Very slowly indeed, like, even if you resist the cogging step as you
turn it, it still seems to work, though my simple transistor decoder
would need a little more gain. But the scope shows a definite pulse,
still, at reduced amplitude. A couple of op-amps would cope nicely, I
think. The quadrature is quite reliable, too, though you want to rely on
the phase relationship between edges and not assume that pulse widths
remain consistent with speed (they don't). But the edge relationships
are quite stable. You can use a monostable to create consistent pulse
widths if you wanted.

> ----------
> From:         John Griessen[SMTP:john_g@CIBOLO.COM]
> Reply To:     pic microcontroller discussion list
> Sent:         Thursday, January 08, 1998 7:00 PM
> To:   PICLIST@MITVMA.MIT.EDU
> Subject:      [semiOT] Stepper as Sensor   was(Best way to read a
> POT..)
>
> I like your idea.  what did you observe the speed of turning needed to
> get a
> definite pulse was?  Will you always reliably get a good pair of
> pulses for
> each movement from one stepper motor step to the next one at a time,
> or do
> you need to keep it moving quickly?  [snip]