Size a stepper motor at the speed it will actually run. Read the torque-speed curve at your working speed, match the load inertia reflected to the shaft against the rotor’s own, then set microstepping for the motion quality you need. Holding torque alone will not tell you whether the axis will lose steps.
Most stepper selections fail for the same reason. The motor gets picked from its holding torque, the figure measured with the shaft held still and rated current in both phases. On a real axis the motor spends its life turning, and at speed the torque available is a fraction of that number. The three sections below work through the three figures that decide the outcome, using the Anaheim Automation range as the worked example.
Three numbers, in this order
Work through them in sequence, because each one constrains the next.
- Torque at speed. The motor’s pull-out torque at your working step rate, with a margin on the load.
- Inertia ratio. The load inertia reflected to the motor shaft, divided by the rotor inertia.
- Microstepping and drive voltage. Set after the first two, because they change smoothness and resolution, not the torque budget.
Step 1: read the torque-speed curve at your working speed
A stepper produces its rated holding torque at standstill and zero step rate. As the step rate rises, the drive has less and less time to force current into the winding inductance, so the current falls and the torque falls with it. The graph of torque against step rate is the pull-out curve, and the point where it bends down sharply is the corner speed.
Two consequences follow.
- Size against the curve, not the headline figure. Take the load torque at the operating speed, add a margin of roughly 30 to 50 percent, and find a motor whose curve stays above that requirement across the whole speed range of the move, acceleration ramp included. A motor that meets the load at cruise speed can still stall during acceleration, when the torque demand is highest.
- Inductance decides where the curve bends. Two Anaheim NEMA 17 motors can share the same 111 oz-in holding torque and behave completely differently. The 17Y408S-LW4 winding is rated 3.0 A with 2.4 mH per phase and holds torque to a much higher step rate. The 17Y402S-LW4 reaches the same 111 oz-in at 0.85 A but carries 29 mH per phase, so its torque falls away earlier. When the axis has to move quickly, the low-inductance winding is the one to choose.
Supply voltage sets the ceiling on speed. A common rule of thumb is to run the drive at around 32 times the square root of the winding inductance in millihenries, which puts a 2.4 mH winding near 50 V and a 29 mH winding near 170 V. Treat that as a starting point for the power supply rather than a motor rating, and check the drive’s own voltage limit first.
Step 2: match inertia, not just torque
Torque gets the load moving; inertia decides whether it arrives where you intended. During acceleration the motor must accelerate its own rotor as well as everything bolted to the shaft. If the reflected load inertia is far larger than the rotor inertia, the motor rings, overshoots and can lose steps at the start of a move even though its torque curve looked comfortable.
Convert the load to the motor shaft before comparing anything. A lead screw, belt or gearbox changes both torque and inertia, and inertia falls with the square of the reduction ratio:
J_reflected = J_load / N²Acceleration torque then follows the usual form, with the total inertia referred to the shaft:
T_accel = J_total × Δω / ΔtHere J_total is the rotor plus the reflected load, Δω the angular speed change in radians per second, and Δt the time allowed for acceleration.
The working rule for open-loop steppers is to keep the reflected load inertia below about five times the rotor inertia. Up to that ratio the axis behaves predictably; well beyond it the move becomes hard to tune and step loss appears under load. Take the rotor inertia from the Anaheim datasheet for the exact winding and stack length you are considering, because it changes with both.
When the ratio will not close, gearing is usually cheaper than a larger motor. A 5:1 planetary gearbox cuts the reflected load inertia to one twenty-fifth and multiplies the available torque, which is often enough to let a NEMA 17 or NEMA 23 frame carry a load that would otherwise demand a NEMA 34. The trade-off is added backlash and a lower maximum output speed, so gearing suits slower, heavier axes rather than fast indexing.
Step 3: set microstepping once the torque budget is fixed
Microstepping divides each full step into smaller commanded increments. A 1.8° motor has 200 full steps per revolution; at 1/8 microstepping the drive commands 1,600 increments, and at 1/16 it commands 3,200.
Microstepping makes motion smoother and noticeably quieter, and it damps the mid-band resonance that shows up as a vibration around a few hundred steps per second. What it does not do is raise positioning accuracy in proportion. The step angle error of the motor is non-cumulative and sets the real limit; Anaheim quotes 1.5 percent typical step accuracy on the 34K series. Resolution is not the same thing as accuracy.
