Torque control servo or stepper motor

Just bear in mind the caveats I've given about temperature rise and current limiting.

I've forgotten where it is, now, but there's a church clock in England that is driven by a stalled motor connected to one of the gears in the gear train. It's actually a mains-powered drill that has been repurposed. It is driven off a nominally 12V DC supply, and the torque it provides can be adjusted by altering the current in it, in order to fine tune the timekeeping.

Thank you guys for all this practical information! My knowledge increased significantly since yesterday!

My evolved thoughts about this project right now:

  1. I can't use a load sensor to measure the load directly. As I mentioned I need to keep this as light and simple as possible and the force applied to the item(similar to a stiff spring) will be about 10-12kN. A load cell this size would be too big and too heavy. I could put somewhere a strain gauge and calibrate it but this is too complicated.

  2. I don't think that the motor will work in a stalled mode while the spring is compressed by a certain load. With my lead screw with 2mm lead, it should work more like a vice and I don't think that by pushing this kind of lead screw you can make it spin very easily.

  3. Let say I will control my servo motor normally with a step/dir signal. With full current the motor will go to its maximum until it gets a position error, but what would happen if I separately would just change the max possible current to the servo drive by a potentiometer not connected to the control loop at all. Would this work? A scaled potentiometer to control torque would be sufficient for this application. Just a thought.

  4. I'm leaving the brushed motor for now. It would be nice to control the initial exact position of the lead screw.

btw this is the best :sweat_smile: If not now, I'm sure I will use this knowledge at some point in life. Thanks!

Here is a simple schematic. As you can see my drawing capabilities are limited :blush:

I can't imagine reading the current would be workable. Maybe I'm wrong, but when I've had a meter on a running stepper the current is all over the place. Admittedly I measured only one wire, maybe the sum of the four is more stable?

Do you know the N/mm rate of the things you are compressing? If so, maybe a distance sensor would be appropriate?

Steppers cannot control or measure torque, so no chance.
A servomotor that can be selected for torque-control mode is the obvious candidate here. You'd need to ensure the actual motor is sized correctly for continuous duty at that torque without moving (motors rely to fan cooling for maximum power handling, if the fan isn't moving the continuous power rating is less).

Industrial servomotors and their controllers are very expensive but very capable with many modes and programmable features.

Cheaper servomotors (such as used for CNC) might be more limited in the modes they support.

Rolling your own servomotor takes a motor, an encoder or tachometer, and a current sensor for torque sensing. And various control loops using PID controllers and a basic understanding of control theory / feedback.

1 Like

OK this is interesting. From what you are saying the only problem with cheap servo is the drive that will not allow you to control torque? Maybe buying a cheap motor and making a diy torque drive would be an option but I think it is outside my capabilities...

I need to do some more research on servo drives but in the meantime I wanted to buy a Arduino board and play with it already. I didn't know that there are so many versions... Which version would you recommend for such application (max two servo drives, some switches and potentiometer, nothing else)?

By the way you can make a crude torque motor by current limiting a standard motor - that might be all you need here - of course you have to check the motor doesn't overheat when stalled at the current you limit it too.

You haven't read the rest of the thread, have you? Best to do that next time, before offering a suggestion that's already been discussed at length.

I like to provide the benefit of my experience, in reply to the OP's questions/replies. I try to focus on that mostly.

Its probably more productive I do that than spend my reading other replies that I may or may not agree with - here the OP wasn't confident about rolling their own servo loop, so I suggested a simpler approach in response to that.

See replies #14 and #18. No servo loop required.

It seems daft to wade straight into a discussion without checking out what's been said already.