Bonfiglioli Planetary Gearbox & Gearmotor FAQ: Answers from a Quality Engineer

Every week, I review technical drawings, test reports, and returned units. I'm a quality engineer at Bonfiglioli, which basically means I'm the person who says “send it back” before something ships. I inspect roughly 200 unique product configurations a year, and in 2024 I rejected 11% of incoming supplier samples, mostly for backlash and sealing deviations. These are the questions I actually answer in emails, site visits, and the occasional LinkedIn thread.

What makes a Bonfiglioli planetary gearbox different from a cheaper planetary gearhead?

Not all planetary gearboxes are built to the same standard. Actually, the biggest differences are the ones you can't see from the outside: tooth flank geometry, case hardening depth, bearing spacing, and sealing concept.

Here's something I learned the hard way. Everything I'd read said backlash is the spec to obsess over. In practice, the first thing that fails on a planetary gearbox in the field is usually the seal, not the gears. A bit of contamination ingress starts a wear chain that leads to noise, lost motion, and eventually a locked unit. When I do an incoming inspection, I spend as much time on the seal and the oil condition as on the gear measurements.

Bonfiglioli 300 Series planetary gearboxes use case-carburized gears that hold within 3 HRC hardness uniformity across a batch. The cheaper units we've benchmarked often vary by 7 HRC or more. Per standard gear rating methods (ISO 6336 and AGMA), that difference directly affects allowable contact stress before pitting starts.

Why buy a Bonfiglioli gearmotor instead of matching a motor and gearbox myself?

Motors and gearboxes are each designed to be connected to something. But that connection is exactly where field-assembled units go wrong.

Bolt a motor onto a gearbox and you have alignment, coupling, and tightening torque to manage. Get any of those wrong and you get vibration, fretting corrosion on the shaft, leaking seals, and premature bearing failure. In 2023, a customer returned a “failed” gearmotor that was actually a mounting problem: the motor adapter bolts were torqued unevenly, and the input shaft had bent by 0.08 mm. That wasn't a manufacturing defect—it was an assembly mistake.

Bonfiglioli gearmotors come out of the factory already aligned and tested as a system. We run each unit and measure noise, vibration, and temperature rise according to IEC 60034-1. That's the prevention part. Field repair always costs more than a factory check that takes five minutes.

Does Bonfiglioli make a NEMA 23 stepper motor?

No. I'll be straight with you: Bonfiglioli builds industrial gearmotors, planetary gearboxes, worm and bevel drives, and servo gearboxes—not stepper motors. So if you searched “NEMA 23 stepper motor,” you came to the right place for gears but the wrong search.

NEMA 23 is a frame size: the face is about 2.3 inches square. Those motors run open-loop in 3D printers, desktop CNC routers, and lab automation, and they're genuinely good for light, intermittent loads. But if your project has a continuous duty cycle, shock loads, or a dusty factory floor, a stepper is usually the wrong layer of the stack. Stepper torque falls off as speed rises, and if the load exceeds it, the motor skips steps and your machine loses position.

Use a NEMA 23 if it honestly fits your application. Just match the motor class to the application, not to whatever came up first in search. If you're building a one-off prototype, a stepper setup is probably fine. If you're moving an 800 kg conveyor load three shifts a day, you need a gearmotor with a real thermal rating.

What about the MG90S servo motor? Is that the kind of servo Bonfiglioli uses?

The MG90S is a 9-gram micro servo. The one with tiny plastic gears and a stall torque around 1.8 kg·cm. It's made for RC cars, drone flaps, and school robotics arms. I keep one on my desk to show visitors how broad the word “servo” is.

An industrial servo motor is a completely different machine: a permanent-magnet synchronous motor with a high-resolution encoder and a matching servo drive. It holds position under heavy, reversing loads for years without a thermal trip. An MG90S will not survive one production shift in that job.

In May 2024, a startup contacted us after ordering a dozen MG90S units for a small indexing table. The first one stuck after three hours of testing. They'd assumed all servo motors were interchangeable. That event changed how I answer this question: don't size industrial machinery from a hobby catalog. If you don't know your duty cycle or the reflected load moment, run it by an engineer before you buy, not after you burn out the third one.

What happens when a linear actuator fails?

There's no single sequence, but here's the one that keeps showing up in our failure reports:

The drive cycles the actuator toward the end stop. The limit switch, if it's working, tells the motor to stop. If that switch was damaged, out of adjustment, or bypassed during setup, the actuator keeps going. The carriage meets the mechanical stop at full motor torque. If the drive's torque limit isn't set, the screw threads strip or the motor thermal trips. After that, the load is unsupported, the rod may bend, and the unit jams in a position your PLC didn't expect.

The consequences depend on the application. In a conveyor gate, product backs up and gets damaged. In a vertical lift, the load drops or hangs uncontrolled. In a wall-mounted actuator, the bracket cracks. In September 2024, I saw a packaging line go down for eight hours from exactly this chain of events; the replacement actuator and downtime cost roughly €14,000.

Prevention beats repair: set the torque limit, test the limit switches on a schedule, and size the actuator with a service factor that matches the actual load. A bigger actuator isn't always the answer, but a safety margin proportional to the risk is.

What quality problems do you reject most often?

On castings, porosity at sealing surfaces. It's invisible until the housing leaks under temperature cycling. In Q1 2024, I rejected about 7% of our incoming castings for that.

On gearing, it's backlash. A supplier quotes 10 arc-min, then delivers a planetary gearbox that measures 18 to 22 arc-min. In a precision indexing application, that extra lost motion makes scrap quietly for hours before anyone catches it. In 2023, we sent back a sample batch of fifty gear sets because case hardness varied 7 HRC across a batch; our spec is 3 HRC uniformity. The vendor said it was within industry tolerance. It wasn't our tolerance. They redid every one at their cost.

Those findings are why I built a 12-point incoming checklist in 2022. Five minutes of verification beats five days of correction. The checklist has saved us an estimated $28,000 in potential rework since then.

What's the one question you wish more buyers asked?

The one I wish buyers asked is not “how much torque?” but “what's the service factor for my actual duty cycle?”

Catalog torque feels precise, but it doesn't tell you whether the unit can handle that torque when it starts and stops thirty times per hour, in a dusty environment, with occasional overloads. Gear rating standards (ISO 6336 and AGMA) and motor duty standards (IEC 60034-1) only help if you feed them the correct load profile.

Most specifications I review list a nominal torque and a speed. The load profile, ambient temperature, mounting direction, and available cooling are missing. Those four factors cause more gearmotor replacements than the gear rating ever does.

I can only speak to industrial applications. If you're building a prototype or a short-run machine, the tradeoffs are different and your mileage may vary. But for production equipment, send the load profile to an application engineer instead of just the torque number you hope will work. It's a five-minute conversation that can save you a two-month redesign after the first failure.