Blog · 2026-07-05 · Motors
Every robot project reaches the same fork in the road: which motor? This guide walks the whole decision — motor type, torque, speed, current, voltage and gearing — ending with a complete worked example you can copy for your own build.
Choosing a motor feels overwhelming because sellers describe motors in six different units, datasheets bury the numbers that matter, and forum advice contradicts itself. But underneath the noise, motor selection is a short chain of questions with definite answers: How much force does the job need? How fast? For how long? Answer those, and the motor almost picks itself. Let's take the questions in order.
You cannot pick a motor from vibes. You need three numbers before you shop:
Torque — the rotational force required, usually in N·m or kg·cm. For a drive wheel it comes from your robot's weight, the surfaces it climbs, and how hard it accelerates. For an arm joint it comes from the payload and the lever arm.
Speed — the RPM required at the output. For wheels: RPM = (speed × 60) ÷ (π × wheel diameter). A robot doing 1 m/s on 80 mm wheels needs about 239 RPM at the wheel — a number worth memorizing, because it tells you immediately that a bare 10,000 RPM motor is nowhere near a wheel without gearing.
Power — torque × angular speed. Power is the great honest number: a motor that must deliver 0.35 N·m at 240 RPM is delivering about 8.8 W of mechanical power, and no clever gearing changes that total. Gearing converts between torque and speed; it never creates power.
If your motor is a drive motor for a wheeled robot, our Motor Sizing Calculator computes all three numbers from your robot's weight, wheel size, speed and terrain — do that first and come back with your spec.
Four motor families cover essentially all hobby robotics. Each trades simplicity against performance in a different place.
A brushed DC motor with an attached gearbox is the workhorse of hobby robotics, and for good reason: apply voltage, it spins; reverse voltage, it reverses; PWM the voltage, it changes speed. Control needs nothing more than a cheap H-bridge driver. The gearbox converts the motor's natural high-speed/low-torque output into wheel-friendly numbers. Downsides: brushes wear over hundreds of hours, efficiency is moderate (60–75%), and there's no built-in position feedback — add an encoder if you need one. For a first robot, or any robot where "it just needs to drive," a brushed gearmotor is almost always the right answer.
Brushless motors replace mechanical brushes with electronic commutation, giving far higher efficiency (often 85%+), power density, and lifespan. They dominate drones and combat robots, where watts-per-gram decides everything. The cost is complexity: they need an electronic speed controller (ESC), and precise low-speed control requires sensored motors or FOC controllers. Choose brushless when you need serious power in a small package — fast rovers, drones, weapon motors — and are ready for the extra electronics.
Stepper motors move in fixed angular increments (typically 1.8° per step), so you can command "300 steps" and know where the shaft is — no encoder required, as long as you never exceed the torque limit. That's why 3D printers and CNC machines are full of them. Their weaknesses: torque collapses at high speed, they draw full current even standing still, and if overloaded they skip steps silently and lose position. Choose steppers for slow, precise positioning: plotters, camera sliders, small CNC.
A hobby servo bundles a motor, gearbox, position sensor and controller into one case that moves to a commanded angle. Perfect for arm joints, grippers, steering, pan-tilt. Standard servos rotate ~180°; continuous-rotation versions behave like slow gearmotors. See our complete servo guide for classes and specs; if you're sizing one for an arm, the Servo Torque Calculator does the math.
| Type | Control | Best at | Weakness |
|---|---|---|---|
| Brushed DC gearmotor | Trivial (H-bridge) | Drive wheels, general motion | Brush wear, no feedback |
| Brushless (BLDC) | Needs ESC | High power density | Cost, low-speed control |
| Stepper | Step/dir driver | Open-loop positioning | Loses torque at speed |
| Hobby servo | One PWM wire | Angular positioning | Limited rotation, ~180° |
Rated (continuous) torque and current — what the motor can deliver all day without overheating. This is the number you size against. Stall torque and stall current — the maximum, at zero speed, for seconds only. A motor held at stall is a heater. Your requirement should sit at roughly one-third to one-half of stall torque; equivalently, choose stall torque at least 2× your continuous need. No-load speed — RPM with nothing attached; expect roughly 75–85% of it under real load. Rated voltage — run at it, or slightly under; overvolting raises speed and wear together. Stall current matters twice more: your motor driver must survive it (an L298N dies at currents a BTS7960 shrugs off), and your battery must supply it — check that with the Battery Runtime Calculator's C-rating output.
