In article , "David Forsyth"
writes:

Would it be possible, though not necessarily practical, to make a reciever
for the modern FM broadcast band, using only pre-WWII tube technology? I
know they had FM on a lower frequency band at that time. What sorts of
tubes could one use from the 1930's to make a reciever that could tune in
the modern FM band? I'm sure I wont actually attempt such a thing,
especially any time soon, but just wondered how difficult it might be. Are
there any schematics or construction articles from the late 30's about
making FM radios that might be adapted over to the new FM band?

I doubt you will find much in amateur radio periodicals. They do
exist in various corporate libraries, though, since the VHF and up
radios were being pioneered in the late 1930s, some of the first being
police mobile radios. Those would evolve into the SCR-300 backpack
"walkie-talkie" of WW2 Army use and those shiny chrome button
channel select "tank radios" used mobile back then. The AN/TRC-1
through TRC-4 of WW2 times was low VHF (70 to 90 MHz) using
octal base tubes, PM with lots of multiplier stages following a MF
crystal oscillator.

Of course there are lead length problems with big bases as well as
the electrode connecting leads within bigger tube envelopes. What
most folks don't consider is the electron transit time within the tube
structure. That is slower than the speed of light (don't have an exact
value handy) and will cause a significant phase change between
grid input and plate output at VHF and above. The "lighthouse" triode
structure (ultimate may be the 2C39) has terribly short tube electrode
spacings allowing operation on up to 2.5 GHz...very quick electron
transit time internally.

Note: Klystrons and magnetrons go much higher in frequency. In the
case of the klystron, an integral tuned cavity structure is an absolute
requirement for operation. The magnetron depends upon both the
diode spacing (it is only a diode) and the magnetic field and the voltage
and the output tuned cavity structure to oscillate at X-band.

Anyone can play games with old, big tubes used at low-VHF and
arrange all kinds of neat tuned circuits to work with long leads and the
long electron transit time. Problem is, the amount of extra components
isn't really worth it. Whenever a tube has to be replaced (happened
often due to filament technology still lagging) then you would need to
do a humongous amount of retuning. About the only thing workable
for the oldie tubes is the distributed amplifier wherein LOTS of tubes
were arranged along tapped delay lines for grid inputs and plate
outputs. [the Tektronix 54x series of oscilloscopes uses such an
arrangment as the final CRT voltage driver for vertical deflection
plates] Very good as a space heater for a residence in winter...

The electron transit time thing is akin to a low f_sub_t in transistors.
Such low f_sub_t bipolars might be okay on low HF but their
characteristics don't allow good amplification or easy oscillation at
VHF and higher. Some of the newer SiGe bipolars have f_sub_ts in
the tens of GHz range.

Electron transit time depends on filament-cathode temperature, tube
geometry (and element spacings), and accelerating potential (screen
and plate quiescent voltage). There's no precise value that fits all
tubes. To view what did work at 70-90 MHz, find an old AN/TRC-1
Technical Manual and see which octal base tubes were used there.
The transmitter box final amplifier was an 829, good for about 50 W,
but not an octal base tube.

Len Anderson
retired (from regular hours) electronic engineer person
[and caretaker of TRC-1s and TRC-8s a half century ago, among
other tube-based radios]