ATM Stepper Motor Woes

[cshaw@ghgcorp.com (Chuck Shaw)]

OK, its SOS time from me to ATM list (again)!! I have two questions/problems
I need some help with.....but first a tiny bit of background:

I currently do most of my ccd work (cookbook 245) with a 14.5"/f5 newtonian
on an equitorial platform.  Two of the weight carrying rollers for the
platform (the two under the sector on the north side) are driven by stepper
motors (1.8 deg that are 1/2 stepped).  Each driven roller has its own
dedicated stepper motor and 400:1 gearbox. The drive ratio for the roller to
sectors is 37:1.    The drive rollers are stainless steel and the sectors
are aluminum.  These materials assure minimal/no slippage, and excellent
wear characteristics that do not degrade the tracking.  Having two seperate
driven rollers assures no "torque" being applied to the platform (by only
drving one roller) and having two seperate gearboxes results in their
periodic errors virtually cancelling each other out.  Each motor is driven
with a Motorola SAA1042 chip (the current for the motor actually goes
through the chip, all you need is a timing pulse and some logic inputs for
choosing direction and full or 1/2 step).  I have been generating the timing
pulses with a 555 chip.....Not a very accurate source, but cheap.....

Now for the 1st question/problem:   In the past I have had EXCELLENT
tracking (up to two minutes UNGUIDED!!) However, I have started to get very
erratic behavior on tracking.....the speed does not seem to be consistant,
and whats worse, the two motors actually "seem" to be running at very
slightly different frequencies (there are LED's in parallel to the power
inputs for each SAA1042 that light when the chip pulls current, and on the
outputs to each phase of each motor (a total of 10 LED's, quite a light show
when running at 40hz).  So, how is this speed difference possible? The SAME
output of the one 555 chip is split and goes to the inputs of both
SAA1042's???  I am assuming that either the line from the 555 to one of the
SAA1042's is intermittant (and I will be checking that).  I also suppose
that one of the SAA1042's could be having some type of problem where it does
not "see" all the timing pulses it receives, or is adding extras???  I have
no scope to measure the exact frequencies of signals being generated various
places in the system......  Any other troubleshooting ideas for this would
be appreciated....  If it were a 555 chip problem, it should affect BOTH
SAA1042's simo, and I also see this problem when using a guider (the guider
REALLY has to work a lot and too often cannot smooth out the changes and
looses it).

And now the second question/problem:  I think I should update the circuit to
incorporate a much more accurate timing source.   So, does anyone have a
SIMPLE crystal timing circuit I could integrate into my SAA1042 system?  It
needs to be able to have its frequency changed by varying a resistance
input, since I want to be able to guide in RA using an old Orion Accutrack
(photodiode detector that varies a resistance output from a Vactrol).  I use
this system now and a friend also uses it, quite successfully!  The were not
much good for the very long film exposures in the past, but seem EXCELLENT
for ccd work, and I also own it....so no $'s are involved!

I think I will also change the driver part of the circuit to use a single
SAA1042, and have its outputs trigger transistors for each motor phase
winding for both motors (all 8).  This makes the SAA1042's life easier (no
current draw through it) and eliminates another component.

So,......whose got a simple crystal timing circuit (and the part
numbers/suppliers) that I can use to build a better two motor driver
circuit??????

Thanks!!!!! (sorry this was such a long saga, but thought more info up front
would minimize the Q&A needed for folks to get a handle on how to help)

Chuck






Chuck Shaw
CShaw@ghgcorp.com (home e-mail)
CShaw@gp801.jsc.nasa.gov (work e-mail)
http://www.ghgcorp.com/cshaw  (humble start at a home page)