[AMRadio] 1930's Old Buzzard Open Rack Transmitter Finally Sees the Airwaves! (Long)

[ne1s]

Hello, 

Some of you may remember that a few weeks ago I posted a question regarding 
optimizing the layout for a push-pull triode HF final deck, given that the 
deck in question was already partially built (by someone else). This was a 
component of a larger project involving completion/restoration of a 1930’s 
vintage homebrew open-relay rack AM transmitter I had initially acquired 
some 15-20 years ago, and has been sitting patiently in my cellar shack 
since this time waiting for me to ge a round tuit. Well, the elusive and 
highly-coveted tuit finally arrived about a year ago, when I started 
drilling and blasting for the modulator deck. I had to take a rather long 
break from itarting last July, when I started a major effort to bring a 
majorly-hacked Johnson Valiant back from the dead for a friend which led to 
the birth of FrankenValiant. But that’s another story for another time. With 
FrankenValiant finally completed, I finally got back to the project at hand 
a month ago, and to make a much longer story less so, the transmitter made 
its maiden voyage on the airwaves last Sunday evening, when I used it during 
the final 2 hours of the AWA AM QSO party, on 75M. 

The final presently uses a pair of 812s, which are modulated by a pair of 
805s. When the debug/shakedown stage is over I’ll switch over to T-55s in 
the final (harder to find, and more expensive, than 812s). There are still a 
few problems, which I hope to rectify next weekend: 

(1) My bias supply for the modulators and finals, built onto the modulator 
deck, failed. I had used a period resistor from the junkbox – it was 
working, now it’s opened up. Replacing it will involve removing the 
modulation transformer from the deck (to make the weight more manageable, 
removing the deck from the rack, replacing the resistor, and assembly in 
reverse order of disassembly. So my on-th-air use was with grid-leak bias 
only on the finals, and zero-bias on the modulators (the 805s were still 
operating within their ratings, though). 

(2) The speech amp has an intermittent noise which modulates the finals 
quite “nicely” when it is present. No surprise: all the caps in the deck are 
original wax-paper. I did check the DC blocking caps for leakage beforehand 
(by measuring the DC grid voltages), but guess I shouldn’t have trusted 
them.  A little benchwork with the ‘scope should trace down the faulty 
component. It’s in the first stage, ‘cuz its intensity varies with the mic. 
gain control. I actually saw it on the scope when I was checking out the 
speech amp initially, and made a conscious decision to deal with it later. 

(3) The link output is directly feeding coax which directly feeds the 
antenna tuner, and I can only load the finals up to ~150 mA for about 160W 
output. Should be able to load them to nearly twice that. I need to install 
a series breadslicer to resonate the link. 

(4) I need to improve the grid drive on 10M. 

(5) I want to use a spare set of relay contacts to disable the speech amp 
(by removing the B+) on standby (receive). 

And, finally (pun intended), I would like to thank all those that answered 
my question concerning the layout of the final PA deck. If you’ll recall, 
What I had to work with was the two tube sockets mounted parallel to the 
front panel, with the output tank 2-section variable capacitor mounted 
perpendicular to the panel between them, and the final tank coil jack bar 
parallel to the panel, near the back of the chassis. The plate tank is all 
above-chassis. There was no grid tank, and I was puzzling over the best way 
to add it. I ended up putting the grid 2-section variable cap centered below 
the chassis with its shaft protruding through the front panel. I mounted the 
5-pin socked for the grid coil a few inches behind that, also below the 
chassis, so that its axis was also perpendicular to the panel (and therefore 
also to the plate tank coil). I had to be careful where I put these, not 
only for the obvious electrical design considerations, but also because the 
cap rotor (when fully open) and coils (when plugged in) extend below 
chassis, and into the area of the exciter deck. But the grid coil is also 
perpendicular to all the coils in the exciter deck. It’s not a perfectly 
symmetrical layout, but it’s pretty close, and I was able to neutralize it 
using the 10M coils. Engineering is full of compromises, and if You've made 
them correctly AND you’re lucky it works when you’re done. I seem to have 
been lucky, at least... 

73,
 -Larry/NE1S