Ask any five engineers how fast can a stepper motor turn and you'll get five different answers. That's because the question is missing the context that matters: load, voltage, driver, duty cycle, and the rest of the mechanical chain. The same is true for specifying a Thomson linear actuator. There is no universal answer.
I'm a motion control applications engineer. I've been handling linear motion and actuator orders for eight years. I've made and documented 14 significant mistakes, totaling roughly $42,000 in wasted budget. Now I maintain our team's pre-order checklist. This article is the guide I wish someone had given me in 2017.
Three Scenarios, One Spec Rule
Before you choose anything, stop working through the checklist in your head. The right actuator, shaft, bearing, and switch package depends on what the machine is doing. In my ordering flow, I split every request into three buckets:
- Scenario A: Compact short-stroke axes. Think pick-and-place, lab automation, packaging.
- Scenario B: High-thrust, high-duty-cycle axes. Think lift tables, presses, machine feed.
- Scenario C: Coordinated multi-axis systems. Think gantries, robotic cells, assembly machines.
Each bucket gets a different answer. Mixing them up is where the money disappears.
Scenario A: Compact Axes, Small Ball Bearings, and Shaft Support
If you're moving under 50 kg over a short stroke, a stepper-driven actuator with small ball bearings riding on a precision shaft is usually the right call. A hardened Thomson 60 Case LinearRace shaft (the classic precision shafting family) is the product line I start with. It's available in many diameters and can be supported or unsupported.
For compact axes, use a supported shaft. Yes, it costs more. I know. But an unsupported shaft deflects under load, and a small ball bearing will bind in the middle of a stroke. I learned this on an order of 12 manual positioning stages in 2018. The vendor shipped them, they worked beautifully on the bench, and then jammed on the customer's machine. The culprit was an 8 mm shaft with no center support. The whole order came back.
A Thomson linear shaft is your starting point. Match the diameter to the load and the unsupported length. For every 300 mm of unsupported span, diameter matters much more than most engineers want to admit. Use the supplier's load capacity table. Don't guess.
Small ball bearings sound simple, but the tolerance between the bearing bore and shaft diameter has to be correct. A too-tight fit creates preload; a too-loose fit creates play. Both cause issues. I once stripped an aluminum housing because I pressed a bearing at an angle. The $15 bearing turned into a $120 mount replacement.
So how fast can a stepper motor turn? The misleading number is 2,000-3,000 rpm. Unloaded, some NEMA 23 stepping motors will spin that fast. Under load, torque drops quickly. For a small 1.8° stepper driving a 5 mm lead screw, I plan around 300-600 rpm at the motor. That gives you roughly 1.5-3 m/min linear speed. If you need more, increase the lead or switch to a servo.
Even on a compact axis, add switches for electric actuators on both ends. Relying on software limits is a mistake. In my first year (2017), I set up a controller with no hardware limit switch because I wanted to save $40. The carriage ran to the mechanical stop and bent a 10 mm shaft. $350 repair, one lost week. Every future order got mechanical switches.
Scenario B: High-Thrust, High-Duty-Cycle Thomson Linear Actuators
If the payload is above 100 kg or the duty cycle is above 40%, a stepper motor usually isn't enough. Use a true electromechanical actuator, a ball-screw or belt-driven Thomson linear actuator, with an AC induction or servo motor. Don't put a tiny stepper on a 2 kN actuator and hope for smooth acceleration. It will miss steps and fault out.
In this scenario, the motor is only half the story. The shaft and bearing system must handle radial and axial loads. A high-thrust actuator with a cantilevered load will wobble. I recommend using a linear guide rail rather than a round shaft if there's any moment load.
Use both mechanical limit switches and drive-based torque limiting. I don't have hard data on industry-wide crash rates, but based on my repair calls, most high-thrust actuator damage comes from a moving carriage hitting something before the controller sees the fault. Mechanical switches for electric actuators are the final line of defense. They kill power independently of the PLC. That matters.
Here's the counterintuitive part: upgrading to a ground ball screw is often the wrong upgrade if your mounting structure flexes. I once had a customer insist on a precision ball screw while the bracket deflected 0.3 mm under load. The machine still produced inaccurate parts. The screw wasn't the problem. Stiffness was.
That's when I adopted the quality-perception view: every visible wobble, every scratched shaft, every noisy bearing gets tied to your brand. A slightly more expensive bearing or a stiffer housing isn't an expense. It's a customer's first impression of your machine. In 2022, I switched a client from a budget actuator line to a Thomson linear actuator for a high-duty-cycle machine. Customer feedback on smoothness improved noticeably. I can't prove the actuator alone caused the repeat purchase, but the repeat purchase happened.
For ball screws, ISO 3408 defines acceptance criteria for lead accuracy. If the machine needs repeatability, choose the right tolerance class. But again, a tight screw won't help if the housing flexes.
Scenario C: CANopen CiA 402 Control with Thomson Linear Actuators
The third scenario is where engineers lose the most time: coordinated multi-axis control. You have a PLC or motion controller, a CANopen network, and a drive powering a Thomson linear actuator. The actuator itself is not CANopen. The drive is. And the drive must speak the right profile.
CANopen is the network protocol. CiA 402, maintained by CAN in Automation, is the device profile that defines motion states and modes. When a customer sends an RFQ for a 'thomson linear actuator canopen cia 402' setup, what they usually mean is: the drive is configured for CiA 402 and controlled over CANopen using profile position or velocity mode.
Set the node ID and baud rate first. I wish I had counted how many 'CANopen isn't working' calls ended with two nodes on the same ID or missing bus termination. Check those before you dig into the object dictionary.
In CiA 402, the drive runs through a state machine. You can't just send a position and expect it to move. You have to enable the drive in the correct order: shutdown, switch on, enable operation. Most software handles this if you set the controlword correctly. But if you're hand-writing PDOs, you'll forget a step. I know because I have.
You also can't skip the homing procedure. And even with homing, keep hardware limit switches. A homing routine needs a reference; don't use a hard stop as home. That will shorten the life of the actuator and the switch.
For gantries, use profile position mode and let the PLC generate the trajectory. If you need true synchronized motion, look at interpolated position mode (pvT) instead of trying to coordinate with individual position commands. A stepper with a CiA 402 drive still obeys its torque-speed curve. The network profile doesn't change physics. If the axis misses steps at 1,000 rpm, slow down or change the mechanics.
Which Scenario Are You In?
Here's how I judge which scenario a customer is in:
- Quick, repeatable cycles under 50 kg: Scenario A. Don't over-engineer it.
- High force, high duty, or visible moment loads: Scenario B. Stiffness wins.
- Multiple axes that need to move in sync: Scenario C. Plan the bus and the drive profile before you order the actuator.
If you're between A and B, choose B. A bigger actuator costs less than a service call.
This isn't a complete engineering guide. I wish I had data on every failure mode, but I don't. What I have is a checklist that has caught 47 potential errors in the past 18 months. The checklist: confirm the load, confirm the duty, confirm the control profile, and never skip the limit switches.
The cheapest component isn't the one that costs the least on the invoice. It's the one that survives the first year in the field. Years ago, I'd have called that a slogan. After rebuilding a customer's actuator at my own expense, I call it a lesson. (mental note: I really should turn this into a public checklist, not just an internal one.)