The message usually arrives after a bad day: "my machine lost position - the part is off by a few millimeters and it only happens sometimes." When a stepper-based machine loses position, the axis did not drift and the controller did not glitch. It skipped steps, and because a stepper motor is an open-loop device, nothing in the system knew - the controller kept counting steps that never happened.
The good news is that losing steps is almost never a motor failure. It is a setup failure, and there is a fixed order to check things in. We have debugged enough of these over the years that the sequence below catches the cause in nine cases out of ten.
The order we check, every time
1. Driver current setting. This is the one. The DIP switches on the driver set the phase current, and if they are left at the factory default while the motor is rated for more, the motor runs with half its torque. The stepper datasheet lists rated current per phase - set the switches to match, and a surprising number of "losing step" machines simply stop losing steps.
2. Speed and acceleration. A stepper's torque collapses as speed rises - most NEMA 23 motors have lost half their holding torque by a few hundred rpm, and the curve is steep, not gentle. If your axis accelerates too fast or runs near the top of the curve, it will stall at the worst moments. Lower the acceleration, or pick a higher-torque motor for the speed band you actually use.
3. Supply voltage. The driver needs headroom. Running a 24 V supply when the driver is rated for 48 V caps your high-speed torque; raising the bus voltage is the cheapest torque upgrade a stepper system can get.
4. Resonance. Every stepper has a resonant speed band where it vibrates and loses steps at light load. Micro-stepping and a higher driver resolution move the problem out of your operating range - a 1/16 micro-step setting usually kills it entirely.
If the machine is still losing steps after all four, the honest answer is that the axis is asking for more than the motor can give - and that is when we talk about a bigger motor or a closed-loop stepper. A closed-loop stepper adds an encoder on the back of the motor; when the driver sees the shaft fall behind the commanded position, it corrects instead of counting on. It is not a servo, but for most cutting and positioning axes it behaves like one for a fraction of the cost.
Our own high torque stepper motor kit is the one we point to when the diagnosis ends at "the motor is undersized" - the 86 frame holds roughly double the torque of a NEMA 23 at the same speed, which buys you the same margin in one swap. And yes, we make closed-loop kits too, with the encoder already matched to the driver, so there is no integration surprise.
If you send us the motor frame size, the supply voltage and what the axis actually does - through the inquiry form - we will tell you which of the four settings is most likely wrong, and if it is not a setting, which motor fixes it. The person answering has walked through this exact checklist more times than we can count.
