[R-390] Adjusting Linearity Stack Collins PTO

[David Wise]

I've done one.  Start by taking data points, expected
vs actual every 25kHz.  If you have enough visual acuity
(or a good lighted magnifier), pick a starting point that
puts the rider exactly on top of one stack element (I
call them "chips").  Turn these data points into a graph
of actual freq divided by expected freq.  If you have the
endpoints set, the graph will intersect 1.0 at each end.
Ideally, it will be a horizontal line.  It won't be!

However, if it's a smooth curve, you may be able to avoid
the tedium of adjusting the stack.  Trimming the shunt
capacitors will bow the curve up or down.  Sorry, I can't
remember which is which, just try it both ways.  Each time,
you'll have to redo the endpoints, but for the first few tries
the only other data you need will be at midpoint.  When it's
close to 1.0, do a full run to see the residual error.
If it's small enough, smile!  Then get ready to fix up the
temperature coefficient you changed when you added or
subtracted capacitance.  By the way, the originals may
not be all that good today even if they were spot-on in
1954.  It's possible to calculate the required tempco
based on the percent change in capacitance you've perpetrated,
but I found that it doesn't match experimental results very well.
Cut and try.  Don't forget that low VFO tempco isn't the goal,
it's low overall tempco.  Your crystals have tempcos too.
All this doesn't matter much if you're one of those
hardheads who run 24x7, but if you aren't, warmup drift
is just another annoyance you can do something about.
I wound up removing the small caps entirely and
substituting a pair of trimmers, one NPO, the other
N750.  By adjusting them suitably, you can get any
capacitance at any tempco.  (Within physical possibility,
of course.  Darn language lawyers!)  To check system
tempco, I focused a tv camera on a frequency meter, and
fed the video to a VCR with a six-hour tape.  After
the run, I'd fast-forward through the tape writing
down readings.  Frequent readings at the beginning,
spacing further and further apart as the radio
stabilized.  I got it down to around 100Hz total
(down from around 800Hz), but I suspect that this
could only be maintained if you ran the radio on a
schedule.  If you leave it for a week or so, it's
off until it's been run a few times again.

Assuming you've read this far :-) the actual stack
adjustment is simple in concept but IMHO one of
those skills you have to apprentice for.  The
nuances can't be described.  The basic motion
is to loosen the hold-down screw which compresses
the stack.  With the pressure released, the chips
can be slid up and down.  You move the appropriate
chips what you think are the appropriate amounts,
close her up, and take your 40 data points again.
Adjacent chips interact, and there's a mechanical
limit to the amount of change possible between
them.  If you set a couple on a steep slope or
try to move one without moving the others, you'll
see what I mean.  It reminds me a lot of adjusting
the delay line on a 53x-54x series Tektronix oscilloscope.

73,
Dave Wise

> -----Original Message-----
> From: Jim M. [mailto:jamesmiller20@worldnet.att.net]
> Sent: Thursday, December 26, 2002 7:27 PM
> To: r-390@mailman.qth.net
> Subject: [R-390] Adjusting Linearity Stack Collins PTO
>=20
>=20
> I am looking for instructions for adjusting the linearity stack in a
> Collins PTO.  I have the PTO in an external jog now with a frequency
> counter and a calibrated scale and have the end points pretty=20
> close, but
> there is still 1-2 khz error midpoint.   I suspect the stack needs
> adjusting.  It seems to be connected to a large screw that is=20
> accessible
> from the rear of the PTO (after opening it.  Any info out there?
> _______________________________________________
> R-390 mailing list
> R-390@mailman.qth.net
> http://mailman.qth.net/mailman/listinfo/r-390
>=20