[Collins] help needed

[Dr. Gerald N. Johnson]

On 1/30/2012 8:41 AM, Carl wrote:
> Good points Jerry. I fight parasitics often in vintage gear as many
> circuits were marginally stable and the fixes of that era were to place
> a paper and foil cap across the tube socket between input and output and
> move wires around until it was stable.....the latter is actually in
> Riders as part of the service instructions. When replacing with film or
> disc caps it was back to step one and pushing wires around wasnt an option.

The long leads between the socket and the active elements in the tube in 
all but loctal and miniature based tubes makes them hard to tame. When 
its a quarter wave of wire from the bypass to the screen, that tube 
socket bypass doesn't have much effect. That's where a trimmer cap for a 
screen bypass sometimes worked to series resonate the stray inductance 
in the screen lead to bring the screen to ground, at least for a narrow 
band HF or VHF amplfier (and in the pre loctal/miniature days, 20 MHz 
was the threshold of "VHF."
>
> What has always worked for me is a 10-27 Ohm carbon comp resistor in the
> grid lead right at the socket, triode or pentode. For the screen a 100
> Ohm carbon comp right at the socket and a .01 or .02 disc cap at both
> sides to ground with minimum lead length. Carbon and metal films rarely
> work in these cases. A few very squirrely early 30's big pin glass tubes
> needed a few turns of wire around the grid resistor to form a complete
> parasitic suppressor. Having a spectrum analyzer handy saved a lot of
> waster time.

Most I've seen like a 47 to 100 ohm on the grid and not putting any 
bypass next to the tube on the screen with that amount of series 
resistor. Film type resistors probably have a bit more capacitance 
between the end caps so a higher value may be needed.
>
> Since the S1 tone oscillator is at a 1.3kc audio frequency the screen
> bypasses should be .1's. discs which weren't economical in that value in
> 1961. Note that C-107 is a paper cap, guaranteed to be leaky, and should
> be replaced along with all other paper caps.

Not only is the paper cap leaky, its inductive. .01 to .05 sometimes 
were series resonant at about 455 kHz though the makers did offer some 
intentionally tuned to series resonance at 455 kHz. The better film like 
the orange drops are extended foil and don't have nearly the inductance 
of the standard paper.
>
>
> While the S Line is a bit newer the tubes have more gain and the cures
> are still the same. Reducing gain in the tone oscillator is another
> point to look at; having 260V on the plate and 268V on the screen is way
> too much IMO. The feedback cap C-111 is another item to look at. I see
> no reason that someone with a bit of time cant make that stage 100%
> stable with any 6EA8 or other similar substitute. This isnt rocket science.

As they did in the 32S-3 with a resistor on the grid, and in the KWM-2 
with both the grid and screen resistors. The feedback capacitor and 
phase shift network is a high pass filter so encouraging oscillation and 
some high frequency where the phase shift is again 180 degrees. Since 
they immediately drop the AC voltage by a voltage divider down to 6.8% 
of the amplitude on the plate, lower voltage should be fine, IF the 
oscillator starts good at the lower voltage. At least in the S-1 and 
KWM-2 this tone is the carrier oscillator for CW. Its at 1.3 KHz so 
harmonics are outside the mechanical filter passband. In the 32S-3 its 
only sidetone and less critical.
>
> While at National in the 60's there were many differences between
> manufacturers in very common tubes such as the 6BE6, 6BZ6, 6GH8, and
> many others. RCA was the worst by the late 60's as the end was
> approaching and cutting quality was common, performance of tubes in the
> same bulk lot were all over the place. We found Sylvania and Telefunken
> to still have integrity.

How much tube manufacture in those numbers was manual and how much was 
automated? Was the poor quality a lack of rejection at inspection or 
multiple automated machines drifting out of adjustment missing the 
mechanical parameters? Or did they grab any prewound grid even though 
each tube had a nearly unique design?

73, Jerry, K0CQ, Technical Adviser to the Collins Radio Association.
>
> Carl
> KM1H
>
>
>
> ----- Original Message ----- From: "Dr. Gerald N. Johnson"
> <geraldj at weather.net>
> To: <collins at mailman.qth.net>
> Sent: Monday, January 30, 2012 12:16 AM
> Subject: Re: [Collins] help needed
>
>
>> There was a factory service bulletin about the tone oscillator having a
>> VHF parasitic oscillation, cured with 47 or 100 ohms in series with the
>> control grid, probably. The 6EA8 has a bit more gain so probably was the
>> part of the problem. The latest 32S-3/3A manuals show 47 ohms there. I
>> didn't find a SB for that, but it might keep the 6EA8 more calm in the
>> 32S-1 6U8 socket.
>>
>> ASAB1004 added 100 pf control grid to ground for the 6U8A in the KWM-2
>> to control parasitics. The latest Rockwell KWM-2 manual shows sometime a
>> 47 ohm resistor was added between the tube screen pin and its bypass to
>> suppress a parasitic. That was in KWM-2/2A SB-10.
>>
>> All the tone oscillator circuits are very similar so parasitic
>> oscillation tendencies could be similar.
>>
>> 73, Jerry, K0CQ, Technical Adviser to the Collins Radio Association.