and that's a hard lesson to learn by invoice.

So when a customer asks “Bosch vs Shimano, which is more durable?”, my answer is: what's the duty cycle? The same drive unit will have a different service life in a flat commuter bike than in a bakery's delivery cargo bike. Comparing brands on peak torque tells you almost nothing about which one will last longer in your application.

Not the answer people want. It's the one that's actually true.

The Gearbox Is Not an Accessory

I learned this one on the industrial side, and it cost me a customer's line.

In 2021, I compared two planetary gearboxes for a servo motor application. Gearbox A was 35% cheaper, same ratio, same backlash class, similar efficiency. Every cost analysis said: buy A. My gut said the torque margin at the actual operating speed was too thin. I ignored my gut, because the spreadsheet was so clear.

Four months into production, backlash had quadrupled and the line was scrapping parts. The motor was fine. The gearbox was the weak link, and it made the whole axis look bad. We swapped in the “overpriced” unit, and it's been running for three years without a problem.

That's the day I stopped treating “servo motor gearbox” as a bolt-on. The gearbox and the motor are a couple. Ratio, inertia, backlash, and thermal limits at the real duty cycle matter more than the motor's peak specs. Change one, and the relationship changes.

And while I'm at it: “what stepper motor should I use” is often the wrong question too. If your load varies and you need position to hold reliably, a closed-loop servo is usually the better answer. A stepper can miss steps under load — right up until the part on the machine becomes scrap. Choosing a stepper because it's cheaper feels like a no-brainer. Until the rejects eat the savings.

What Getting This Wrong Really Costs

Let me give you two numbers that still irritate me.

$9,400 and one very awkward apology email

In January 2023, a customer needed four servo axes re-specified during a plant shutdown window. We had 48 hours. Normally I'd spend a week collecting load profiles and cycle data. There was no time, so I made the calls based on peak torque and told myself it was fine.

Six weeks later, during the first full-production test, three of the four motors tripped on thermal overload. The rework — new motors, new gearboxes, labor, and an apology email I still cringe at — ran to about $9,400.

In hindsight, I should have pushed back on the timeline. But the shutdown window was real, and the company was waiting. I made the best call I could with incomplete information, and the call was wrong.

$700 and a relationship I nearly threw away

Last summer, a small fabricator called about replacing stepper motors on an aging machine. Their order would have been about $700. Because they were small, I gave them a quick email answer instead of asking the basics: cycle rate, load variation, ambient temperature.

Two weeks later, their controller boards were cooked and their line was down. The actual fix was a closed-loop system at four times the original budget. And the part that stings: I didn't run my own checklist. I was in a hurry, and the order felt small.

I used to rank inquiries by order size. I don't anymore. Here's what I tell new engineers in training: when I was starting out, the suppliers who took my $500 questions seriously are the ones I still send $20,000 orders to today. I've now lived that lesson from the other side of the table.

Small orders are just big orders that haven't happened yet. And small mistakes are just big invoices waiting to happen.

The Checklist That Fixed It

After the January 2023 failure, I sat down with nine years of failure notes and wrote a pre-order checklist. It's not elegant. But since July 2023 — 18 months as of January 2025 — it's caught 47 potential failures. Most of them before anyone spent a dollar.

We use it for everything, from a $400 evaluation unit to a $400,000 production line. Here it is:

1. Define the continuous load first. Give me the RMS torque over the real cycle, not the peak. If you can't measure it, add 25% headroom to your honest estimate.
2. Write down the duty cycle. Cycles per hour, loaded time, ambient temperature range, and the duration of the worst case. Thermal endurance is the lifespan story; duty cycle is its plot.
3. Check the gearbox at the operating point. Ratio, backlash, and torque rating at the actual RPM — not at the motor's max speed. If you change the motor, re-check the gearbox.
4. Compare field data, not spec sheets. For e-bike drives: warranty terms, exchange programs, service network. For industrial motors: failure rates, encoder robustness, and how the controller manages heat.

That's it. Four items. The list is short because the physics is not complicated. The marketing just talks us out of it.

So, Back to the Original Question

Bosch vs Shimano — which e-bike motor is more durable? I've stopped pretending there's a universal answer.

Here's how I evaluate it now. Bosch's Performance Line CX is a fully integrated system: motor, battery, controller, display, connectivity, and optional ABS. It's backed by a two-year warranty and a drive-unit exchange program. You can verify all of that on bosch-ebike.com. Shimano has its own strengths, especially in weight and rider feel. Which one is “more durable” for you depends on your route, your load, your climate, and which service network you can actually reach when something breaks.

I know that's not the neat answer people want. But it's the right one. Durability is an outcome of matching the system to the duty cycle. It's not a trophy you pick up at the checkout.

Bottom Line

The longest-lasting motor isn't the one with the biggest peak numbers. It's the one that matches your continuous load, your duty cycle, and your service network. I've got $47,300 in spreadsheets that say the same thing.

Ask the right questions before you order. It saved us a ton of money and spared me a few awkward apologies. I'd rather you learn it from this article than from your own failure log.