Stepper motor feedback

Hello all. I have a problem, and I do not know enough about stepper motors to solve it (nor can I find it on Google). Is it possible to take a unpowered stepper motor, and read how many steps the motor has been spun? What I mean is can I connect some form of device to the wires of a stepper motor, manually rotate the motor, and have the device tell me how many steps the motor took. I would like to do this with a Arduino at some point, however obviously I need to know if its possible first. Does anyone know if this is possible? If so do you have a resources or information as to how it would be done?

The usual way to get feedback is to add a rotary encoder to the motor.

I guess if you are expert with electronics you could detect the voltage pulses produced in the coils of the motor as the magnets pass them.

Why do you need to do this?
It is not usual to design a stepper application that needs to know what happens to the motor when it is powered off.
In most situations a limit switch is used to detect when the motor (or whatever it is moving) gets to the "home" position. The Arduino then bases all its step counting from that position.

...R

I need to measure a somewhat precise number of steps (in the 10 thousand range) over a somewhat short range. A rotary encoder is probably not sensitive enough.

setra:
A rotary encoder is probably not sensitive enough.

Why not?
You can get them with more steps than the stepper motor.

...R

The output from a stepper motor has a voltage proportional to the speed, so low speed rotation
will fall into the noise background. High speed rotation will generate high voltages that
you'll need protection circuitry like a zener-resistor divider to protect against.

However if rotation speed is within a reasonable range you ought to be able to take the
two phases and put them through 2 comparators to get clean quadrature signal.

So what your saying is that I wont be able to measure a noticeable spike of voltage every time the motor takes a step (the un-powered motor that is)?

The motor will generate roughly sinusoidal quadrature signal, with ampltide proportional
to the angular velocity - if you have a 'scope this is easy to observe.

I need to measure a somewhat precise number of steps (in the 10 thousand range) over a somewhat short range. A rotary encoder is probably not sensitive enough.

Rotary encoders are designed for exactly this task, and they are accurate to better than +/- 1 count.

I haven't tested this but it should be possible to measure the inductance of the motor winding. This should change depending on rotor position because of the reluctance of the magnetic path changing.

setra:
I need to measure a somewhat precise number of steps (in the 10 thousand range) over a somewhat short range. A rotary encoder is probably not sensitive enough.

Rotary encoders are available with many millions of counts per revolution at upto
30,000rpm or more, so there will be a rotary encoder sensitive enough. Cheap
enough is another matter...

What problem are you actually trying to solve BTW?

setra:
I need to measure a somewhat precise number of steps (in the 10 thousand range) over a somewhat short range. A rotary encoder is probably not sensitive enough.

It just crossed my mind to make the point that with a 200step/rev motor, 10,000 steps would be 50 revolutions. That does not seem to be a short range.

...R

This might do the job