Motor Speed torque Curve

[W@ttson]

There are a lot of really smart people on here and especially ones that used to work in custom motor shops. I just wanted to see if anyone can describe the procedure that motor shops do to develop a motors Torque speed curve?

I have seen how Locked Rotor torque is done. I have seen an arm attached to a motor of a certain length and then perpendicularly a stick is set on a scale. Can't remember if full voltage is applied or not but the time allowed was crucial. We were talking less than 5 seconds to take the value down.

How are the rest of the values captured? Motor attached to a dynomometer, turn across the line voltage on and load increases and you watch the torque go up until it starts going down?

Are there any details that can be shared?

Thank you

 

[Jraef]

Generally, yes, a dynamometer is used to confirm the design.

Motor Testing & Sensors, Torque Measurement, Hysteresis Dynamometer, Load Cells | Magtrol

Manufacturer of motor testing dynamometers, hysteresis brakes and clutches for tension and torque control, torque transducers, load-force-weight measurement systems, displacement transducers, rotary transmitters

www.magtrol.com

 

[winnie]

There are quite a few ways to skin that cat, but they all boil down to having a way to vary the speed of the motor and having a way to measure the torque.

The setup that I typically use consists of the motor under test, a separate motor controlled by a VFD, and an inline torque sensor between the two. The VFD lets me set the speed of the 'load motor', and thus control the speed of the whole system. The inline torque sensor pretty much measures how much a steel shaft twists under load, and then transmits the data wirelessly to a data logger.

Other ways to do this:

Use any sort of mechanical load on the motor, but measure the shaft torque indirectly, by measuring the 'reaction torque' on the stationary part of your mechanical load. When a system isn't accelerating, all torque must be balanced on every part of the system. For example if you have a brake, the torque slowing down the rotating part must be balanced by the same torque preventing the brake pads from rotating. With a motor, the shaft torque must be perfectly balanced by torque that prevents the stator from rotating. So instead of having to take a rotating torque measurement you can take a stationary torque measurement on the frame of some bit of your system.

The brake usually converts the input mechanical power into heat. A very common system is a 'water brake' where that heat goes into circulating water, very nice for carrying the heat away. One important practical nuance: if you want to test a motor on the 'back' side of the torque/speed curve (the part of the torque speed curve where torque _drops_ as speed goes down), then you want a mechanical load who's torque drops even faster. Otherwise the mechanical load will simply stall the motor under test.

Acceleration: instead of using a mechanical load to apply torque at a fixed speed, use a flywheel, and measure the acceleration curve. You can even test the motor at greatly reduced voltage and just use the inertia of the rotor itself as your flywheel. Take your 480V motor and instead apply say 24V 60Hz and just measure shaft speed vs time.

The list goes on and on. If you can measure or control speed, and measure or control torque, you can develop a torque/speed curve.

 

[W@ttson]

Generally, yes, a dynamometer is used to confirm the design.

Motor Testing & Sensors, Torque Measurement, Hysteresis Dynamometer, Load Cells | Magtrol

Manufacturer of motor testing dynamometers, hysteresis brakes and clutches for tension and torque control, torque transducers, load-force-weight measurement systems, displacement transducers, rotary transmitters

www.magtrol.com

Thank you

 

[W@ttson]

There are quite a few ways to skin that cat, but they all boil down to having a way to vary the speed of the motor and having a way to measure the torque.

The setup that I typically use consists of the motor under test, a separate motor controlled by a VFD, and an inline torque sensor between the two. The VFD lets me set the speed of the 'load motor', and thus control the speed of the whole system. The inline torque sensor pretty much measures how much a steel shaft twists under load, and then transmits the data wirelessly to a data logger.

Other ways to do this:

Use any sort of mechanical load on the motor, but measure the shaft torque indirectly, by measuring the 'reaction torque' on the stationary part of your mechanical load. When a system isn't accelerating, all torque must be balanced on every part of the system. For example if you have a brake, the torque slowing down the rotating part must be balanced by the same torque preventing the brake pads from rotating. With a motor, the shaft torque must be perfectly balanced by torque that prevents the stator from rotating. So instead of having to take a rotating torque measurement you can take a stationary torque measurement on the frame of some bit of your system.

The brake usually converts the input mechanical power into heat. A very common system is a 'water brake' where that heat goes into circulating water, very nice for carrying the heat away. One important practical nuance: if you want to test a motor on the 'back' side of the torque/speed curve (the part of the torque speed curve where torque _drops_ as speed goes down), then you want a mechanical load who's torque drops even faster. Otherwise the mechanical load will simply stall the motor under test.

Acceleration: instead of using a mechanical load to apply torque at a fixed speed, use a flywheel, and measure the acceleration curve. You can even test the motor at greatly reduced voltage and just use the inertia of the rotor itself as your flywheel. Take your 480V motor and instead apply say 24V 60Hz and just measure shaft speed vs time.

The list goes on and on. If you can measure or control speed, and measure or control torque, you can develop a torque/speed curve.

In the dyno setup, the motor under test is just connected across the line with a breaker and overload right?

You turn that on and then by varying the speed on the dyno motor you force the motor under test to go through its speeds.

How do you choose what torque reference you send the dyno? Or do you not send any torque reference and whatever torque is required to force the motor under test to go the speed you command is the correct torque reference?

If that is the case, how much larger does the dyno need to be as compared to the motor under test?

 

[winnie]

In the dyno setup, the motor under test is just connected across the line with a breaker and overload right?

You would set the motor up for whatever state you want. But if you want a typical torque/speed curve, then yes.

You turn that on and then by varying the speed on the dyno motor you force the motor under test to go through its speeds.

How do you choose what torque reference you send the dyno? Or do you not send any torque reference and whatever torque is required to force the motor under test to go the speed you command is the correct torque reference?

You can do things either way; if you have a load motor VFD with torque control capabilities, then you tell that motor to run at your target torque and then measure the resulting speed. But it is probably more common to tell the load motor to run at a particular speed and measure the resulting torque. You can even just send the load motor a particular drive frequency (open loop control, you aren't even trying to directly control either speed or torque) and let the system settle to whatever speed and torque results from the torque/speed curves of the two motors.

If that is the case, how much larger does the dyno need to be as compared to the motor under test?

Depending on how clever you are the load motor can be the same size as the motor under test; if you you need to run your UUT in heavy overload you will probably want to select a load motor with a rated torque = UUT peak torque.