[Prev][Next][Index][Thread]
Re: Pulsed Tube Coil Work (with updates)
"Hi Ed,
I use a blocking oscillator for my 833 A coil to pulse it, and I
will
try grid leak pulsing (John calls it sputter mode) on a larger 2-tube
833C coil that I am now working on. The grid leak resistor is not very
large for the 1 tube coil, only 40 kOHM, with a .001 microF cap in
parallel, and according to my scope, the coil pulses around 20 to 30
times per sec. The rate of pulsing depends on the value of RC and also
the voltage that the capacitor gets charged to, which is dependent on
grid coil no. of turns and coupling. The frequency of the pulses is
unfortunately not very stable for me, and as the main tuning cap heats
up
and other changes occur, it will slow down and deliver as few as 2 or 3
pulses per sec. I think there is another effect besides slowing down the
repetition rate when the grid cap is made bigger, because when I use
.004
microF, i get loud, crackly, more conventional bluish sparks as opposed
to a thick orangy spark. I think the only real advantage is that its
simpler than an external pulsing circuit. From my experience, it is
difficult to determine beforehand what values of R and C will deliver
pulses of a certain frequency. If you know any pointers on this, it
would
be a big help to me.
-Carl"
The resistance you are using seems mighty low to me. I'm looking at an
old Navy primer on radar (RADAR SYSTEM FUNDAMENTALS, NAVSHIPS 900,017,
dated 1944) which shows simplified schematics for several self-pulsed
UHF transmitters using tubes like the VT-127/227/327. The values of
grid leak vary from 1 to 20 megohms, and the grid capacitor is typically
0.003 ufd. Operating plate voltage was of the order of 10 to 15 kV, and
typical peak power output for a 4-tube transmitter was of the order of
50 to 100 kW, at a frequency of around 150 MHz. These particular
oscillators probably had pulse lengths of the order of 4 microseconds or
so, which is way too small for a TC driver, and a pulse repetition
frequency of the order of 500 Hz, which might be about right. I would
think you would like the pulses to be 100 microseconds or so, to allow
time for the secondary to "ring up" (depends on coupling factor, of
course) but perhaps some of the tube gang can comment. To lengthen the
pulse you would need to use a larger grid capacitor.
To me the advantage of the blocking oscillator would be in the use of
very high plate voltage to give high peak powers, and blocking to keep
the average power input within reason. I am particularly attracted with
the idea of rectifying an NST to get the plate voltage.
Ed
Ed