[R-390] VFO squirreliness solved FOR GOOD

[Dan Merz]

Tim,  same here on the "I like to graph things".  Thanks for the hard data
on the amounts.  I measured the pto on my Collins R-390 a few minutes ago...
The 390a Cosmos isn't as easy to get to at the moment. I found:

The Collins pto output spans about 997.8 khz, or 2.2 khz short of the 1000
khz range in going from 000 to 1000 on the dial.

The error in dial readout is almost all due to that problem.

If I correct the readings  as though I stretched the pto span by adding a
turn or whatever it would take  to the coil,  I estimate the non-linearity
of the pto to be in the range of 0.1 to 0.2 khz,  pretty darn good for
something that's probably not been messed with as far as I know. I'll check
my Cosmos unit sometime to see how it fares.  I'm not going to try to
improve this particular unit,  which spans 997.8 khz instead of 1000 khz.
Dan.

-----Original Message-----
From: Tim Shoppa [mailto:shoppa_r390a at trailing-edge.com] 
Sent: Sunday, March 05, 2006 12:27 PM
To: shoppa_r390a at trailing-edge.com; r-390 at mailman.qth.net;
mdmerz at verizon.net
Subject: Re: [R-390] VFO squirreliness solved FOR GOOD

"Dan Merz" <mdmerz at verizon.net> wrote:
> Tim,  thank you for taking the time to relay your findings.  I never 
> messed with the linearization on my Cosmos because it looked messy.  
> And I don't recall really cleaning mine like you did when I had it 
> open to take a turn off,  which might be worth the effort.

I doubt that the grease-monkeys at Raytheon put so much oily goo inside your
PTO's, Dan :-).

> I wonder how much the linearization scheme improves the linearity.  
> How non-linear are the other pto's that don't have this feature? Or 
> how much correction are you making as you tweak these?

One of my two units has a nonlinearity of about +/- 1.2kHz without linearity
correction, and the other one (the one Raytheon gooped all up inside) is
about +/- 2.0kHz without linearity correction. (By "without linearity
correction" I mean that all the little screws have been adjusted to midrange
and I've already optimized the endpoints.)

The maximum range of linearity correction by turning the little screws is
about 4kHz at the 3.55MHz end (and 2.something kHz at the 2.455 MHz end), so
even the extreme one is still within correction. That's about two turns of
the little screw. To align at the 200Hz level the screws need to be adjusted
to the tenth of a turn, and this is probably about as good as can be done.
(I'm sure somebody's gonna chime in here about how they linearized theirs to
the 10Hz level...!)

I don't think my PTO's are "cream of the crop" with respect to accuracy
before correction.

Getting them much better than 200Hz is probably not feasible because the
linearity screws aren't flat to that level on the back.

Hypothetically, I think the correction range could be extended somewhat by
taking turns (there aren't that many to start with!) off the endpoint
inductor and moving them to the linearity inductor.

Just moving some of the flying wires inside the PTO can will change
linearity at the 1-kHz level.

As to "overanalyzing" this, I've got a few dozen tables of frequencies vs
dial positions as I tweak the endpoint, sometimes at the 25kHz points but
usually at the 100kHz points. It's interesting to an extent (but also
realize that I spent many many years in grad school graphing data and
fitting curves so I'm not actually averse to the subject.)

Tim.