[R-390] R-392 Dubrow PTO endpoint

[David Wise]

> From: Tom M. [mailto:courir26@yahoo.com]
> 
> What you'll need to do is remove the cover, find the PTO 
> endpoint adjusting
> coil, and remove one turn from it, and then start over.
> > 
> > The endpoint is about 3 kHz off, and I am at the limit with the coil
> > adjustment.

Unlike most of the tuning slugs I've turned over the
years, the R-390A's PTO slug seems to lose permeability
over time.  (Usually I see them gain.)

When you tune a PTO down in oscillator frequency, you
are changing the coil from an air-core coil to an
iron-core one.  With a slug that's less permeable
than it's supposed to be, this changes the coil inductance
less than it should, giving you a smaller than expected
frequency spread*.

The spread is trimmed by the endpoint coil, which is
in series with the main one.  Their inductances add.
The endpoint coil has the largest effect when its
inductance is the largest-possible fraction of the
total.  When is this?  At 3455 (i.e. dial 0), because
the main coil is at minimum.  (This is analogous to the
more common capacitive tuning scheme, with a trimmer
in parallel with the main cap.)  The subnormal
inductance spread of the main coil dictates a small
endpoint inductance.  That's why you remove a turn
instead of adding one.

It might also be possible to restore the spread by
changing the size of the tank capacitor.  To get
a 3.455:2.455 frequency spread, the inductance has
to change over a 1.9806 ratio (3.455/2.455, squared).
For any given frequency, a larger capacitance dictates
a smaller inductance, which is achieved by translating
the slug's 10-turn range in a direction away from the
coil.  With the inductance tending towards smaller values,
the max/min ratio would be larger and you'd get a larger
frequency spread, which is what you want.

I HAVE NOT TRIED THIS.  The straight-line-frequency
variable pitch of the R-390* PTO coil may throw my reasoning
out the window.  But since you're already hacking,
it's less intrusive to tack on a cap than to
chop a coil.  Note that with the main slug in a
different relation to its coil, you will have to
do the dreaded linearity calibration.
But the slug has already changed (that's why you
ran out of endpoint), so you might need to anyway.
Take a linearity reading, add a cap, read again.

Note also that the Temperature Coefficient of the
added cap is important, but don't freak out.  My PTO
turned out to pretty far off, which means that yours
might be too.  It's not surprising; if a slug changes
permeability, why not TC?
Experiment with different TCs until you get
satisfactory overall stability.

* It seems to be accepted jargon here to call this
"long", but as I mentioned to Tom, it makes me flinch,
because I've always compared the electrical spread to
a 10-turn mechanical baseline.  The electrical spread
is "short".  If instead you take 1000KC as the baseline,
you have to turn the dial "long"er than 10 turns.
Say what you like; I'm going to footnote my usage from
now on.

73,
Dave Wise