How-to
Following a Master Axis With a Stepper Traverse
Short answer
Gear the traverse stepper to measured spool rotation, not to a fixed speed ratio: commanded traverse speed is v_t = p × n_s, so pitch stays constant as the spool fills. The follower ratio comes from pitch, screw lead and drive resolution — 25 step pulses per 64 master counts in the worked example — and master acceleration must be limited so the amplified slave ramp cannot stall the open-loop stepper.
On this page
- What should the traverse follow: spool speed or wire speed?
- How do I calculate the step pulse frequency?
- What master-to-slave ratio do I program?
- How do I connect an AC-driven spool to a stepper drive?
- Will the stepper stall when the spool accelerates?
- How does the traverse reverse at the end of each layer?
- When should I use a servo instead of a stepper for the traverse?
- Questions asked
Gear the traverse to measured spool rotation and the winding pitch holds itself through the whole fill. Mount an encoder on the spool shaft (or take any signal that counts true spool revolutions), treat it as the master, and command the stepper so the guide travels the pitch p per spool revolution: v_t = p × n_s. As the spool fills at constant wire speed, spool rotation slows and the traverse slows with it, so no diameter calculation ever enters the ratio. This walkthrough sizes the step pulse frequency, sets the electronic-gear fraction, links an AC-driven spool to a pulse-driven stepper, and checks the one failure an open-loop axis cannot report: a stall on acceleration.
What should the traverse follow: spool speed or wire speed? #
Follow measured spool rotation. It is the only reference that holds winding pitch without a diameter estimate, because v_t = p × n_s in distance per minute and n_s already contains every diameter change. Everything else derives spool speed and inherits estimation error.
If wire delivery speed is held constant and the effective winding diameter D doubles, spool speed n_s = v_wire / (π × D) halves, and traverse speed halves with it — 600 mm/min at the core, 300 mm/min at twice the diameter in the example below. Following the AC motor shaft instead of the spool is correct only if the drive ratio between them is fixed and slip-free; put the reference where the wire actually winds.
| Master reference | What you must know | Result for pitch |
|---|---|---|
| Spool shaft encoder | counts per spool revolution C | constant by construction — use this |
| Wire speed plus diameter | v_wire and effective D | correct, but D error becomes pitch error |
| Wire speed alone | line speed only | pitch drifts as D grows — not sufficient |
How do I calculate the step pulse frequency? #
The frequency is f_step = p × n_s × G × N / (60 × L) in pulses per second, where L is screw lead in distance per screw revolution, G is motor revolutions per screw revolution, and N is the drive's commanded pulses per motor revolution.
Worked example, with every input listed:
| Input | Meaning | Example value |
|---|---|---|
| p | winding pitch | 2 mm per spool revolution |
| n_s | spool speed | 300 rev/min |
| L | ball screw lead | 10 mm per screw revolution |
| G | motor rev per screw rev | 1 |
| N | commanded pulses per motor rev | 2000 (ten microsteps on a 1.8° motor) |
Traverse speed: v_t = p × n_s = 2 mm/rev × 300 rev/min = 600 mm/min.
Motor speed: n_motor = v_t × G / L = 600 × 1 / 10 = 60 rev/min.
Pulse frequency: f_step = 600 × 1 × 2000 / (60 × 10) = 1,200,000 / 600 = 2000 pulses per second.
Cross-check: 60 rev/min is 1 rev/s, and 1 rev/s × 2000 pulses/rev = 2000 pulses/s. The millimetres and revolutions cancel and the factor 60 converts per-minute to per-second, leaving pulses/s. This is a calculated example for these inputs, not a measured result.
Read lead as travel per screw revolution — on a multi-start screw, thread pitch and lead are not the same number. Raising N (finer microstepping) smooths the slow traverse but multiplies f_step, so keep the result under the drive's maximum input frequency and the controller's output rate, both from their manuals.
What master-to-slave ratio do I program? #
Program the fraction p × G × N / (L × C) in slave pulses per master count: with C = 1024 counted increments per spool revolution, that is 2 × 1 × 2000 / (10 × 1024) = 4000 / 10240 = 25/64, or 0.390625 pulses per count.
Enter it as a gearing numerator and denominator (25 over 64), never as the rounded decimal: the controller must carry the fractional remainder forward, so 64 master counts produce exactly 25 pulses and nothing is lost per layer. C is the controller's effective count — the decoding mode multiplies an encoder's line count, and any belt or gear ratio between the encoder mount and the spool divides it.
Use position-based gearing, which accumulates master counts and commands the corresponding slave displacement, rather than a speed-only follower. A speed follower holds the ratio while running but drifts through ramps and stops, and pitch registration across reversals is exactly what drifts.
How do I connect an AC-driven spool to a stepper drive? #
Through a motion-capable controller: its high-speed input counts the spool encoder and its pulse-and-direction output drives the stepper; the AC drive keeps running the spool and only needs its acceleration ramp limited. An analog speed signal cannot command a pulse-input stepper without conversion.
The AC drive's 0-10 V speed monitor output is a legitimate reference for a drive that accepts one, but a stepper needs pulses, so the controller does the conversion. Commissioning sequence:
- Mount the encoder on the spool shaft, configure the counter's decoding, and confirm the count rate at full speed: 300 × 1024 / 60 = 5120 counts per second.
