I have an actual TT-3 in front of me.
I've got one too. Best tube pr0n ever!
On page 91 there
is a chart showing the effect of frequency on output power
and efficiency and also allowable plate voltage up to about
150 mhz. Its obvious the tube falls off rapidly beyond its
full rating frequency of 30 mhz but most of the other tubes
available at the time would do no better although
plate-to-filament capacitance of the contemporary Eimac
triodes was a lot lower. Their maximum rating frequency
limit was 40 mhz.
I'm not exactly saying that RCA lied in their graphs, but somehow what
is published there is not particularly useful for ham amp use.
I know that 833A's work in diathermy machines, and I think that those
were circa 27MHz at least in the 30's and 40's. So obviously in the
right kind of circuit I think they are useful (if maybe not top
efficiency) at those frequencies.
But my experiments with them in ham amps showed that I was unable to
get any usable output above 20M with a single tube. Meaning, I was
putting more power into the tube than was coming out. I think in push
pull some of the strays could cancel but bandswitching all the
inductor sections (as opposed to plug-in coilsets) for push pull is
onerous.
I will not really believe a 80M-10M grid-driven 833A actually working
usefully on the upper bands. I'm sure the OP has a very nicely
constructed box. I just won't believe that it's useful on 15M or 10M
until I see it working there. And even on the lower bands the problem
of neutralizing the circuit across band changes looks hairy. With plug-
in coilsets that changed the neutralization out with each band change,
I can see it. Bandswitching all that seems completely ungainly.
All just IMHO. I fully share everyone's fascination with grid-driven
big triodes and love playing with my 833A's and their link-coupled
output and neutralization networks in that spirit. But grid-driven
triodes just do not fit into the "twist the knob on the front panel to
change bands all the way from 80M to 10M" picture that modern hams
have of a linear amp: neutralization across just a single band can get
hairy, never mind across several octaves of RF.
That 60's article showing the 833A in a 80-40-20M amp is kinda clever
in sidestepping the neutralization issue by just swamping the grid
with a honking huge 50 ohm noninductive power resistor, but my
estimation is that more than half the drive power gets turned into
heat and never hits the grid. And note that they don't claim it'll
work well above 20M
There's good reasons why from the 50's onward, the beam power tubes
and grounded grid triodes have ruled: neutralization is a zillion
times easier, bandswitching from 80M to 10M is not so hard, linearity
is useful for SSB, etc.
Tim N3QE