[Central-Electronics] 20A QRO #3 Neutralizing Project

[Jim candela]

	Group,

	I have some news about that 20A that I worked on last year for Yogie
KC5MIP. Shortly after Yogie took the rig home, something else happened, so
he gave it back in hope that I could get it going again. Unfortunately it
sat undisturbed in my garage for 6 months while I was doing other things.
Last week I fired it up, and the audio was very low, and distorted. I found
the B+ on the 6U8 triode plate at 50 volts, and it should have been around
300. I thought the primary of the 1st audio transformer was bad since I saw
15K resistance from the plate to the 300 volt tie point, and a few hundred
ohms was expected. Fortunately the wire to the 300 volt tie point was the
problem. I never did find the location of the open, so I cut out the wire,
and replaced it. OK got audio now, but the voltages are all out of whack on
the 6BA7 mixer circuit. Turns out that the 15K WW resistor (part of divider
with a 10K) was open. Replace that, and it's back to where I left it last
June.

	Whoever installed 6DQ6 sweep tubes to replace the 6AG7's did a really poor
job. Anyway the interelectrode capacity with these is much higher than
6AG7's. They worked fine on 160 meters, but on 75 meters, it was possible to
tune through a parasitic. It was still useable, but tuning was tricky. I
hadn't tried 40m yet. I decided to try a neutralizing circuit. I first put
in a pair of 6CA7's instead of 6DQ6's as these have even higher Cgp
capacitance. With these I could not operate 80 meters at all due to
parasitics, and 160 meters would work, but was a little touchy. I took a
small piece of coax (bare braid over double insulated antique mic cord)
above chassis between the output plate tuning capacitor (soldered each end
of braid directly to chassis), and the bank of driver circuit variable
condensers. This 20A had the later 12BH7 cathode follower circuit with the
tapped cathode resistor where the tap connects to the grid resistor. The
bank of variable condensers has one side bypassed to ground with a .005 uf
ceramic disc capacitor. I reduced this to about 470 pf, and connected the
center of my coax to the bussed side of the variable capacitors. The other
end of the coax connects to a 2-10 pf variable capacitor, and then to the
output plate tuning capacitor.

	I have bad memories neutralizing triodes for years past. I decided to
connect the PA tube plate & screen voltage inputs to ground, and look at the
RF output of the 20A (scope looking across a 50 ohm dummy load) when turning
on the drive, and peaking the tuned circuits. Sure enough, I could get a
null. Had to keep repeaking everything (Mixer, Plate, and 40m driver coil),
but I did get a null from the Neutralizing capacitor. Hook everything back
up, and now the damn thing tunes up, and is free from parasitics 160 - 15
meters! I cannot hardly believe that.

	I got a lot of work yet to do, but I wanted to make a preliminary report as
to my progress. Now this effort has far reaching ramifications. It means
that 20A QRO#2 will someday make an appearance on 20m SSB where all you guys
with your fancy smancy 100V's hang out!

Anyway this is enough rambling for now.

73,
Jim candela
WD5JKO







June 15th, 2003:

	Group,

	Today was a interesting day in 20a land. A friend, Yogie, (kc5mip) brought
over his ailing 20a. It had a few surprises. First the QT-1 was home-made,
and there were various pots & switches on the front & back that CE didn't
install. A little more looking, and the high dollar relay (the one with
10,000 turns of #99 wire, or so it seems) was replaced with a hermetically
sealed unit. The best was noticed last, the 6AG7's had been replaced with a
pair of 6DQ6 horizontal sweep tubes. The 5U4 had been replaced with a wimpy
5Y3.

	The effort to get this thing on the air was significant, and hour by hour
progress was made. After about 6 hours the rig relented, and things were
working fairly decent. It is interesting that the sweep tubes offered no
more than stock 20A power on 80 meters. I could get 13 watts with a 5U4
rectifier, and 19 watts with a solid state 5U4 replacement.

	The 6DQ6 sweep tubes need lower screen voltage than the 6AG7's, and the
original modifier simply used a screen dropping resistor. This is OK for
class C finals, but for AB1 linear amps this is taboo. Making a good stiff
150 volt screen supply takes effort, and more time than I had. I tapped into
the voltage divider for the 6BA7 screen, and bypassed the junction with a 22
uf @ 450 wvdc capacitor. The screen voltage starts out at about 150 volts,
and sags to about 125 volts at max CW output. This is limiting the maximum
power. With a 6 hour time window, what other option was there?

 The grid bias (G1) was still stock (-10 volts). Key the rig, and the 6U5
practically goes out(very dim). The B+ dropped from 380V to 240V from PA
idle current. I paralleled a 100 K resistor in the bias network with 180 K
to bump the bias to -15 volts. Now the key down B+ is a more reasonable 300
volts with 5U4 rectifier. Maybe the idle current was still too high, but no
easy way to measure it.

	We could now make decent AM at 1 Khz audio, but it sounded terrible (tinny)
on my receiver. More modifications to weed through. The 12AT7 two stage mic
preamp had coupling capacitors reduced from .005 Uf to 720 pf. After putting
that back to stock, it started sounding decent. Still at 2 watts out, and
sine wave 1 Khz modulation input, the + mod % was ~ 50% at the same time as
100% negative. All stock 20A's do this more or less on AM. My fix here is a
little negative cycle loading. This loads the audio feeding the balanced
modulator during the 180 degrees that the 1n34A balanced modulator diodes
are back biased by audio. Adding this network provides a more uniform load
on the audio amp over the full audio cycle. A 130 ohm resistor, and 1n34A
diode in series, and connected to an RCA plug (cathode to pin) and then plug
this into the rear RCA audio out (used for scope XY pattern). Must take out
for SSB. Now at 2 watts, the 20A would modulate to about 80% (+) @ 100% (-).
Drop to 1.5 watts, and it is symmetrical. This is a little worse than a
stock 20A with the negative cycle loading plug, and much better than stock
without the negative cycle loading RCA plug.

	Time to try SSB. Very poor SSB rejection. Shucks. Looked at the audio phase
shifts at red & green 9 mhz RF coils. I could vary amplitude, but not phase.
Getting late so tried another PS-1 from  another 20A awaiting
restoration/modification. Nice. I got a complete null of the unwanted
sideband just from the two audio pots. Switch from SB1 to SB2, and still
perfect! No way, this can't be. I have spent days trying to get it this good
before, and settled for less! Better lucky than good I guess. Never had to
adjust any of the coils..

	Look for Yogie on 3880-3890 from San Antonio for his "one of a kind" 20a.

Now I need to fix that original PS-1 that someone "adjusted"....

Regards,
Jim Candela
WD5JKO