torque sensing with stepper driver

[subsearobot]

Hi all! I hope I am posting in the right place- please let me know if not. and please forgive- I'm a mech E so caviots all around!

We have some consultants designing a test fixture for us. there is a stepper motor driving a mechanism rotationally for cycle testing. We are interested in monitoring the degradation (if it exists) of the frictional interface. I expect this to show itself as increased torque resistance of the joint at a constant rotational speed.

I asked for precision torque tests, either insitu or even an intermittant measurement outside of this test fixture. I'd like to see in detail, when the torque increases by 20%, so I think that minimum resolution should be better than 10% on initial torque- I'd prefer to see resolution at 1/100 of initial.

The consultant proposes that we can measure torque with the motor controller by monitoring the drive current. I realize that stall-detect is feasible, but is torque sensing? Please comment on whether this sounds feasable. I think we need a precision torque transducer, but I'd prefer the simplest fixture that gives me the informatino that I seek....

Can a motor controller act as torque transducer with a reasonable accuracy?

thanks!

 

[Skogsgurra]

There is a better indirect way to do that.

If you have the same set up and want to see changes it is quite easy to measure the power (DC) fed to the drive. Since active power is the sum of mechanical load plus losses, it is usually fairly easy to add a "calibration run" and then compare to what you get when running at different speeds during the time the experiment is running.

DC amps times DC voltage equals active power and that makes it possible to forget about any worries abut power factor or anything going on between drive and stepper. Especially with the very modest resolution/accuracy you seem to need.

Keeping DC input voltage constant makes it even easier - all you need is to measure DC current.

Gunnar Englund

http://www.gke.org

--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.

 

[MikeHalloran]

I expect this to show itself as increased torque resistance of the joint at a constant rotational speed.

Is the DUT ordinarily driven by a stepper?

Steppers, IMHO, will provide 'constant' rotational speed at basically one speed. ... where the drive is resonating with the remainder of the system to turn on the stepper coils at the exact instant where some torque is needed.

At any other speed, steppers basically beat the crap out of whatever they are connected to.



Mike Halloran
Stratford, CT, USA

 

[Skogsgurra]

I think (or hope) that the motor is micro-stepped. That makes it run smoother than a DC motor. If it is not - make it.

Gunnar Englund

http://www.gke.org

--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.

 

[cswilson]

Stepper motors is are AC synchronous motors. When run open loop (as they almost always are), they use a constant current magnitude. As Skogs pointed out, you likely have a microstepping drive, which provides waveforms to the phases that are at least approximately sinusoidal, making the performance very much like a classic AC synchronous motor.

In open-loop control of an AC synchronous motor, the torque produced for waveforms of a given current magnitude is proportional to the sine of the angle between the rotor and stator magnetic field orientations. (For the small angles where typical operation is, this can be well approximated as proportional to the angle itself.

If the load torque increases, the rotor angle falls further behind the stator angle, and torque is automatically increased. Note that there is NO increase in the current magnitude when this happens -- the change is in the phase difference.

All of the above is an explanation as to why it appears to be a fundamentally flawed approach to use current magnitude as a proxy for torque in open-loop stepper control.

Contrast this with closed-loop control of a brushless servo motor (which is also an AC synchronous motor) using a rotor angle sensor such as an encoder or resolver. In this control mode, the sensor info is used to keep the angle difference at 90 degrees always for maximum torque per unit current. The servo loop automatically adjusts the magnitude of these waveforms to match the load torque. In this mode of operation, you often can use the current magnitude as a proxy for the load torque. We have many customers doing it -- I refer to it as the "poor man's" torque sensing.

It is possible to operate a stepper motor with an attached rotor encoder in closed loop mode, running it as a high-pole count (typically 100 poles) servo motor. Again, we have many customers doing this. But you must make sure the controller and drive are doing real servo control.

Curt Wilson
Omron Delta Tau

 

[Skogsgurra]

We don't have to use an encoder. If you measure MOTOR current - yes. But all that conservation of energy stuff makes measuring Power to the drive a good torque indicator. And if the DC supply is stabilized, it is easy to calculate actual torque from power (i.e. current drawn by the drive) minus known losses divided into known speed. You can get below a few percent uncertainty that way.

It should work well if your error budget is 5% and especially if you need to track different friction in a mechanical contraption at different speeds and over time.

Gunnar Englund

http://www.gke.org

--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.

 

[cswilson]

Yes, Gunnar, you are correct. I was focused on the current from the drive to the motor, which is what our servo customers use as a torque proxy. This would not work for open-loop stepper control.

But the input current to the drive would vary deterministically with the torque. it is not directly proportional, due to losses, particularly the resistive losses in the stepper motor (which can vary significantly with temperature). But with a little effort, these can be understood well enough to subtract out pretty accurately.

Curt Wilson
Omron Delta Tau

 

[benta]

Ummm... the stepper motor drives I've worked with use current drive for the coils.

 

[cswilson]

benta -- Please clarify your point.

 

[sreid]

I agree with all that cswilson says but I'd suggest using a Slotless Brushless Motor with No shaft Seals. Current = Torque.

https://www.machinedesign.com/technologies/slotless-motors-smooth-way