[Milsurplus] GO-9 Project; "Which way did he go, George?"

[David Stinson]

Got the decks cleared of the BC-669 and the RAK/RAL projects, so it's 
time to start the "Winter Hair-Puller," the GO-9 transmitter.  As we saw 
before, the power supply is almost completely stripped and an 813 
shoe-horned in where an 803 should be nesting.  The power supply and 
transmitter display the sad and regrettable fate that awaited 90% of 
"conversions:"  a lot of wire-cutting  and gutting which was not 
finished- the poor thing never having done another "dit" from the time 
it left Service until now, whereupon the "convert-er," converted a 
beautiful piece of engineering into a corpse on the junk pile.   It's 
not all bad news; the transmitter came with the shock mounts, which is 
great because without them, there are things on the bottom of the 
transmitter that would be easy to break (not every engineering decision 
is a good one).  A stripped 800Hz power supply is a perfect excuse to 
build a 60Hz supply, so we don't have the 800Hz boogie-man to whip.   
Here is a simplified schematic of the transmitter for those who wish to 
follow the discussion:
https://photos.app.goo.gl/Z3sLRrEWgFRpUtVaA

We'll start by surveying and trying to understand the changes made to 
the rig, so "Which way did he go, George?"  starting with the 
almost-a-power-supply.  It's nearly stripped and a start made on a 
filament supply deck.  A large transformer supplied 10VCT for the 813.  
Two small ones were each 7.5VAC, wired in series to provide 15VCT, 
intended for the 837 MO and 837 IA.  A bunch of big mystery resistors 
here and yonder.  I decided to "begin at the beginning" and removed 
almost all the original and "conversion" parts, leaving us a 
nearly-clear cabinet in which to build our supplies, controls and 
modulator.   As it turns-out, our "converter" did put some thought into 
the changes he made to the transmitter itself, if one assumes he did not 
have access to an 803.  The 813 is wired in such a way as it would 
work.  The 813 socket is larger than the intended 803 socket, so the 
socket mounts had to be warped a bit to fit it.  He did power the 
transmitter up at some time, because the 813 is shorter than the 803, 
the plate lead touched a ground point and arced to ground at that 
point.  Insulating varnish on the wire and the ground point (just in 
case) addressed this. The original .01 uFd screen bypass capacitor is 
missing and a large mica 600 pFd capacitor mounted to a small drill 
hole.  No idea why on that one.    Reference the grid circuits of the 
three tubes. Through pin 17 of the connection between the Transmitter 
and Power Supply chassis, relay K201 closes the MO/IA gird circuits to 
ground during keying.  The original modifier lifted the grid circuit of 
the PA from ground at the bottom of meter M302 and connected this point 
to pin 17.  This allows the used of grid-block keying for the 
transmitter.  I may retain this, since the thought of the PA producing 
waste heat and running the filament-life clock, being unbiased during 
key-up, is unacceptable to me.

The "Filament Resistor" Question.
As you remember, we discussed reasons why a 4.5-Ohm dropping resistor 
was in the filament lead to the Intermediate Amp.  After some 
investigation and insights from our membership, the answer is the IR 
drop in the two .5 mH chokes in the filament leads of the MO.
https://photos.app.goo.gl/5dEd6kpZZ9Bg8wA19
R-312 is included to equalize the two IR drops, which is then made-up by 
supplying about a higher voltage to the MO/IA filament buss.  To get 
12.6 V to the tube filaments at 60 Hz, one must supply about 15.5 VCT to 
the buss.  While 15 VCT transformers can be found surplus, one of my 
goals in doing this work is the hope that others will get their rigs off 
the garage floor where it's sat for 40 years and breath life back into 
it.  Really didn't like the idea of using a non-standard transformer.  
There are a lot of 12VCT, 3 Amp transformers out there.  Radio Shack 
used to sell them by the basket and you see them at hamfest for a couple 
of bucks all the time.  So began looking for another solution.  Here is 
a photo of the MO stage.  The two filament chokes, wound basket-weave on 
a wooden core, are easy to see on the right:
https://photos.app.goo.gl/gRbTQGYbwVE7zHqT6
Modern ferrite-wound chokes can deliver significant inductance without 
large IR loses, so I decided a non-destructive solution was to bypass 
these two chokes with ferrite-loaded coils and bypass R-312, allowing 
standard 12.6VCT supply to the MO/IA filament buss.  R312 is the center 
of the three clip-mounted resistors:
https://photos.app.goo.gl/YQsntABB46pMorJRA
https://photos.app.goo.gl/GsoKN6iTZ9pqhBVQ6
https://photos.app.goo.gl/b7ncNWyhQhtt669XA

Replacing the 813 with the correct 803 hit a small snag; the 803 socket 
I had was the same size as the 813 socket- too big.  Luck was with me, 
because I'd acquired a fried-transformers GP-7 transmitter at HamComm 
many years ago to keep as a parts donor. It had the correct, smaller 803 
socket, but there was yet another catch.  There had been arcing around 
the socket ( high power at alititude? ) and there were nasty carbon 
tracks between the filament tubes and from filament to the screen pin.  
Cleaning didn't touch them.  Used plumber's emory cloth to remove the 
tracks completely, then covered the areas with a good coating of High 
Voltage Insulating varnish:
https://photos.app.goo.gl/niUjad3wVDQAGaLQ7
The 803 was originally mounted with rubber grommets at each of the 
mounting screws.  The old ones were long gone and the ones from the GP-7 
parter were crumbling and unusable.  Did some searches and asked a local 
car mechanic if hi-temperature grommets were available.  Nothing I found 
was satisfactory.  Decided to go with silicone faucet washers (that 
might come back to haunt me, HI). They were slightly too large, 
preventing the tube from seating the last 16th of an inch or so, but 
that doesn't seem to have been a problem.  In fact, it might be a 
positive as this will allow air flow under the tube.
https://photos.app.goo.gl/ULsbPa295Q5DG8Bn9
The original screen bypass cap was missing and a "tacced-in" 600 pFd 
replacement installed.  I removed this cap and installed a .01, 1000V 
ceramic disc.
https://photos.app.goo.gl/fT61C5rcthXsDjm3A

After the filament circuit changes and replacing the 803, I decided to 
run a first test at low voltages.   Supplied 420 VDC to all High Voltage 
busses.  The rig took right off and produced 15 W out with clean 
grid-block keying.  We'll call this "prototyping" because "haywiring" 
(a.k.a. "the truth") sounds so... yeah. That:
https://photos.app.goo.gl/TTm68nZ2YEApKvr96
Still much to do.  BIg power supply build ahead.  Fitting the 
transformer deck:
https://photos.app.goo.gl/f4sfAcDb5eciFZjB9
The big black transformer is 1200VCT, which with a bridge rectifier 
should deliver about 1700 VDC to the 803.  The "Economy" center tap is 
too "hot" for the Low B+ to MA/IA at 800+V, but I have a series  cap to 
tone it down a bit, followed by a MOSFET regulator to fix this buss at 
525V.  We'll see how it goes.

TNX OM ES 73 DE Dave AB5S



-- 
This email has been checked for viruses by Avast antivirus software.
https://www.avast.com/antivirus