I Almost Spec'd the Wrong Harmonic Drive for a Telescope Mount

The Backstory

Last spring, our lead engineer walked into my office with a bill of materials in one hand and a half-empty coffee in the other. "We need a harmonic-drive telescope mount," he said. "The client wants the drive integrated into the DEC axis, with a stepper motor already built in."

I'm the procurement manager at a 15-person motion control company. I've managed a component budget of about $480,000 per year for six years, negotiated with 40+ vendors, and documented every order in our cost tracking system. This one was new to me. Most of our work is industrial robotic arms. A telescope mount sounded simple—until I started pricing harmonic drives.

My Initial Misjudgment

When I first started buying precision gearboxes, I assumed the lowest quote was always the best choice. This project nearly proved that assumption right—and then proved it very wrong.

We asked for quotes from eight harmonic drive manufacturers. The published spec sheets all mentioned near-zero backlash and reduction ratios from 30:1 to 160:1, but the practical details varied widely. The lowest quote came in 28% below the median. My first instinct was to accept it. Honestly, the number looked great. But when I compared the technical specifications line by line, the deal fell apart.

The low quote was for a standard 50:1 harmonic drive. We needed 50:1 for the right ascension axis, but the DEC axis required 100:1. The vendor said they could quote that separately. The separate quote was not low. We also needed a hollow shaft version to route cabling. The low quote vendor didn't have a hollow shaft model available within our lead time.

A lot of buyers focus on the unit price and completely miss lead times, configuration options, and support. The question I hear most often is "what's your best price?" The question they should ask is "what's included in that price?"

What's a Stepper Motor? And Why Not an Induction Motor?

Then the client's engineer asked a question that reset the whole project: "What's a stepper motor, and why not just use a standard induction motor?"

Fair question. An induction motor is cheap, durable, and widely available. But a telescope mount is not a fan. It needs to track at sidereal rates, hold position with high resolution, and do it without constantly burning current in a brake.

An induction motor is not designed for precise indexing. You could add an external encoder and a brake, plus a high-ratio gearbox, and it might work. But that adds cost, wiring, and failure points. A stepper motor, by comparison, moves in discrete steps. With micro-stepping and a 50:1 harmonic drive, you can get arc-second-level resolution. In many designs, no brake is needed for a balanced mount.

(Yes, a servo with an absolute encoder would also work. For this small-batch order, it was overkill—and the lead times were worse.)

So that simple question, "what's a stepper motor," was more useful than "which drive is cheapest." It forced us to think about the whole motion system instead of a single component.

The Linear Bearing Rail Plot Twist

While the harmonic drive quotes were coming in, I also had to source a linear bearing rail for the mount's altitude adjustment. That part should have been boring. Two vendors, one week, done.

But the rail vendor asked what we were building. When I said "telescope mount," he paused. "Ah. So you need the preloaded carriage and the low-vibration coating, right?"

I had quoted the standard rail. $210. The preloaded version was $340. It added four days to the schedule. But it saved us from a field failure. That was a small lesson inside a bigger one: the spec details matter more than the sticker price.

Small Order, Real Support

The harmonic drive order was for seven units. Seven. Not a huge number, but not a sample size, either. Two of the eight manufacturers didn't respond to follow-up emails. One had a 50-unit minimum order quantity. Another said "we'll send you a quote" and then went silent. Not ideal.

The vendor we eventually chose was a harmonic drive manufacturer with a slightly higher per-unit price. But they offered no MOQ, a configurator that matched our actual mounting dimensions, a 3D model that was accurate, and a human being who answered the phone. That last one mattered more than you'd think.

The price difference from the lowest quote was $1,350 on the full order. But the low quote didn't include the hollow shaft option, the motor adapter, or the lead time. When we compared on total cost of ownership, the "expensive" vendor saved us roughly $4,200—mostly by preventing a redesign when the motor wouldn't fit.

Small doesn't mean unimportant. It means potential. When I was starting out, the suppliers who treated my small orders seriously are still the ones I trust with big orders. Same logic here.

Result and Reflection

The mount shipped in January 2025. The customer reported clean tracking and no resonance issues. I'm not an engineer—my job was to keep the project under budget and avoid surprises.

Our TCO spreadsheet now includes an integration cost line. That wasn't there before this project.

What did I learn? A few things.

• Unit price is not cost. Integration support and lead times are part of the real price.
• The motor choice should come before the gearbox choice. Ask "what's a stepper motor" early.
• Harmonic drive manufacturers are not fungible. A quote for one ratio says almost nothing about the rest of their catalog.
• Small orders are not a favor. They are one side of a long relationship.

Everything I'd read said harmonic drives were expensive and sensitive. In practice, for this application, they were the most cost-effective path—because the alternative required more parts, more labor, and more risk. That's the kind of lesson that doesn't show up on a quote. It only shows up after you've signed one.