[R-390] PTO Rebuild - and Core Disaccommodation and other effects

[John KA1XC]

Hi Miles and group,
that's very interesting about what you have found about the PTO caps and
corrector adjustments. No doubt the handful of old caps that are in that
circuit contribute to many PTO problems but there are also another couple of
ageing mechanisms at work, one of which can easily mimic the symptoms you
describe and may be the "core" of  some problems.

I've got a friend who is an old time Bell Labs engineer and fellow R-390
aficionado. Having worked there for a couple years well after it became
Lucent I learned how long ago the facility used to be a place of expert
manufacturing of *all* the parts used to build the telecom equipment; they
did their own metal work, cut their own crystals, grew their own quartz,
wound their own coils and transformers, silicon wafer processed to make
their own transistors and ICs, and mixed and fired their own powdered iron
and ferrite magnetic materials. They literally had train loads of raw
materials unloading at one end of the factory while finished equipment was
trucked out the other end; they were true masters of engineering and
materials science. Well, that was then and this is now and most everything
that one needs to build anything is available from a catalog but some of the
knowledge remains.

Anyway, he was very familiar with the magnetics part of the operation and
told me of  the effect of magnetic core material disaccommodation which
simply put is a naturally occurring decrease in the permeability of all core
materials over time. The permeability change follows a log curve - it is
rapid starting with the manufacture of the part but its rate of change
sharply falls after a relatively short period of time but it does continue
forever. In very critical designs the effects of disaccommodation are
considered into the design and are translated into a typical lifespan for
the part where it will stay in "spec". These PTO cores (as well as most
other parts) are now 30, 40, and 50+ years old so it's safe to assume they
are well out of their intended lifespan and can cause havoc and hair pulling
when attempting precision measurements and adjustments of such things as
linearity corrector stacks. It is his opinion that the R-390 PTO cores are
affected by this effect and I also concur.

What the actual composition mix of the PTO core material consists of is
anyone's guess but it's probably safe to presume it's some type of iron
powder mix since its characteristics provide the best stability for LC
resonant circuits, however the industry literature states that its
disaccommodation factor is much higher than ferrite core materials and thus
has more long term drift. What this means is that as the PTO core ages its
permeability decreases, so the PTO coil inductance decreases and the
oscillator frequency increases. This is also the same symptom of an
oscillator trimmer cap decreasing in value, so it makes sense that it could
be compensated for by adjusting the value of the trimmer capacitance upward
as Miles described.

There is also a second and perhaps even larger factor in considering PTO
inductor value drift and according to the literature it involves the
mechanical stability of the coil winding as it ages over time. Coil forms
and even the binder material used to anchor the windings to the form do go
through small dimensional changes over long periods of time. My *guess* is
that the coil form shrinks which would mean a tightening of the windings
and/or a decrease in the coil diameter thus a probable increase in the
inductance. This is *counter* to the effects of disaccommodation but may
explain why some PTO's tune "long" in span and why turns or partial turns
have to be removed from the series-connected endpoint adjusting coil in
order for it to get back in range. (BTW this endpoint adjustment coil is
also susceptible to all the various long term drifts as well.) In this case
a downward adjustment of the PTO trimmer capacitance might bring things back
in the right direction.

High temperatures accelerate the effects of disaccommodation and coil
dimensional drift, so maybe even whether a PTO was run with its ovens ON for
a period of years had an effect on the various internal parts as well.
Hopefully these various aging drift effects (and counter effects) and
unknowns about the operating environment may shed some light on the all the
crazy PTO symptoms seen and make them appear to be a little more science and
a little less voodoo :^)

Lastly, one of the interesting things about disaccommodation which sets it
apart from a pure aging process is that it is a resettable and repeatable
effect. If the core material is raised above its Curie temperature (or even
severely mechanically shocked) the core will reset to its *original*
permeability value which will then start to decrease again at a logarithmic
rate. If someone wants to win the R-390 Scientist of the Year Award a very
interesting experiment would be to take an old PTO core and measure its
permeability, and then bake it until it hits its Curie temperature (possibly
a few hundred deg. C) and let it cool. Then let it age undisturbed for a
period of a few months until it reaches the slow moving part of the
disaccommodation log curve and measure it again. What you should now have is
a PTO core which has been restored to its original factory permeability!

For the technical definition of Disaccommodation Factor see  #12 of :
http://www.elnamagnetics.com/library/catalogs/TGL/TermsDefinitions.pdf


73,
John
KA1XC

----- Original Message ----- 
From: "Miles B. Anderson" <mbalaw at optonline.net>
To: "R-390 Mailman" <R-390 at Mailman.qth.net>
Sent: Wednesday, April 06, 2005 9:20 PM
Subject: [R-390] PTO Rebuild


> I'd like to caution against jumping to the conclusion that the corrector
> stack needs adjustment whenever the PTO can't be made to track at the
> intermediate points. I tried once to adjust the corrector stack on a
Collins
> PTO and made such a hash of it that the whole assembly wound up in my junk
> box.
>
> I later discovered that the real problem is often much easier to solve.
The
> problem in my case was that one or more of the little unencapsulated mica
> capacitors went west.
>
> There is no way to adjust the shunt capacitance of a Collins PTO. All you
> can do is adjust the start point of the tuning slug and the little series
> inductor that is used to trim the end point. This means that there is a
> unique shunt capacitance which will make the tuning equation come out
right
> at both ends and the middle. If that shunt capacitance changes because of
> aged components, no amount of fiddling the inductances will make the
tuning
> linear anywhere except at the end points.
>
> The procedure I followed was to adjust the start point and the end point
as
> per the manual. Then tune the PTO to the midpoint (500). If the oscillator
> is high at the midpoint, add more shunt capacitance. If it is low, remove
> shunt capacitance.
>
> Adjust both end points again per the manual and check the error at the
> middle. If it is still off, repeat the process.
>
> We are only talking about 10 to 30 pf difference, but that small
capacitance
> difference can knock the daylights out of the linearity by putting an
"error
> bulge" in the middle.
>
> When you think you have the middle and both endpoints "spot on," check the
> tuning error every 100 kHz. If there are two "error bumps" (at, say, 300
and
> 700) these can be washed out by simply over-compensating the middle so the
> error runs in the other direction.
>
> By the way, I replaced the inner shield cover before each measurement.
>
> The main thing to bear in mind is that the corrector stack was set
properly
> at the factory. The main coil is heavily doped and is not likely to move
or
> change. What IS likely to happen is a change in the shunt capacitance.
>
> The shunt capacitors are intended to be temperature compensating, but this
> never bothered me much. Ordinary NPO ceramics seem to work fine.
>
> Miles Anderson, K2CBY
>