Almost every robot needs reduction between motor and wheel. You can build gear trains yourself, but for drive systems the sensible move is buying a gearmotor with the ratio already integrated: N20 micro gearmotors for tiny bots, 25 mm gearmotors for lightweight platforms, 37 mm gearmotors for 3–10 kg robots. To find the ratio your build needs — and see how much torque it multiplies — use the Gear Ratio Calculator, and read Gear Ratios Explained for the theory.
Let's size motors for a realistic build: a 4 kg indoor robot (including battery and payload), two driven wheels of 90 mm diameter, target speed 0.8 m/s reached in 2 seconds, occasionally climbing a 12° ramp, running on wood floors (rolling resistance coefficient ≈ 0.015).
Force. Incline force = 4 × 9.81 × sin(12°) = 8.16 N. Rolling resistance = 0.015 × 4 × 9.81 × cos(12°) = 0.58 N. Acceleration force = 4 × (0.8 ÷ 2) = 1.6 N. Total worst-case force ≈ 10.3 N.
Torque. Wheel radius = 0.045 m, so total torque = 10.3 × 0.045 = 0.46 N·m. With a 1.5 safety factor and split across 2 motors: 0.35 N·m ≈ 3.5 kg·cm per motor, rated.
Speed. Wheel RPM = (0.8 × 60) ÷ (π × 0.09) = 170 RPM.
Power. Per motor with safety factor: (10.3 × 0.8 × 1.5) ÷ 2 ≈ 6.2 W.
Shopping translation: a 12 V, 37 mm brushed gearmotor with roughly a 1:30 gearbox, listed around 170–200 RPM and ≥ 3.5 kg·cm rated torque (so ≥ 7 kg·cm stall). That is an abundant, inexpensive class of motor — which is precisely the point of doing the math: you discover you don't need the exotic option. Check stall current (likely 3–5 A each), pick a driver rated above it, and size the battery for roughly 10 A peak.
Shaft diameter — do wheels or couplings exist for it? Mounting — is there a bracket pattern, or will you be improvising with zip ties? Encoder option — if you'll ever want PID speed control (you will; see the PID Visualizer), buying the encoder version now is cheaper than retrofitting. Weight — motors are often a robot's heaviest components after the battery, and that weight feeds back into the torque requirement you started with. And buy a spare: hobby motors from the same batch vary, and a matched spare saves a build week later.
Work the chain in order — force → torque → speed → power → type → specific motor — and motor choice stops being a gamble. Ten minutes with a calculator replaces weeks of "it hums but doesn't move." Start with the Motor Sizing Calculator, and if the torque units still feel foreign, the next article — Motor Torque Explained — untangles N·m, kg·cm and oz·in for good.
The same 6 W, 170 RPM, 3.5 kg·cm specification is available at three price points, and it's worth knowing what changes. Bottom tier (₹150–400 / $2–5): yellow TT motors and unbranded gearmotors — plastic gears, loose specs, unit-to-unit variation of 10–20% in speed (why cheap two-motor robots curve), fine for learning and prototypes. Mid tier (₹700–2,000 / $8–25): 25/37 mm metal-gear gearmotors from reputable hobby suppliers — honest datasheets, encoder options, the sensible default for any robot you intend to keep. Top tier ($40+): Maxon/Faulhaber-class or quality planetary BLDC — precision, efficiency and lifespan for machines that must work daily. The worked example above lands squarely in the mid tier, and that's the general pattern: doing the sizing math usually reveals the mid tier is enough, which is precisely the money the math saves.