Measuring dynamic torque

[Caito]

As part of my Master Thesis I'm doing some brushless motor control and other cool stuff.

I made a quick setup to measure static torque and, as expected, it behaves quite linearly with Iq (the perpendicular) current. Here's a quick graph for you to see:


I was wondering if you know of any experiment to measure dynamic torque, a similar graphic but with the motor turning at speed. I have access to a 3d printer and can make pulleys and stuff, I also know the angle (hence speed/accel) of the motor and can control the current, speed, or position.


Thanks!!

[Tim.Wright]

2 common ways of measuring dynamic torque is using a brake (can be a friction, hydraulic or eddy-current) which is reacted by a loadcell. The loadcell force multiplied by its reaction arm length is the torque.

The other way is to hook up the motor to a known inertia and calculate the motor acceleration from your motor encoder. Your motor torque is then the inertia times the acceleration.

The theoretical difference between the two methods is that the first method can calculate a steady state condition wheras the second one can't. The second method will theoretically read a lower value due to inertial reaction of the motor's own rotor inertia. The bigger your "test-inertia" the smaller this error will be.

[Tim.Wright]

P.S. Since you know the current/torque relationship of the motor, cant you just measure the current to give you the torque?

[Caito]

Tim, thanks for your answer!

P.S. Since you know the current/torque relationship of the motor, cant you just measure the current to give you the torque?

I'm actually measuring and controlling current all the time (at 20kHz). What I want to know is how can I relate current to torque. At no speed I know the relation is really linear, but does it hold with a moving motor?

The load cell arrangement is a little too complicated and would take too long to get the parts. But the other alternative should be doable. I will try to get some known inertia

Cheers,
Andrés

[Greg Locock]

A method we used for electric motors is to mount a large pulley on the motor, lay some string on it, put a weight pan on one end and a spring balance on the other. Not often you can do useful engineering with equipment scrounged from the kitchen!

[Tim.Wright]

From what I've been told about bldc motors is that the torque is pretty much always proprtional to the current. I've never confirmed it for myself though. I think a bit of research should get to the bottom of it.

I use a couple of bldc motors at work to simulate steering torque and I know that the steering torque request is simply divded by the torque constant of the motor and sent to the drive.

[Tommy Cookers]

@ Caito

you seem to have convinced yourself that your electric machine is a reliable transducer of static torque when calibrated for this

the usual manufacturer's efficiency plot may suggest how it might or might not be reliably calibrated as a transducer of steady dynamic torque
but such data avoids representing the effects of temperature on magnetic flux density etc which will vary the machine's properties
eg traditionally magnet material was chosen according to the machine's purpose eg whether caterorised as a 'servo motor' or not
eg http://revue.elth.pub.ro/upload/427181art9.pdf

and there may be another question, regarding the transducing of unsteady dynamic torque