[Collins] 30L-1 Repair

[Dr. Gerald N. Johnson]

On 7/5/2014 7:08 PM, Carl wrote:
>
> ----- Original Message ----- From: "Dr. Gerald N. Johnson"
> <geraldj at netins.net>
> To: "Carl" <km1h at jeremy.mv.com>
> Cc: <collins at mailman.qth.net>
> Sent: Saturday, July 05, 2014 3:02 PM
> Subject: Re: [Collins] 30L-1 Repair
>
>
>>
>>
>> On 7/5/2014 7:18 AM, Carl wrote:
>>>
>>> ----- Original Message ----- From: "Dr. Gerald N. Johnson"
>>> <geraldj at netins.net>
>>> To: <collins at mailman.qth.net>
>>> Sent: Friday, July 04, 2014 5:10 PM
>>> Subject: Re: [Collins] 30L-1 Repair
>>>
>>>
>>>> The cathode circuit and band switch is indeed suspect, but so are the
>>>> tubes and the grid bypassing. These tubes from the late 1930 have
>>>> relatively long leads on the grids and the grounded grid circuit
>>>> depends on getting those grids effectively grounded. A change of grid
>>>> bypass capacitor type or lead length may contribute to the problem or
>>>> 811A made primarily for audio applications may contribute.
>>>
>>>
>>> ** The symptom points to an input circuit problem. Any disturbance in
>>> the actual grid or parasitic circuits wouldnt surface on 20M but rather
>>> 10M.
>>>
>>
>> Yes it does but the 220 pf on the grid was probably chosen to series
>> resonate the grid to ground so the grid was at ground better than the
>> socket pin. Its often necessary to adjust the grid circuit even in a
>> planar tube at VHF to get adequate isolation from the grid in a
>> grounded grid stage.
>
> ** Collins is floating the grid for DC and some Jr engineer likely came
> up with that 220pf method which is very frequency dependent. The stage
> is also poorly neutralized. Did Art stick his nose into this also?? (-;

They switch the bias to the tubes to cut them off for receive and to let 
them cool and the whole PA to cool.

When its three inches from the socket pin to the middle of the grid, 
grounding the socket pin doesn't ground the grid, especially above a few 
MHz and for sure not at 30 MHz, but series resonating is one way to get 
better stability. Neutralization is very difficult when the grid leakage 
from not being bypassed AT the grid is different for each band.

Art probably could have stuck his nose into that but that's too 
technical for his knowledge base.
>
>
>
>>> 20 to 30 MHz was high
>>>> enough in the 30s that the power rating of the 811A was already
>>>> reduced from its low frequency rating.
>>>
>>> ** While the 811A is a 811 with a more robust anode it is still capable
>>> of good performance to 6M in a well designed circuit. I regularly
>>> convert various 572B amps to 6M and it is simply a 811A with a thick
>>> graphite anode.
>>
>> Providing the grid is resonated.
>
>
> ** This is GG, no resonating involved. Bypass the grids hard with a .01
> and neutralize.

That won't fly with 811A, the wire to the grid is too long inside the tube.
>
>
>>>
>>>
>>>>
>>>> I note in the Collins 30L1 manual that they want a 20.5 foot long
>>>> piece of RG58C/U between the exciter and the linear. Rumor has it that
>>>> had some effect on linearity, but it also might have a big effect on
>>>> match, being about 1/2 wavelength at 20 meters in polyethylene coax,
>>>> not foamed coax. And will improve the SWR seen at the exciter just
>>>> from the loss in that coax. The round trip loss would be about 0.8 dB
>>>> reducing the reflected power a bit. And interacting with the input
>>>> tune circuit to affect the match seen at the exciter end of that coax.
>>>
>>> ** The input circuit was poorly designed as was the norm in the 50's and
>>> the specific length of coax was adopted by many.
>>
>> It may not be a benefit with a driver other than a pair of 6146 with
>> limited load range as is typical of KWM-1, KWM-2, and 32S. But it
>> might have an effect when the SWR is so poor. I'm not sure the reason
>> for the cable length was ever made public, its been argued a few times
>> since then, probably helped intermod performance of the 32S 30L-1
>> measurement on the band it was tested on and really didn't help any
>> other bands. But it might be a useful test to install it in this case.
>
>
> ** There are loads of 6146 and sweep tube rigs that still need that
> cable at least in some 30L1 production runs.

What are the symptoms of its need? Hard to tune the exciter? Excess 
intermod, (hard to detect on the air with the poor intermod performance 
of the sweep tube rigs)? Low PA gain?
>
>
>>>> As noted recently on this forum mica capacitors are pretty good but
>>>> can fail.
>>>
>>>
>>> ** Thats fairly common when the bandswitch is set to a different
>>> frequency than the exciter. Those micas arent built to be part of a
>>> 50-100W dummy load.
>>
>> Actually they have to carry the RF plate current/cathode plus grid
>> current. They are handling the RF plate current of a KW worth of
>> tubes. They are working hard. Silver micas dipped or molded won't
>> survive there, the current carrying surfaces are too thin. They need
>> old fashioned foil and mica or large low temperature coefficient
>> transmitting ceramics.
>
> ** The subject was about input network caps going bad as I described.
> Failure of the grid bypasses is a different cause/subject.

I'm talking cathode circuit capacitors they get the same RF current as 
the other end of the electron stream at the plate. They don't get the RF 
current in the plate to grid capacitance.

The grid capacitors get that RF current in the plate to grid 
capacitance, which is most of the tube output capacitance in grounded grid.
>
>>
>> What can be as hard on the cathode circuit is to be on the right band
>> where they are resonant, but the plate voltage isn't on for some
>> reason, then the driver can develop high voltage in the one next to
>> the PA cathode and damage that capacitor.
>>>
>>> Check to see if the whole slug in in the coil, if it got broken
>>>> off or a wrong slug inserted it won't tune properly either.
>>>> Unfortunately the parts list doesn't give the measured inductance
>>>> range of the coil to check for the proper slug.
>>>>
>>>> 73, Jerry, K0CQ, Technical Adviser to the Collins Radio Association
>>>
>>> ** While the 30L1 isnt the easiest amp to work on it is still rather
>>> basic. If not already done its a good time to replace the various
>>> electrolytics and check the other high failure items as mentioned on
>>> several web pages.
>>
>> For sure.
>>>
>>> Carl
>>> KM1H
>>>
>>>
73, Jerry, Technical Adviser to the Collins Radio Association