Four or five times a week a nameplate photo lands in our inbox with one line under it: what servo driver do I need for this? Sometimes the motor is new, sometimes it came off a machine that stopped mid-job, sometimes it has no paperwork at all. The photo is usually blurry, but a nameplate tells us almost everything.
There's no such thing as "the driver for this motor." Sizing a servo motor driver means matching three things - bus voltage, current and encoder feedback. Get those right and most industrial drives will run the motor. Get one wrong and the pair fails in a way that looks like a motor fault, which is why so many mismatches end with a new motor being ordered for nothing.
Why the nameplate power rating is the worst place to pick a servo driver
Nobody reads the power rating first. Everybody reads the power rating first.
It usually goes like this: someone sees 1.5 kW on the motor, buys a 1.5 kW drive, and calls it a match. It isn't one, it's a coincidence. Power is derived - torque times speed - and it says nothing about whether the drive can deliver the current the motor wants when it wants it. A 1.5 kW motor pulling 6 A continuous and 18 A while accelerating doesn't care that its driver is also "1.5 kW" if that drive peaks at 12 A. It trips the first time the axis reverses at speed. Two motors with the same kilowatt rating can have very different current ratings, too - one 3000 rpm design, one low-speed high-torque unit behind a reducer.
Step 1: match the DC bus voltage to the motor's rated voltage
A servo drive rectifies incoming AC into a DC bus that feeds the output stage. That bus, not the mains, is what the motor sees. The question isn't "220 V or 380 V on site," it's "what bus does this winding expect."
| Motor rated voltage | Typical supply | Rectified DC bus |
|---|---|---|
| 48 V class (closed-loop stepper, small servo) | Single phase 220 V | approx. 65–70 V |
| 220 V class servo | Single phase 220 V | approx. 310 V |
| 380 V class servo | Three phase 380 V | approx. 530–560 V |
Put a 220 V motor on a 380 V drive and you stress the winding insulation and the output stage with a bus about 1.7 times higher than intended. Put a 380 V motor on a 220 V drive and it turns, but never reaches rated speed - the bus can't overcome the back-EMF. We've seen routers and plasma tables that "got slower over time" when they were mismatched from day one. Stuck with 380 V on site and a 220 V motor? Use a step-down transformer. A 380 V drive with the current limit turned down doesn't lower the bus.
Step 2: size on continuous and peak current, not rated current
This is where the real decision gets made, and where most people asking how to size a servo driver trip up.
Every datasheet gives a continuous current and a peak current. A 750 W 220 V servo typically runs 2.6 A continuous and 8 A peak - roughly three times the continuous figure, normal for small frames with no fan cooling the winding. A bigger 3.8 kW servo sits near 15 A and 45 A. Read your own datasheet, not our averages.
Then check two things against the drive:
- Continuous rating ≥ motor continuous current. Add headroom for ambient - at 45 °C in a cabinet, derate by 10 to 20 percent.
- Peak rating ≥ motor peak current, lasting long enough to cover a real ramp, typically two to three seconds. A drive that peaks for 200 ms and folds back will still alarm.
One frame up is common and harmless. Undersizing kills drives slowly, by thermal cycling, until the output module is gone. And a lowering load on a vertical axis pushes energy back into the bus - a regen resistor problem, not a current problem.
Step 3: encoder type, the thing everyone forgets
Voltage and current are engineering. Encoder matching is archaeology.
An incremental encoder - 2500 lines, A/B/Z differential, 5 V line driver - is the classic setup and works with a huge number of drives. An absolute encoder is a different animal: 17-bit or 23-bit, single or multi-turn, carrying position on a serial protocol. Protocols vary by maker and by series, and suppliers sometimes change encoder models without changing the motor part number.
That's why swapping in a drive from another series so often ends in a confusing failure. The drive powers up, the motor holds position, then it hunts, or jogs a few degrees when it should be still, or refuses to commutate. Nothing is broken - the drive just doesn't speak the encoder's language. Signal level is a second trap: 5 V into a 24 V input, or the reverse, quietly kills a feedback board.
Before buying anything, read the encoder label, not the motor. If it's absolute, confirm the protocol is on the drive's supported list. Our own absolute-value servo driver line exists partly because this question came up so often that matched pairs beat explaining protocol tables over email.
A worked example: 1.5 kW axis on a gantry router
A customer in Germany sent us this one for a two-metre router - two motors on X, one each on Y and Z.
| Parameter | Motor value |
|---|---|
| Rated power | 1.5 kW |
| Rated speed | 3000 rpm |
| Rated torque | 4.77 N·m |
| Rated voltage | 220 V class, single phase |
| Continuous current | 6.0 A |
| Peak current (2.5 s) | 18 A |
| Encoder | 17-bit absolute, serial |
Three steps. Single-phase 220 V supply, so a 220 V class drive with a ~310 V bus is right, no transformer. Continuous current 6.0 A means the drive must clear 6 A continuously; the cabinet measured 42 °C in summer, so we recommended a unit rated 7.5 A. Peak current 18 A means the drive's peak must clear 18 A; the same frame peaks at 22 A for three seconds. The encoder is 17-bit absolute on a serial protocol, so the drive has to list that protocol - and it does.
That leaves roughly 25 percent margin on continuous current and 22 percent on peak - our target headroom. More and you pay for copper you'll never use; less and you're trusting the cabinet cooling. The 1.5 kW "matching" drive specified originally would have tripped: its peak was 12 A.
Two extras came out of that build. The leads ran 9 metres alongside the spindle cable, so we quoted shielded motor cable with the shield bonded at the drive end only. And because Z is vertical, the drive got an external regen resistor.
Sizing comes down to three numbers you can read off the nameplate and one question most buyers forget to ask: what encoder is on the motor.
Send us the motor nameplate photo and the encoder type through the inquiry page and we will come back with the driver rating, including the current margin we would use. Two minutes on your side.
The questions we actually get asked
Can I run a 3.8 kW motor with a smaller driver?
Only if you accept never using the motor's full output. It needs about 45 A at peak, and a 2 kW frame tops out well below that. If the load doesn't need 3.8 kW, buy a smaller motor instead.
Will a 380 V drive run a 220 V motor if I lower the current limit?
No. The limit doesn't change the bus. You'd be feeding a 220 V winding from a 530 V bus, and the insulation takes the hit. Use a step-down transformer, or use a 220 V motor.
My motor doesn't have an encoder - can I add one?
Usually not worth it. Retrofitting means pulling the rear cover, matching the shaft taper and mounting, and getting commutation alignment right. Sensorless mode exists for fans and pumps, not positioning - if the axis has to hold a position, the encoder isn't optional. New matched servo motors with the encoder already fitted cost less than the labour.
Why does my drive trip at 3000 rpm but run fine at 1500?
That's a voltage ceiling, not a current one. Above rated speed the drive runs out of bus and loses control of the current loop. Check that the bus matches the motor's rated voltage, and that the drive's maximum output frequency suits 3000 rpm.
If you're holding a nameplate photo and working through this list in your head, send it over instead: the plate, what the machine does, and your site voltage. We usually come back with numbers the same day, through the inquiry form or by email.
