Difference between throttle-torque-control, and throttle-speed-control type system.

Good day everyone,

I'm hoping that I could get a clear and concise explanation of each of these types of controller operation using a throttle.

Specifics on the how, why, and what is need or measured, on each system. Advantages/disadvantages for each. And a bonus if you have used both, with an explanation on why you like one over the other.

Thank you for your time.

Regards,
T.C.

A throttle-speed-controller sets the duty cycle of the PWM directly, the resulting current is only limited, not controlled.
This has the effect, that the wheel speed is proportional to the throttle position (without load).

A throttle-torque-controller controls the motor current actively proportional to the throttle position. So the wheel will spin up to maximum speed with no load with nearly closed throttle, as there is almost no torque necessary to accelerate the wheel without load.
The torque is directly proportional to the motor current. Remember, the motor current is different from the battery current:

So you need at least a battery current shunt for both strategies. FOC controllers have shunts in all three phases normally, see chapter 5 of this STM paper:

I think there is no better or worse, it's just different.

regards
stancecoke

Thank you for the reply stancecoke,

Would you be able to take this a step further and explain under real world riding conditions how each reacts?

I.E. With the throttle set in a static position, how would each react to a rise or lowering of the grade upon which your traversing?

TommyCat said:

With the throttle set in a static position, how would each react to a rise or lowering of the grade upon which your traversing?

Click to expand...

With the speed throttle, the speed will increase until the voltage from the controller (duty cycle * battery voltage) is almost the same as the BEMF. If you increase the load, the wheel will spin a little slower, so the difference between the BEMF and the voltage from the controller will get higher and the current will increase, so you get more power output. This is a kind of "self controlling" of the speed, but no real, active speed control.

With the torque control, the speed will decrease with higher load, as it is controlled to the setpoint. On a higher grade you will have to open the throttle a little more, to keep the speed.

Thanks for the added details… the more the better!

stancecoke said:

but no real, active speed control.

Click to expand...

In your opinion, how could this best be implemented?

TommyCat said:

In your opinion, how could this best be implemented?

Click to expand...

The VESC, MESC, EBiCS projects are nesting two control loops. The base is the current control loop, that is running in every mode. On top they set a second control loop, that increases the setpoint for the current, if the motor is too slow, and decreases the setpoint for the current, if the motor is too fast.
It works, but is a bit tricky to tune, to avoid oszillation.

stancecoke said:

So you need at least a battery current shunt for both strategies. FOC controllers have shunts in all three phases normally, see chapter 5 of this STM paper:

Click to expand...

lol, ouch this made my head hurt. But very informative if you're at that level of knowledge.

stancecoke said:

On top they set a second control loop, that increases the setpoint for the current, if the motor is too slow, and decreases the setpoint for the current, if the motor is too fast.

Click to expand...

This was along my line of though also... but with perhaps just an easy digital output external speed sensor with multiple magnets for the input?

stancecoke said:

It works, but is a bit tricky to tune, to avoid oszillation.

Click to expand...

Roger the "hunting" issue.

TommyCat said:

but with perhaps just an easy digital output external speed sensor with multiple magnets for the input?

Click to expand...

That's not necessary, the controller knows the recent rotor speed exactly, as it has to power the coils according to the rotor angle.

regards
stancecoke