- Wire the controller's pulse and direction outputs to the stepper drive; set N and confirm the resulting f_step stays inside the drive's input limit.
- Enter the follower ratio as the fraction 25/64 in the controller's electronic-gear registers.
- Set the AC drive's acceleration ramp to reach 300 rpm in 2 s, the ramp the stall check below sizes.
- Program the traverse stroke limits and the reversal behavior at each end.
- Run slowly and verify measured travel against counted spool revolutions.
- Run at maximum spool speed and through several accelerations, re-verify travel, then wind real product.
Before step 6 crosses to live wire, run with guards closed and interlocks proven as part of the machine's risk assessment under ISO 13849-1, which grades a guard-interlock function by the risk it controls; this setup procedure is not a safety validation.
Will the stepper stall when the spool accelerates? #
Only if the master's ramp is too short. The gearing amplifies acceleration by the same displacement ratio p × G / L = 0.2 motor revolutions per spool revolution, so a 2 s ramp to 300 rpm demands 1000 pulses/s² of step frequency, while a 0.2 s ramp demands 10,000 pulses/s².
Master acceleration of 300 rpm over 2 s is 2.5 rev/s², which becomes 2.5 × 0.2 = 0.5 rev/s² at the motor. The torque to produce it is T = J × 2π × α; with an illustrative combined inertia of 1.2 × 10⁻⁴ kg·m² (rotor plus reflected screw and carriage — take the rotor value from the motor datasheet), T = 1.2 × 10⁻⁴ × 2π × 0.5 = 3.8 × 10⁻⁴ N·m, negligible next to friction. The deciding checks are therefore elsewhere: starting without any ramp lands the motor above its start-stop region, and some speed bands resonate. Read pull-out torque at 60 rpm from the torque-speed curve in the motor's datasheet and require it to cover load plus acceleration torque with about a 2:1 margin — a rule of thumb, not a rating.
The deeper limit is informational: an open-loop stepper reports no following error, so a stall is silent. On follower axes that do have feedback, documented practice sets the deviation tolerance near 10% of master resolution (for instance 2000 pulses on a 20,000 pulse-per-revolution encoder) and watches the peak at maximum master speed. With your open-loop axis, the measured-travel verification after the acceleration test is the equivalent check.
How does the traverse reverse at the end of each layer? #
The controller does it, not the gear ratio. Track accumulated traverse position, and when it reaches a stroke limit, flip the direction signal or mirror the geared displacement about that limit; a fixed master-to-slave ratio never reverses anything by itself.
Because position-based gearing accumulates counts across reversals, pitch registration survives the turn. The velocity sign flip at reversal passes through zero at only 600 mm/min of carriage speed, so the demand is mild. What needs care is the layer edge: a short dwell or a set overtravel at each limit forms the turnover, and if the controller cannot mirror inside its gear function, trigger a short profiled move per reversal from a position compare instead.
When should I use a servo instead of a stepper for the traverse? #
Swap when an undetected lost step costs more than the closed-loop hardware: high line speeds you cannot ramp gently, a stroke that pushes f_step deep into the stepper's torque droop, or a wind where pitch error scraps product or jams the machine.
Decision for this class of machine: at 2000 pulses/s with a 2 s master ramp, a stepper sized from its torque-speed curve with the margin above is the economical fit, and the spool motor never needs to become a servo just to serve as master — the encoder on the driven spool shaft is the master. If the process forces fast catch-ups (splicing, thread-ups) that break the ramp limit, or the verification travel ever comes up short at speed, move the traverse to a servo or a closed-loop stepper so the drive reports following error and stops the wind on fault instead of drifting.
Questions asked
Do I need to measure the spool diameter to keep the pitch constant?
Not if the stepper follows measured spool rotation. Diameter enters only when you derive spool speed from wire speed, n_s = v_wire / (π × D), and any error in D becomes pitch error directly. With an encoder on the spool shaft the diameter effect is already inside the measured rotation, so the ratio stays 25/64 for the whole wind.
Can I use the AC drive's 0-10 V speed output to run the stepper?
Not directly: a pulse-input stepper drive needs step pulses and a direction signal, not a voltage. Feed the analog signal to a controller that generates the scaled pulse train, or use it only as a speed indication. Analog scaling also drifts, so a position-based encoder follower holds pitch better over a long wind.
How do I prove the follower ratio is set correctly?
Count spool revolutions with the controller and measure the carriage travel. For a 2 mm pitch, 25 spool revolutions must move the guide 50 mm: 25 × 2 mm = 50 mm. Repeat at maximum spool speed after the acceleration test; a shortfall there means the motor lost steps on the ramp, not that the ratio is wrong.
What happens if the stepper loses steps during a wind?
Nothing tells you: an open-loop stepper has no following error to report, and the layers drift from the point of the stall onward. If that scraps wire or jams the machine, add an encoder on the traverse for monitoring or stall detection, or change the axis to a closed-loop stepper or servo so the control stops the wind on fault.
How do I choose the microstep setting N?
N is a drive setting: 2000 pulses per motor revolution is ten microsteps on a standard 1.8° motor. Finer settings smooth low-speed motion but multiply f_step, so keep the result under both the drive's maximum input frequency and the controller's output rate, taken from their manuals, at your highest spool speed.