What Motor Do You Actually Need? Lessons From Bafang Hub Motors, Servo Projects, and VFD Mistakes

Eighteen minutes. That's how long the first "bafang 3000w hub motor" lasted in a customer's cargo-trike build before it stalled hard enough to let the magic smoke out. He returned it, furious. I pushed the replacement through. A month later that replacement came back too, this time with the smell of overheated varnish inside the hub. Same route. Same rider. Same motor class. Same result.

I handle application support and order review for a motor and motion-components distributor. Six years in, I've documented 14 significant mistakes from my own recommendations—mistakes that added up to roughly $16,000 in returned goods, replacement shipping, and lost engineering time. This article is about the pattern behind those mistakes, because it was never really about bad motors.

The surface problem: "it says 3000W, so why did it die?"

The customer in that first story wasn't careless. He'd searched for the biggest hub motor he could afford, found a listing described as a bafang 3000w hub motor, and assumed the wattage number was the answer to his real question: will this pull a heavy load up a long hill?

He was right about one thing: the motor could handle a lot of power. What he didn't check was for how long and under what conditions. He was also comparing the wrong set of specifications, because that's what the listings encourage you to do.

Here's the uncomfortable truth about hub motors in that power class: a direct-drive hub motor is efficient at cruising speed and at moderate acceleration. But hold it near stall speed—lugging a heavy vehicle up a steep grade, for example—and the motor turns a large share of that electrical power into heat instead of motion. The heat has nowhere to go quickly. Eventually the winding insulation gives up. That wasn't a counterfeit motor or a defective one. It was an application mismatch.

I have mixed feelings about wattage labels, by the way. They're useful for comparing the size of a system. But a max wattage number tells you almost nothing about low-speed torque, sustained duty, or whether the motor is matched to the controller that feeds it.

Deeper: watts are always attached to conditions

The mistake that cost me the most was not asking one question before recommending a motor: What is the actual duty cycle? A duty cycle is just a careful way of saying how often the motor can work hard and how long it needs to cool down. Most product pages won't write it in plain English. They'll list peak power, rated power, and maybe a temperature limit. If you don't read the fine print, you'll buy a motor that's great for 15 minutes and useless for 15 hours.

This applies even to newer hub-drive models like the bafang g900 hub motor that customers often ask about for small electric utility vehicles. The spec page tells you the motor's voltage range and rated output. It does not tell you whether that motor is appropriate for your vehicle weight, your grade, your controller settings, or your riding pattern.

In 2023, a customer asked me whether the bafang g900 hub motor would "fix" a conversion that had already burned one hub motor. My honest answer after the postmortem was: it depends on what burned the first one. A different motor isn't a different plan. If the load and duty cycle stay the same, a bigger motor just fails later—and sometimes fails more expensively.

"Arduino servo motor" is three different things hiding under one name

The second expensive lesson came from a prototype build. A customer wanted a small robotic arm for a classroom product. They asked for an "arduino servo motor," which is a completely reasonable thing to search for. But that search returns at least three different categories:

• Hobby positional servos, which rotate to a commanded angle and stop.
• Continuous-rotation servos, which spin forward or backward at a commanded speed.
• Industrial servo motors, which are useless without a matching servo drive and encoder.

They sound similar. They are not interchangeable. The hobby servo has a controller and gearbox built into a plastic case. It's meant for light loads and low precision. The industrial servo motor is a component of a much larger closed-loop system. You cannot plug one into an Arduino and expect it to behave like the little plastic ones.

According to Arduino's Servo library documentation, the library controls a standard servo by sending a pulse roughly every 20 milliseconds, and the pulse width sets the position. That works great for small hobby servos. It does not make an industrial servo motor suddenly function. That difference seems obvious after you've been burned by it, but I've watched otherwise smart engineers learn it the expensive way.

To be fair, the terminology is genuinely confusing. The word "servo" describes a control concept, not a physical size or a connector type. If you ask for a servo without specifying the control signal, the feedback device, and the drive, you haven't fully specified what you need.

What motors are compatible with a VFD? Not the ones you might think

This question arrives in my inbox more than any other: "what motors are compatible with vfd?" The short answer is that a standard general-purpose variable frequency drive is designed to run three-phase AC induction motors. That's its home turf.

People get into trouble when they assume the VFD is a universal motor driver. It isn't.

