First, I'll answer your posting, Paul, and then I'll show in detail how
I once built a very good and safe coil winding machine from a scrapped
dot-matrix printer stepper motor, that could be revived as a PIC project
imho.

On Sun, 26 Jul 1998, Paul B. Webster wrote:

> Peter L. Peres wrote:
>
> > I deduce from the use of the word 'motor' in the original posting that
> > the speed of the unit might slightly exceed the capability of micro-
> > switches.
>
>   I wouldn4t have thought the winding bobbin was directly driven by the
> motor, so it shouldn4t be *that* fast.  My presumption was that the
> lifetime of a *reed* switch would cover the number of turns on a
> reasonable number of custom-made transformers or whatever is being
> considered.

Hmm. Assuming 1E6 life, and 1000 turns total per unit to be done, the
switches (being two of them) would be about dying after 700 units done.
Ok, that's a lot for amateur use, but then if one has very few to do, then
one can wind by hand and count ;) Also, I suppose that the user would go
on hearing click-click click-click in his head for a long time after
stopping the unit.

>   Actually, I made further assumptions.  The reference to the need to
> count back suggested that the number of turns was critical.  This
> further implies the number is relatively low and not particularly fast.
> The tone of the posting suggested a non-commercial application (i.e.,
> you don4t cut costs on sensors on factory equipment ;-).

The one sound use for counting back on windings is when doing accurate
balanced transformers and hybrids for high-end valved audio. Nobody winds
600 ohm bal/unbal transformers by hand anymore. Also, metal detector
loops, but those are done by hand on a jig due to size ;)

- snip -
>   I think of coil-winding machines in the same fashion as lathes (comes
> from using one for the other) in which environment optical sensors are
> a no-no!

imho, 2 big mistakes here:

1. Lathes are NOT coil-winding machines, unless the shaft can be equipped
with a torque limiter. This is not a matter for joking or experimenting,
take it very seriously. This includes small table lathes, unless
hand-driven. Foot-driven ones are dangerous for coil winding. Also, no
lathe should EVER be operated with an improvised jig to hold ANYTHING in
it, unless the jig is correctly machined and tested for strength.
Human-operated lathes are among the most dangerous work machines in use.
They have all the work piece and the tool exposed to touch and in case of
accidental fragmentation the probability that the operator is hurt is in
the 30-40% range. If, additionally, the operator can be pulled by a tether
(such as, winding wire) to the moving parts, things can get even worse.
Very, very, very bad.

2. Optical sensors are used a lot in lathes, both reflective and
transmissive, but only in locations where they are capsulated and
protected from debris. Hall sensors are 'out' near the parts and work head
in most fine mechanics part processing machines because many times pieces
and materials get to be processed that cannot be exposed to magnetic
fields of any kind, or are themselves strongly magnetic, and the sensing
tolerance of hall sensors is not good enough. Instead differential HF
proximity (Q) sensors are used, and more recently, low power laser-based
sensors that can be placed pretty far from the work piece and head (in CNC
only). Precision position sensors with great travel are almost always
optical (quadrature or Gray using Moire effect, precision-geared to the
moving part), and there is no replacement in sight. Most optical sensors
use infrared, except high precision, which use visible but are totally
capsulated.

  Last, I'd like to say that I once made a coil winding machine using a
large stepper motor from an old printer. It worked out to be very good,
durable, simple to build, and could be a PIC project. The motor was driven
by a 555 timer, a CD4017 decoded decade counter used to count to 8
(obtained by connecting the 9th output to the reset pin of the 4017), and
4 Darlington transistors. 8 1N4148 diodes turned the 8-phase output from
the 4017 into 4-phase half-stepping drive (chosen for smoother torque) for
the motor. Power was 12 Vdc / 0.8 A or so, and the machine was very simple
mechanically, yet powerful enough to wind 1 mm wire on a 25 mm core. The
speed range was small but the stepper motor gave constant speed (unlike
the runaway of an uncontrolled DC motor), solid torque, including holding
torque, and is an inherent torque limiter, all this without using any kind
of gears.

  The speed control was a potentiometer in the 555 circuit. Its switch
would cut out the output of the 555 completely by opening the charge
cap. resistor circuit completely. This gave stopped motor+holding torque,
without glitches at switching.

  To reduce the torque, I simply turned the bench PSU voltage that powered
this unit down to 9 or 6 V. The circuit continued to work well, with some
speed change from the 555, and lower max. speed.

  The motor was biphase (6 wires, 2 center-tapped coils), and I used BDX53
Darlingtons and 1N4007 kickback damper diodes for each phase coil.

  Mechanically, I used a piece of 1mm mild steel sheet to hold the motor
in a small desk vise. The sheet was cut in a rectangular piece as wide as
the motor and three times longer. The sheet was screwed to the motor using
its mount holes at one end of the sheet and the protruding 'tail' of the
sheet was twisted 90 degrees to have the axle parallel with the desk edge
when the lower end was held in the vise. The motor would rest on the tail
end of the vise by bending the sheet towards the desk. The electronics
were on a bread-board and connected to the bench PSU and to the motor, the
latter using the original connector from the printer.

  Jigs to hold coils were affixed with 2 or 4 machine screws to the
cylindrical aluminium hub that was originally on the motor shaft. I just
bored and threaded the 4 holes symmetrically placed on a circle at the
front end of the hub, for 3 mm machine screws. I used to make jigs out of
rigid PVC tubing, cardboard and Balsa wood. I did not use a counter, there
was a paint dot on the hub ;) The shaft attained 4-5 rot/sec (300 rpm)
which is easy to count ;).

  This could be a PIC project too. A PIC and a ULN 2803 would be all
that's needed. And it's MUCH safer than a misused electric drill or lathe.

hope this helps, and sorry for the long posting,

        Peter