There is a torque cost per increment as well. The incremental torque of a microstep is roughly proportional to the sine of the angle it commands, so a single 1/8 microstep produces only about a fifth of full-step torque, and a 1/16 microstep about a tenth. The motor still holds its rated torque under load, but each individual increment is weaker, and a slipping load will jump further when it lets go.
Practical settings are therefore a compromise:
- 1/8 or 1/16 for general machinery, where smoothness and low noise matter more than the last increment of torque.
- 1/2 or full step where maximum torque at speed is needed and the mechanics is already smooth, or where the controller cannot generate a high pulse rate.
- Higher microstepping on slow, precise axes, paired with an encoder if position has to be guaranteed.
Microstepping also multiplies the pulse frequency for a given speed, so a 1/16 setting needs sixteen times the pulses of full-step operation.
The Anaheim range at a glance
The table below summarises the frame ladder. Torque, current and step-angle figures are taken from Anaheim Automation’s published series data. The range as a whole runs from NEMA 08 to NEMA 42 and from 2.5 to 5,700 oz-in, in 1.8° and 0.9° step-angle variants.
| Series | Frame | Holding torque | RMS current | Step angle | Suits |
|---|---|---|---|---|---|
| 17Y | NEMA 17 | 24–111 oz-in (0.17–0.78 N·m) | 0.3–3.0 A | 1.8°, 0.9° variant | Light positioning axes, laboratory and bench equipment |
| 23Y | NEMA 23 | 175–425 oz-in (1.2–3.0 N·m) | 0.7–3.5 A | 1.8° | Machine-tool and router axes, medium loads |
| 34K | NEMA 34 | up to 2,790 oz-in (19.7 N·m) | up to 5.6 A | 1.8° | Heavy gantries, large screws, servo replacement |
Individually specified examples make the spread clearer. A 17Y202S-LW4 is a NEMA 17 at 50 oz-in and 0.85 A, an entry point for a light axis. A 17Y408S-LW4 is the same frame at 111 oz-in and 3.0 A. A 23Y204S-LW8 is a NEMA 23 at 262 oz-in and 1.4 A, and a 23Y310S-LW8 takes the same frame to 425 oz-in at 3.5 A. The 34K series reaches up to 2,790 oz-in and, with four stack lengths available, is quoted as giving roughly 25 percent more torque than the 34Y and 34N series. Anaheim also lists it as compatible with the Pacific Scientific K34 series, which matters when an existing K34 is being replaced. One availability note: Anaheim’s own 34K series page currently carries a rare-earth supply notice, so confirm the position before designing that series in.
Drives matter as much as motors. For the 17Y series Anaheim lists the MBC15081 and MBC25081TB drives, and an integrated motor-and-driver option in the same frame; for the 34K it lists the MBC05641, MBC12101, MBC10641, MBC082561 and MLA10641 series. The range also includes IP65 sealed versions such as the 17Y65 and 23Y65 for washdown and humid areas. Encoders, 24 VDC brakes, MiniFit Jr. connectors and planetary gearboxes are available as factory adders.
Which one for which job
- Low-inertia positioning, light screws and belts, loads up to about 0.2 N·m: start with the 17Y series. A 17Y202S-class winding covers a moderate duty; step up to the 17Y408 windings when the axis needs torque at speed.
- Machine-tool and router axes with medium loads, or any NEMA 17 duty that runs out of torque: use the 23Y series, from 262 oz-in (23Y204S) to 425 oz-in (23Y310S).
- Heavy vertical axes, large ball screws, or a servo being replaced on torque density: the 34K series, up to 2,790 oz-in.
- Reflected inertia ratio above five to one: add a planetary gearbox instead of stepping up a frame size.
- A vertical axis that must hold position on power loss: specify the brake option rather than relying on holding current.
- Washdown, condensation or a humid plant: specify the IP65 sealed variants.
Send the numbers, not a model number
If you have an existing motor, the nameplate plus a photograph is enough to start. For a new design, four figures let us size it in one pass: the load torque at the working speed, the maximum step rate, the reflected load inertia or the mass, screw lead and reduction ratio needed to work it out, and the supply voltage you have available. From those we can name the frame, the winding and a matching Anaheim drive, and confirm holding torque and rated current against the datasheet before anything is ordered.
- Browse the Anaheim Automation range we carry.
- Compare the wider motors and gearboxes we supply.
- Read more selection and cross-reference guides in the Insights library.
- Send the details straight to the inquiry form.
Every motor and drive we supply is original and new, bought through the manufacturer or established distribution channels, with procurement documentation for verification on arrival.