• Single-phase AC induction motors don't belong on a standard three-phase VFD output.
• Brushed DC motors and universal motors don't belong on a VFD output either.
• Brushless DC motors—including most e-bike hub motors—are permanent-magnet motors designed for a dedicated BLDC/PMSM controller, not for a general-purpose VFD.

I'm not a power electronics engineer, so I won't pretend to walk you through reflected wave voltages or dv/dt filters. But from a compatibility standpoint, I can tell you this: just because a motor has three phase wires does not mean a VFD will run it correctly. A hub motor's three phases need to be commutated in sync with the rotor position. A general-purpose VFD expects a different kind of motor and different control parameters. If you have to ask whether a random motor is VFD-compatible, you probably don't have the documentation needed to set it up safely.

One more layer here: even for true three-phase induction motors, "compatible" and "good idea for the long term" are different questions. Running a standard induction motor across a wide speed range without inverter-duty insulation can shorten its life. NEMA MG-1, Part 31 is the section people reference for definite-purpose inverter-fed motors. If your application needs sustained low-speed torque, look for a motor that's rated for inverter duty, or add external cooling. The VFD is not a magic box that makes thermal physics disappear.

When the "motor" problem was actually a linear-guide problem

The failure that taught me the most didn't involve high wattage at all. It involved a small servo-driven gantry. The customer's motor kept tripping on overcurrent. We swapped the motor. It tripped again. We swapped the drive. It tripped again. Someone finally checked the linear guide systems and found that the carriage was binding at one end of travel because the rails weren't parallel.

The motor wasn't weak. It was being asked to push against a mechanical bind every single cycle. The servo drive was doing its job by refusing to burn itself out. The real problem was in the mechanical system.

Linear guide systems fail differently than motors, but they show up in the motor's behavior. A misaligned rail adds friction, which increases current, which heats the motor, which eventually trips the drive or damages the winding. I still remember the embarrassment of authorizing a second replacement before asking the question I should've asked first: does the axis move smoothly when you turn the motor by hand?

This is one of those cases where I get why people blame the motor. The motor is what makes the noise. The motor is what smells hot. The motor is what stops working. But the motor is usually the last component in a chain of decisions about mechanics, control, and duty. It just happens to be the part that takes the blame.

What 14 documented mistakes actually cost

Let's make this concrete. Across six years, the big costs were not the motors themselves. They were:

• Return freight and restocking fees on products that were never defective.
• Replacement drives and controllers damaged by running with mismatched motors.
• Engineering time spent debugging what turned out to be a mechanical or application issue.
• Credibility. That one is harder to put a dollar figure on.

One order in particular stands out: a $3,200 mistake on a small production line where I recommended a motor based on rated power alone, ignoring the fact that the application required precise positioning at low speed. The motor had plenty of power. It had almost no holding torque at the required resolution. That error cost a full week of delays and a very uncomfortable phone call.

The good news: once I started writing down the failures, the fix got simpler. I stopped asking "what motor do you want?" and started asking "what does the machine actually do all day?" The motor is the last decision, not the first.

The pre-buy checklist I use now

If you're choosing any motor—a bafang hub motor, an arduino-compatible servo, an industrial AC motor, or anything else—run through this list first:

1. Define the motion, not the part. What torque do you need at what speed? Is that torque continuous or just for a few seconds?
2. Name the control system. Do you have a servo drive, a VFD, a BLDC controller, or an Arduino? Each motor type expects a specific controller. If you haven't chosen the control system, you haven't chosen the motor.
3. Find the continuous rating. Look for rated current, rated torque, ambient temperature, and duty cycle. Ignore the peak wattage until you've understood the continuous number.
4. Check the cooling reality. Does the motor rely on shaft speed for cooling? Will your application spend time at low speed under load? If so, plan accordingly.
5. Question the mechanical system. Move the load by hand. Check alignment of rails, couplers, and lead screws. If anything binds, no motor will fix it.
6. Ask a human. A good application engineer will ask you questions you didn't expect. That's a good sign. The wrong answer is the one that just says "yes, it'll work."

There is something genuinely satisfying about the first build you spec this way: no smoke, no callback, no overnight freight charge for a replacement that shouldn't have been needed. The motor finally does what it was designed to do, because someone actually told it what that was.

One last note: specs change, and listings can be sloppy. Verify current ratings and compatibility on the manufacturer's official page before you order—if you're looking at Bafang products, that's bafang-motor.com. The time you spend reading the fine print is cheaper than the time you'll spend explaining the smoke.