simulate torsion spring with motors???

[TheTick]

I am looking into ways to simulate torsion spring behavior using electric motors.

Basically, I need a motor (or system of motors, etc.) that can apply torque that varies linearly (at least approximately) with respect to the angular position of the motor shaft through about 90 degrees of motion.

If the ladies don't find you handsome, they should at least find you handy.

 

[Skogsgurra]

TheTick,

That is fairly easy. Get a configurable drive (ABB and Siemens will work and probably many other)and a motor with an encoder and counter for position feedback (or even simpler: a potentiometer). Configure the speed controller to do only proportional gain and set the gain quite low. Now, when you try to move the motor away from its rest position, the angular difference will be amplified by the P-configured speed controller and feed the torque (current) controller with a signal that is proportional to the angle. The result: a restoring torque that is proportional to the angle. Just like a torsional spring. The spring constant can be set with the P gain.

You can use a DC drive or a vector controlled AC driva. The scalar types will not work very well.

Of course you will need to think about cooling and such things since the motor does not turn and cool itself. Also remember that a DC motor has a commutator that can get distorted by uneven heating if allowed to stay stationary with high current.

But the rest is classic drive technology.

 

[TheTick]

The size of the system would be fairly small (small enough to carry, anyway). We want a way simulate a wide range of spring curves and preloads moving masses less than 1kg.

I know the components and such are probably obtainable off-the-shelf, but I am not familiar with motion control tecjnology. Your reply will help me figure out where to start looking. Thanks.



If the ladies don't find you handsome, they should at least find you handy.

 

[jbartos]

Suggestion: Visit

http://mechanical-components.global...ea?Comp=271&state=all&SortBy=0&RegEvent=login

(requires the registration)
etc. for more info

 

[Skogsgurra]

The Tick,

The thing gets a lot more complicated if you want to simulate dynamically as well (which I think that I can read between the lines in your recent posting). You should use a small servo system for this and you really have to think about the equivalent mass. It can be fairly well simulated as long as your movements are band-width limited. But do not expect to get true simulation if you require more bandwidth than 300 - 500 rad/sec.