[Collins] 32S-1 band oscillator -- fixed, thanks

[Dr. Gerald N. Johnson]

In this circuit the capacitor from screen to ground is not really a 
bypass, its part of the feedback divider. Its like a Colpitts oscillator 
with the ground at the cathode instead of the screen.

The cathode capacitor is a new problem, but I think I can explain it. It 
takes only 11 nhy inductance to resonate a .01 capacitor at 15 MHz, 
where the capacitor and leads are series resonant and that gives the 
lowest impedance for the bypass. The value of .015 and that same 11 nhy 
moves the resonant frequency down to 12.4 MHz and its inductive above 
that resonant frequency with the impedance rising as you go higher in 
frequency. That capacitor inductance adds to the cathode lead inductance 
in the tube and transforms through feed back to a lower input resistance 
at the tube control grid which loads the crystal more. Add in crystals 
that have more loss than when new and the crystal doesn't oscillate 
besides when the cathode isn't solidly grounded the tube gain is reduced.

By the parts list in my 32S-1 manual, that capacitor has a tolerance of 
-20% +100%, so its acceptable by that specification if between .008 and 
.02 mfd. Clearly its not so acceptable if the leads are wired on the 
long side for the oscillator bypass.

Likely new tubes with more gain that spec and new crystals it has worked 
a long time, that the capacitor hasn't changed, just wasn't quite right.

On the other hand, its conceivable that the capacitor has developed some 
series resistance in the metalization that didn't change the capacitance 
or show up as shunt leakage on the megger but that added resistance to 
the cathode for inverse feedback and reduced stage gain. That added 
series resistance would only show if checked for that in an RF bridge.

73, Jerry, K0CQ, Technical Adviser to the Collins Radio Association.

On 1/22/2011 3:37 PM, Al Parker wrote:
> Hi everyone,
> Many thanks for the responses, some were "off list", some on-list were
> later followed up privately. Jerry J & Ed K hit it on the head(s)
> "Cathode bypass and current path are critical to oscillator gain."
> After all I'd checked out, the only things left were the screen and
> cathode bypass capacitors to check, and that required some
> desolder/solder work. I decided to do things one at a time so we would
> know just what it was.
> Russ D said that these things are 50 yrs old, we gotta suspect
> capacitors. I have found a few "real" micas that are bad, and maybe a
> cupla disc ceramics in the past 5 yrs or so, so age is getting to be a
> problem for all of us ;-)
> It was the cathode cap, not really a failure, but it had moved a bit
> high, from 0.01 to 0.015 mfd. I was using a handheld meter, and suspect
> it's accuracy, so I checked several new ones and found 0.008 to 0.012,
> so I selected one close to 0.10. That did the job, good oscillation on
> 15 & 10m with several different xtals that hadn't worked before.
> The "bad" cap was 50% high in value, reducing the oscillator gain I
> guess, tho' I'd expect that the Xc at that frequency would so low that
> it wouldn't make much difference if the cap was that much higher. Shows
> how little I know of the process. The capacitor, a disc ceramic, showed
> 100's of megohms resistance at 500vdc on the megger. Not what you'd
> expect to cause a problem, but that's all that was changed and we're
> back in operation.
> Thanks again to all who lended a hand, this is a great hobby with great
> people.
> 73,
> Al, W8UT
> www.boatanchors.org
> www.hammarlund.info
>
> "There is nothing -- absolutely nothing -- half so much
> worth doing as simply messing about in boats"
> Ratty, to Mole
>
>