On Sun, 8 Feb 2004 12:27:12 -0500, Josh Koffman =20 wrote: > My question is this. I had an idea for a project that would use a > stepper for positioning. So unlike the discussions above, the primary > function of the stepper would be positioning, not sensing. This means I > would have to drive it with enough current to get it to move a load, an= d > I would likely have to keep one winding energized for holding torque. Assuming, that is, that you don't want to half step or microstep the moto= r=20 in which case the rotor could halt at a position where both windings are=20 energised. > Given that circumstance, is it still possible to be able to sense > position using just the stepper as an encoder? Well you could certainly measure the emf in the other winding and count=20 the pulses to get an idea of how far it had moved, and if you were able t= o=20 sense the polarity of the induced voltage then you would know which=20 direction it had moved. However the problem would come when someone moved= =20 it whilst the motor itself was moving as the changing voltage in the=20 driven winding would induce a similar change in the sense winding - added= =20 to the fact that you would have some inteligent driver circuitry that=20 knows which winding it is driving and which one is the sense winding.=20 Although possible i don't think it would be particularly reliable as ther= e=20 would be a good chance of you missing some pulses. I have done a project not entirely dissimlar to this which was to use a=20 small stepper motor as a cheap laser graphics galvo - one winding was=20 driven with a constant current to make it centre around one particular=20 rotor position, and because it was current driven, velocity feedback was=20 availiable by monitoring the voltage across this winding. The other=20 winding had a variable drive signal applied to it. An error signal was=20 derived from the difference between the two and PID control done in the=20 usual manner. It worked surprisingly well for something costing in the=20 region of =A320 compared with =A31000 for pro systems. If both coils in your situation were current driven you could very=20 probably derive exactly the same feedback information from it as if were=20 unloaded with a few OpAmps. Regards Alex Rice -- http://www.piclist.com hint: The PICList is archived three different ways. See http://www.piclist.com/#archives for details.