[Collins] KWM-1 manual

[Dr. Gerald N. Johnson]

On 4/27/2011 8:10 AM, AI2Q wrote:
> Thank you, Brian. I just signed up.
>
> BTW, I got my relay contacts burnished and cleaned with the help of a residue-free contact cleaner that W1PIE gave me (it's a Miller-Stephenson product). Seems like good stuff.
>
> I also methodically removed every last Black Beauty cap. When I put these caps on my old Eico capacitance bridge and applied variable high voltage, the green tuning eye on the tester closed right up. It was Leakage City. I now have a nice little collection of nasty old Black Beauties!
>
My experience has been the same for molded paper capacitors. I have 
concluded that its not worth the effort of trying to remove them in 
condition to put them back because the overwhelming majority will not 
pass my leakage test. Sprague made those molded black beauties and for 
the replacement market they made similar looking capacitors with a mix 
of paper and plastic dielectric that weren't leaky. I don't know what is 
the fundamental flaw, just that they are excessively leaky. So I 
recommend replacing them when seen without testing. It makes the 
workbench time more productive in getting the radio to work. I observed 
those molded paper capacitors excessively leaky nearly 40 years ago.

The most sensitive circuits for leakage are the AGC time constant and 
bypass capacitors plus the audio coupling capacitors. Leaky AGC 
capacitors cause less AGC control than needed and that shows up as 
distorted audio from overdriving the detector. Leaky audio coupling 
capacitors upset the bias on the following stage leading to distorted 
audio and hot tubes. In the long run hot tubes fail and loss of bias 
(due to a faulty 8 mfd electrolytic in 75S receivers) can also burn up 
the audio output transformer. Spending a buck on a coupling capacitor 
can save a $100 plus rewind and a less common output tube.

The other most common place for black beauties are screen bypass 
capacitors. While the circuit impedance reduces their effect, a few 
stages with leaky bypass capacitors can lower the screen voltage a 
little, and cause the screen dropping resistors to heat a bit more. The 
gain of a pentode is directly proportional to the screen voltage, so a 
small reduction of screen voltage means a similar reduction in gain. 
Over several stages that can add up to loss of receiver performance that 
is hard to find by troubleshooting because its distributed over many 
stages. The carbon composition resistors used tend to drift higher in 
value with age and that is accelerated by heat, then they contribute to 
overall reduced gain.

As far as I can tell, up through the S-line, Collins did not use IRC 
resistors, the resistor specification on the vendor line said "QPL 
except IRC." That's probably because IRC resistors were film resistors 
in a molded case with different RF, thermal, and aging effects compared 
to the standard carbon composition resistor. In modern film resistors, 
the resistive element is just under a layer of paint. In the IRC it was 
on a glass or quartz tube less than 1/2 the diameter of the molded part 
so the heat was more concentrated than in a carbon composition resistor. 
That heat causes thee molded insulation to char and its subsequent 
conductivity to lower the resistor value. IRC claimed better RF 
characteristics, though they also sold film resistors with only paint 
over the film for even better RF characteristics (lower inductance and 
lower distributed capacitance). So IRC tend to drift lower with age and 
heat while every other carbon composition resistor tends to drift higher 
from age, heat, and humidity.

Resistors drifted out of tolerance are something else to check when 
working on getting a radio up to specifications. Each color coded 
resistor (except 20%) has a tolerance fourth band. They way they were 
made, odds are poorest for wide tolerance resistors to have resistance 
close to the marked value. The manufacturing process made a wide range 
of resistances around some target value, then the resistors were sorted 
into the brackets for each resistance value and tolerance. There are 
more standard values for 10% than for 20% and for 5% than for 10%. The 
standard values were selected so that there are no reject resistors from 
a batch of random distribution, every resistor falls into a bin for 
color coding with a tolerance. I think the resulting distribution may 
have depended on the orders in the resistor plant.

> The replacement Orange Drop caps, and some military quality caps I had, showed wide open green eye tests on the Eico all the way to 600V. I installed them, and then aligned the rig and balanced the balanced modulator.
>
> Last night I had a lot of fun on 20M fone with the KWM-1, making a number of contacts in Europe and South America. I even got one unsolicited report of "excellent audio."  HI HI.
>
> Vy 73, AI2Q, Alex
> Member: ARRL, FOC, RSGB, CWops, QRP-L, Antique Wireless Association, Wide Area Amateur Radio Network
> http://home.roadrunner.com/~alexmm
>
73, Jerry, K0CQ, Technical Adviser to the Collins Radio Association.