Energized motor driven in opposite direction?

Question

This question is regarding brushed permanent magnet DC motors.

Suppose that the motor's output shaft is connected to a torsional spring which was at its unstretched state when the motor was OFF. After turning the motor on, the motor tries to rotate in forward direction while the spring constantly opposes the motor's motion. The opposing torque of the torsional spring will increase with the total angle turned by the motor's shaft and at some point it will balance out motor's torque causing the motor to stall.

Now the motor is in stall condition. I suddenly decrease the input DC voltage to my PMDC motor by decreasing the PWM duty cycle which should decrease the average input voltage to the motor. With the decrease in input voltage I expect the motor's stall torque to decrease proportionaly. Now the motor's torque is decreased, but the shaft is instantaneously at the same position as before - which means that the spring's torque is the same. So essentially the spring's torque has become greater than the motor's torque.

Will the spring be able to drive the motor in the opposite direction?

I was actually experimenting with an electronic throttle body. Using an Arduino UNO, I first gave a duty cycle of say 16% which opened the throttle to some extent. Next I suddenly decrease the duty cycle to say 12% but there was no change in the position of the throttle. The throttle does not go back no matter how much I decrease the duty cycle. When I decrease the duty cycle to very low value then after a certain threshold the throttle valve just closes shut. This assymetric behavior of throttle is very confusing to me.

I give 12% duty cycle then the throttle opens say 45% and when I increase the input PWM duty cycle to 16% duty cycle then the throttle opens to, say, 80%. When I go in the opposite direction (when I decrease the PWM duty cycle) then the throttle does not respond at all.

Has this behavior got something to do with electromagnetic inertia of the motor? Maybe when the motor is energized it takes a huge amount of torque to drive the motor is opposite direction. Am I correct regarding this? Are the concepts of "holding", "cogging" and "detent" torques relevant here?

EDIT : No overheating issues have been observed in the motor yet. Also the electronic throttle body is new and unused, so there is no point of carbon buildup or other non-idealities associated with long term use of throttle.

EDIT 2 : I am aware of closed loop control of throttles. My doubt is regarding the asymetrical behavior of throttle - why isn't the spring able to drive back the motor when PWM duty cycle is reduced?

Answer 1

Will the spring be able to drive the motor in the opposite direction?

Yes it will. Consider the scenario where the voltage is removed completely. Assuming no frictional or stiction, the motor shaft will return to the neutral position.

In any case, friction can stop the motor returning to the original position and stiction can prevent movement back to the original position when the motor voltage is reduced. Of course, your motor will have some element of cogging that can cause it to not glide back perfectly.

Answer 2

Such a system you are describing above is commonly referred to as a Series Elastic Actuator (SEA).

The motors are not perfect torque sources. Motors with permanent magnets have cogging torque, which is the interaction of the permanent magnets in the motor with the slotted electrical steel on the stator. This interaction creates a passive torque that resists movement from some equilibrium position. If you want a motor with no cogging torque, look up slotless motors.

Also the magnets interaction with the electrical steel creates eddy currents when the magnets are rotated. The eddy currents act as a torsional damper and slow any motion.

Answer 3

There are a lot of mechanical and magnetic effects present in motor-spring interaction. Besides throttle bodies, electrically-controlled hydraulic valves also fall into this category.

For this very reason, throttle bodies typically contain a sensor, possibly a potentiometer or LVDT (linear variable differential transformer) to create a closed-loop control strategy. The motor is energized based on setpoint vs. pot; the error of which drives the motor and forces it to match the setpoint.

Such control may utilize P.I.D. (Proportional, Integral, and Derivative) maths or logic to dictate how "wildly" the motor can respond. Setting the PID values properly allows the control to respond quickly, yet suppress undesirable artifacts like oscillation and overshoot.