How to Size a Motor for Your Application Without Guessing

Picking a motor from a peak torque number on a datasheet is the fastest way to learn the wrong lesson on the bench. The motor that "should work" stalls under real friction, cooks in continuous duty, or trips a supply that looked fine on paper.

You do not need a full electromechanical PhD to size correctly. You need a short list of numbers, a worst case you actually believe, and a way to stress them before DigiKey ships. This post walks that path.

Start with the job, not the catalog

Write down what the motor has to do before you open a manufacturer site:

1. What load moves? Mass, moment arm, spring force, cutting friction, wind load, whatever applies.
2. How fast, and how often? Continuous motion, intermittent hops, or rare peak events.
3. What fails if it is short? Missed cycle time, overheat, stalled actuator, breaker trip.
4. What does the rest of the product allow? Voltage rail, current budget, gear train volume, noise, cost.

If you cannot answer those four, any motor you pick is still a guess. Catalog pages will not invent your load profile for you.

Convert the job into torque and speed

Most sizing math collapses to continuous torque, peak torque, and speed at the shaft you care about.

Load torque is the steady force (or moment) the motor fights after gearing. For a lever or arm, that is force times radius. For a lead screw, convert linear force through pitch and efficiency. Prefer measured friction or a conservative estimate over an optimistic CAD number.

Acceleration torque is Jα: inertia at the motor shaft times angular acceleration. Include reflected load inertia through the gear ratio. Short moves with heavy loads often care more about peak than continuous.

Required speed is the shaft rate that still meets cycle time after gearing. Spec continuous RPM with margin; do not size to the absolute maximum the motor can spin unloaded.

A workable first pass:

T_req ≈ T_load + T_accel + T_margin

Margin is not superstition. Friction rises with wear, grease cold-thickens, and manufacturing tolerances stack. Ten to thirty percent is common early; raise it when the cost of under-sizing is a respin.

Match continuous vs peak to the duty cycle

Datasheet peak torque is almost useless alone. Ask:

• How long does the peak last?
• How often does it repeat?
• What heatsinks or airflow exist in the enclosure?

Continuous torque is the number that keeps winding temperature inside the rating for your duty. Peak torque is the number you may touch briefly if RMS current over the cycle stays legal and the supply can deliver it. A motor that hits peak for 200 ms every few seconds can be fine. The same peak held for minutes is a different part.

If the duty is intermittent, estimate RMS torque or RMS current over a representative cycle. If continuous, size to continuous rating with thermal margin, not to the marketing peak.

Check the electrical side early

Torque without voltage and current is incomplete:

• Stall current (or peak driver current) must fit your supply and FET / driver rating.
• Voltage must still produce usable speed under load at the end of cable drop and converter sag.
• Controller limits matter: software current caps, folding-back supplies, and USB-PD negotiation all change what the motor actually sees.

A motor that meets torque but demands 8 A continuous on a 2 A rail is not sized. It is postponed failure.

Gearing is part of sizing

Gearboxes multiply torque and divide speed. They also add backlash, efficiency loss, and reflected inertia. Before you upsize the motor, ask whether a different ratio clears continuous torque without blowing the current budget. Sweep ratios on paper or in a model: required motor torque vs shaft speed vs supply current. The sweet spot is often a smaller motor with sensible gearing, not the largest catalog SKU that fits the envelope.

Validate with a system model before you order

Once you have candidate parameters (Kt, resistance, inertia, gear ratio, supply limits, load profile), run them as a small system: voltage source, motor, gear/load, and your proposed current limit or duty cycle. Sweep load torque and supply voltage. Watch shaft speed, current, and settling under the ugly cases, not the happy path.

That is where guessing ends. You learn whether stall under max load trips the rail, whether continuous duty exceeds continuous current, and whether a slightly different ratio or rail fixes it without another purchase order.

With a Python system model you can change gear ratio or supply voltage in a parameter edit and re-run, instead of waiting on another courier. Tools like Hardwave are built for that workflow: typed components, a graph, and local steady-state or transient solves so motor + supply + load live in one place before hardware arrives.

A short checklist you can reuse

• Load torque and peak torque with stated assumptions
• Speed that meets cycle time after gearing
• Continuous vs peak duty called out explicitly
• Supply voltage and current budget under sag
• Thermal reality of the enclosure
• Gear ratio swept, not frozen on the first guess
• Worst-case run in a model before the order

Physical prototypes still matter. Friction curves, seal drag, and firmware timing will surprise you. Size so the first bench session validates a design that already survived the electrical and mechanical corners, not so it discovers that a 12 V hobby motor was never going to move the load.