Gear Ratio Calculations: Mastering Torque-Speed Trad...

Why Gear Ratios Matter in Motor Control

Motors rarely drive loads directly. A small DC motor might spin at 6000 RPM but produce only 0.02 Nm of torque — fine for spinning a fan, useless for opening a heavy door. Gearing lets you trade speed for torque (or vice versa), matching the motor's characteristics to what your application actually needs.

The math isn't complicated, but getting it wrong means either a system that stalls under load or one that moves painfully slowly when it doesn't need to. And then there's efficiency — real gears have friction, and ignoring it leads to motors that overheat or can't deliver the torque you calculated on paper.

The Basic Relationships

The gear ratio is simply the number of teeth on the driven gear divided by the teeth on the driver gear:

A ratio greater than 1 means speed reduction and torque multiplication. A 48-tooth gear driven by a 12-tooth pinion gives you a 4:1 ratio — output speed drops to one quarter of input speed, but torque increases fourfold (minus losses).

Output speed follows directly:

Torque multiplication is where efficiency ($\\eta$) enters the picture:

That efficiency factor is doing real work. A typical spur gear stage runs 94-98% efficient. Planetary gears are similar. Worm gears? Often 50-90% depending on the lead angle. Stack multiple stages and those losses compound.

A Real Design Example

Let's work through an actual scenario. You're designing a motorized camera slider that needs to move a 2 kg payload at 50 mm/s. The drive pulley has a 20 mm diameter, so you need the output shaft spinning at:

Your motor is a small brushed DC unit rated at 3000 RPM no-load, with a peak efficiency around 2400 RPM and 0.015 Nm of torque at that speed. You need a gear ratio of:

That's a big reduction. You could do it with a two-stage spur gear train — say, a 10:1 first stage (10T to 100T) followed by a 5:1 second stage (15T to 75T). Each stage at 96% efficiency gives you:

Now for torque. Your motor provides 0.015 Nm at the input. After the gearbox:

Is that enough? The friction force to overcome is roughly μ × m × g. With decent linear rails ($\\mu \\approx 0.01$) and a 2 kg payload:

Torque required at the pulley:

You've got 0.69 Nm available — massive overkill for steady-state motion. But you also need to accelerate the payload. If you want to reach full speed in 0.5 seconds:

Total force during acceleration is 0.4 N, requiring 0.004 Nm at the pulley. Still well within your 0.69 Nm budget. The design works.

Let's check power to make sure nothing's thermally ridiculous. Input power:

Output power:

The gearbox dissipates about 0.3 W as heat — not a concern for this application.

Open the Gear Ratio Calculator to run through your own numbers.

Choosing the Right Ratio

Most engineers pick a gear ratio by dividing their motor's rated speed by their desired output speed. That works, but it ignores where your motor actually operates efficiently.

Brushed DC motors have peak efficiency at maybe 70-80% of no-load speed. Running them near stall for high torque is thermally brutal — you're dumping most of your electrical power as heat in the windings. Running them near no-load wastes the motor's torque capability.

The sweet spot: size your gear ratio so the motor runs in its efficient operating region during typical use, not just peak demand. If your application has variable load, bias toward the heavier end.

For stepper motors, the calculus is different. Steppers lose torque rapidly as speed increases due to inductance limiting current rise time. A higher gear ratio keeps the stepper in its high-torque, low-speed region where it's happiest.

Efficiency: The Number Everyone Gets Wrong

Datasheet efficiency figures are best-case numbers. A gear rated at 95% efficiency achieves that under optimal load, proper lubrication, and after break-in. During initial operation, with marginal lubrication, or at very light loads, real efficiency can drop to 85% or worse.

Worm gears are the worst offenders. A worm gear datasheet might claim 85% efficiency, but that's at the optimal load angle and with the worm driving the wheel. Try to backdrive it (wheel driving worm) and efficiency can drop below 50% — or the gear becomes completely self-locking. Sometimes that's a feature. Often it's a surprise.

For multi-stage gearboxes, multiply the efficiencies:

η_total = η1 × η2 × η3 × ...

Three stages at 94% each? You're down to 83% overall. Four stages drops you to 78%. That missing power becomes heat in your gearbox, which affects lubricant viscosity, which affects efficiency further. It's a feedback loop.

Planetary Gearboxes

Planetary gear sets are popular in robotics and precision applications because they're compact for their ratio and handle high torque well. But they're not magic. A cheap planetary gearbox from the usual overseas suppliers might claim 90% efficiency per stage, but I've measured units that barely hit 80% when new. The good ones (Harmonic Drive, Maxon, etc.) actually achieve their specs, but you'll pay for it.

Common Mistakes and Gotchas

I've seen these trip up experienced engineers, not just hobbyists.

Ignoring reflected inertia. When you gear down, you don't just multiply torque — you also multiply the load's apparent inertia as seen by the motor. The reflected inertia scales with the square of the gear ratio:

J_reflected = J_load / (Gear Ratio)²

A 50:1 gearbox makes a 1 kg·m² load look like 0.0004 kg·m² to the motor. Usually this is fine, but if your motor's rotor inertia is tiny (like a coreless DC motor), the reflected load inertia can still dominate and affect acceleration performance.

Forgetting backlash. Every gear mesh has some backlash — dead zone where the output doesn't move when you reverse the input. Cheap spur gears might have 1-2° of backlash per stage. Through a 50:1 gearbox, that's 50-100° of slop at the input shaft. For positioning applications, this matters enormously. Preloaded planetary gearboxes or anti-backlash split gears solve this, at a cost. Using no-load motor speed. Your motor's rated speed is at no load. Under load, it slows down. A DC motor might be rated at 6000 RPM but only turn 4500 RPM when delivering its rated torque. If you sized your gearbox for 6000 RPM, your output speed will be 25% lower than expected. Thermal oversight. The power lost in the gearbox has to go somewhere. Small gearboxes in enclosed spaces can get surprisingly hot. I've seen 3D-printed gearbox housings soften and fail because someone didn't account for the 2W being dissipated in a tiny volume. ABS and PLA start getting soft around 60-80°C. Doesn't take much. Belt and chain "gearing." Pulley ratios work the same way mathematically, but the efficiency characteristics differ. Timing belts are typically 95-98% efficient and don't need lubrication. V-belts are worse, maybe 90-95%. Chain drives fall somewhere in between. The calculator works for these too — just adjust your $\\eta$ accordingly.

When High Ratios Become Problematic

There's a temptation to use very high gear ratios to get massive torque multiplication from a tiny motor. Past about 100:1 in a single gearbox, you start hitting practical limits.

Backlash accumulates. Efficiency compounds downward. The gearbox itself becomes a significant fraction of system cost and weight. And the output shaft moves so slowly that any control loop has terrible bandwidth — your motor might be able to change speed in 10 ms, but through a 200:1 gearbox, that translates to glacial output response.

For high ratios, consider a belt reduction stage followed by a gear stage, or use a cycloidal or strain wave (Harmonic Drive) reducer designed for the purpose. These hit 100:1 or 160:1 in a single compact stage with minimal backlash.

Try It

Plug your motor specs and gear tooth counts into the Gear Ratio Calculator to see your output speed, torque, and power. Adjust the efficiency to match your gear type — 96% for quality spur gears, 90% for budget planetary, 70% for worm drives. The calculator handles the math so you can focus on whether the numbers actually make sense for your application.