Why does Kt vary with torque?

[SPIGUY1]

I'm using a 24V 3 phase BLDC servo motor with trapezoidal commutation to provide force feedback to a steering wheel. Torque should be greatest when the angle between the steering wheel centerline and the machine centerline is maximum (cranking the wheel for a hard turn). Since torque and current are proportional, shouldn't the torque constant not vary unless it's due to coil heating or current limiting? I'm seeing the Kt at max. when the wheel is first being turned.

 

[sreid]

The torque should simply be

Torque = Kt x Current

It should be (approximately) constant vs. angle.

As long as you are controlling the current, the only reason torque will change with temperature is that Neo RE magnents lose flux with increaing temperature (it recovers when they cool).

Is your Trap commutation actually working?

 

[Clyde38]

I think that you should look at “Kt” as a design goal constant. As mentioned, temperature of any magnet will cause variation in both magnet performance and winding resistance. What I believe you are really interested in is “torque per amp”. This is not a constant. Temperature as well as things such as armature reaction, electrical time constant, pwm, phase advance etc. will affect the actual torque available for the current applied.

 

[Clyde38]

As mentioned, temperature of any magnet will cause variation in both magnet performance and winding resistance.

Sorry, I meant to say that any variation in temperature will affect both magnet performance and winding resistance.

 

[blacksea]

First, with sinewave motor windings and trap type of commutation you have 13% of moment changes in the same Hall states with constant current(well-known torque ripple).
Second, torque constant may have tolerance +/-10% from Data Sheet.
Third, pay attention how motor manufacturer define torque constant - for trap or sine commutation?

And finally why don't use standard torque sensor directly? - see Futek for example.

 

[SPIGUY1]

This motor uses a PWM signal with 50% as the neutral point at which the machine and steering wheel are aligned (thus no or minuimum torque).

 

[cswilson]

A couple of things to look at:

1. The PWM signal modulates voltage -- the current is only an indirect result of the modulated voltage. Are your values for torque/current based on actual current measurements, or are you assuming the current is proportional to the modulated voltage?

2. A lot of motors can be driven into magnetic saturation at high current levels. A few years ago, I ran into a brushless DC motor whose Kt at 90% of rated (instantaneous) current limit was about 20% of the Kt at low current levels. This caused all sorts of problems in the application, because the users were expecting a constant Kt.

Curt Wilson
Delta Tau Data Systems