## Measuring Stepper Motor Torque

9 posts / 0 new
Author
Message

I need to measure the torque of some small stepper motors (NEMA 17 and 23) when run from one of my boards. I don't need absolute values so the 'string and weight around a pulley' is over the top; I need a more compact solution to do relative measurements and characterise the drive waveforms.

I was thing about coupling a stepper motor to a DC motor and feeding the DC motor to something like a constant current load. Memories of DC dynamos on my bike suggest this will work. However, when reading up on Dynamometers it seems like the actually feed the motor from a drive rather than just absorb the generated current.

Instead of a DC motor I also thought about using another stepper as the generator.

Thoughts?

"This forum helps those that help themselves."

"How have you proved that your chip is running at xxMHz?" - Me

"If you think you need floating point to solve the problem then you don't understand the problem. If you really do need floating point then you have a problem you do not understand." - Heater's ex-boss

I'd just use the load on the DC motor--usually works decent (though perhaps  not highly calibrated, other than "more" load means "more" torque) In a DC motor, neglecting other errors, torque is proportional to current.

If you drive the DC motor, then you might have issues if your stepper cannot sink power (it can certainly source it).  Note that with a current load, the DC motor will have a voltage developed as it acts like a generator  (which you could measure with a DVM), so from that perspective it "appears" the motor is being driven  (say with 8v, 12v or whatever the DVM measures)  Of course it is acting as a sink, not a source.

Beyond this, the steppers can have a complex torque pattern, and resonances, etc.  So measuring "torque" might not be a simple as in a DC rotating motor.

see here...see graph on page 63, also do a document search for torque & it has many tidbits of useful info

When in the dark remember-the future looks brighter than ever.

Couldn't you just look up the specs of the steppers?

Unless there are no markings on the motors to begin with.....THEN, thats not gonna work

JIm

If you want a career with a known path - become an undertaker. Dead people don't sue! - Kartman

Atmel Studio6.2/AS7, DipTrace, Quartus, MPLAB user

The actual torque relies heavily on the driver circuit. A simple driver that just switches the windings won't give a much torque as a driver that uses high voltage and/or current control methods. Also, measuring torque is a good method of determining if the driver circuit is working properly.

How accurate do you want to be?

What's wrong with  the string on pulley, why would that not be compact enough?

I actually think this is quite a neat solution, and it can be made pretty compact.

How many motors do want to measure?

Stationary measurement, or during rotation?

Max stepper motor torque depends on a lot of var's: Speed, intertia, dampening.

Have you thought about "reversing" the measurement.

With this I mean not measuring the torque on the output shaft, but mounting the motor in a way that it can rotate.

Then you can measure the torque by the force the motor exerts on the mounting frame.

The simplest way to do this is to put a ball bearing on the motor shaft and hold the outside ring of the bearing firmly.

Then put a screw through one of the mounting holes and you can attatch someting to that screw to measure the force.

For more accuracy, add a longer lever, integrate a loadcell from a cheap kitchen scale (with an AVR and a HX711).

For more compactness you can  mount the motor in the center of a lazy susan or use a "thin section bearing" for better quality.

You can probably als make quite a nice dynamometer out of an old HDD.

Mount the magnets (movable) near the platters to generate an adjustable amount of eddy currents.

I hope those platters are made out of aluminimum. Copper would work better.

Some HDD's have glass or ceramic platters and then it won't work.

Paul van der Hoeven.
Bunch of old projects with AVR's:
http://www.hoevendesign.com

Last Edited: Tue. Sep 5, 2017 - 07:57 AM

avrcandies wrote:

I'd just use the load on the DC motor--usually works decent (though perhaps  not highly calibrated, other than "more" load means "more" torque) In a DC motor, neglecting other errors, torque is proportional to current.

Thanks. KISS is definitely the way to go I think.

jgmdesign wrote:

Couldn't you just look up the specs of the steppers?

I know what the motor should do; it's my driver circuit and software that I'm interested in.

Kartman wrote:

The actual torque relies heavily on the driver circuit. A simple driver that just switches the windings won't give a much torque as a driver that uses high voltage and/or current control methods. Also, measuring torque is a good method of determining if the driver circuit is working properly.

I'm using a 'proper' stepper controller chip with all the bells and whistles.

Paulvdh wrote:

How accurate do you want to be?

No accuracy at all; just repeatability.

Paulvdh wrote:

What's wrong with  the string on pulley, why would that not be compact enough? I actually think this is quite a neat solution, and it can be made pretty compact. How many motors do want to measure?

Workshop space is an issue. This will be part of a test jig for an ongoing project so will likely spend ost of it's life on a bench somewhere.

Paulvdh wrote:

Have you thought about "reversing" the measurement. With this I mean not measuring the torque on the output shaft, but mounting the motor in a way that it can rotate. Then you can measure the torque by the force the motor exerts on the mounting frame...

I hadn't but that's quite a neat idea. My only reservation is that I think I also need to keep an eye on dynamic performance. I use various velocity/acceleration profiles to drive the motor so probably ought to check the drive torque under acceleration.

Thanks for the ideas people.

"This forum helps those that help themselves."

"How have you proved that your chip is running at xxMHz?" - Me

"If you think you need floating point to solve the problem then you don't understand the problem. If you really do need floating point then you have a problem you do not understand." - Heater's ex-boss

Brian Fairchild wrote:
Workshop space is an issue.

I see at least 2 options here.

Build something quick and easy and throw it away after your're finished

(Take some pictures notes in case you have to rebuild)

Or put some care into it and build something small and compact.

Most kitchen scales have pretty nice loadcells in them. Something like these:

https://www.aliexpress.com/whole...

With a decent load cell amplifier (which costs more than the scale :) you can meaure the force and put it on the scrren of a DSO or dedicated logger (ADC of AVR?)

An "HX711" (About USD 3 from Ali / ebay) also work pretty well. I got 14 bit of usable accuracy without averaging. But this amplifier/adc maxes out at 80 samples/s.

Brian Fairchild wrote:
so probably ought to check the drive torque under acceleration.

Good Idea, Excellent even.

With a (big) flywheel on the stepper motor it won't have to deliver any torque at a steady rpm, but the amount of torque delivered into the flywheel is easily calculated from the accelleration parameters you put into the stepper motor.

Look for "David Austin" for a good way to calculate constant accelleration parameters for a stepper motor

This will be very accurate & reapeatable because there is no unknown external friction.

Gotcha's:

Need to correct for inertia of the motor itself.

Motor parameters are always of a full system. Anything connected to the motor shaft influences resonant frequencies etc.

With a big flywheel the resonances are probably swamped by the inertia of the wheel but they can stick up their ugly 7 heads in a "real" aplication.

Use 2 (or more) different flywheels. Double check yourself.

If you really want to go fancy and measure resonances and such on the motor shaft you can easily do that with a "continuous rotation" potentiometer with "conductive plastic".

Paul van der Hoeven.
Bunch of old projects with AVR's:
http://www.hoevendesign.com

Isn't there on the market a torque-o-meter? I have (or had) a spanner thingy that measured the amount of torque applied to the head of the screw for some project a long time ago.

John Samperi

Ampertronics Pty. Ltd.

www.ampertronics.com.au

* Electronic Design * Custom Products * Contract Assembly