[AMRadio] Legal limit AM amplifier, homebrew

[Bernie Doran]

Hi: Rick": push pull can be run in any class, that is a function of the 
bias/ conduction angle.   A pair of  304TLs will produce 1800 watts at 3KV 
in class AB2 and 2400W in class C, A Pair of 833s will do about the same and 
2700W  in class B at 4 KV.   a pair of 813s are significantly cheaper and 
good for  1 kw in on CW and 800 in on AM.   They would last a long time 
running at a little lower input. The CCS ( continuous commercial service) 
rating on plate mod service shows 1600 V   150 MA and 180 w out for one tube 
and they can be turned off in standby periods of longer than 15 minutes and 
reduced to 80% during non transmit periods.  Those are RCA recommendations. 
A pair of 811A would be excellent modulators at that power level and a pair 
of 813s Would produce the maximum of 1500 PEP allowed.    The whole thing 
with the PEP rating came about because the old 1 KW input rating was 
difficult to measure on a plate meter jumping around following voice 
waveforms. So the FCC changed it to PEP and made it impossible!!!   The so 
called 375 W carrier is based on 100 % modulation, if you run less than 100% 
you can run more carrier, 120% modulation means you have to legally run less 
carrier.

813s require screen modulation, not a big deal at the lower voltages, just 
use a dropping resister from the plate supply. at higher voltages it 
"wastes"  a lot of power. Best way to do that is combine the bleeder 
resistor into the plate dropping resister.    Now at the 1 KW input level 
there are a lot of triodes that will work very well and do away with the 
screen issue, 1 KW in makes one big CW signal on a good antenna and the 
drive requirement matches your DX100 much better.   813s only need about 10 
W for a pair.    For 1940 era transmitter that would be one way to go and 
forget about the full 1500 W out on CW.  810s were mentioned earlier and 
would make a great pair for a vintage design as would a lot of other 
triodes.    Stay with low mu triodes as they modulate better than high mu. 
Ignore some of the alleged designs that show triodes with Pi net outputs and 
a grid neutralization, use only push pull and the old criss cross 
neutralization.  Very difficult to make grid neutralization work because of 
the extreme unbalance that occurs when one side of the grid tank has the 
entire load.


 four 833s is way over kill, even two is unusual. however a pair would do 
about a 400 W carrier as an AM linear and then an easy 1500 out on cw. If 
you run SSB also that does become an interesting  way to go.  833s do have a 
socket, available sometimes on ebay $$$$ but you can simply use the fined 
heat radiating gismos on the filament terminals and mount them on an 
insulating board. ( leave them loose on the board to allow for expansion) 
Gates did that.

Plate voltage---  every thing gets big when you go over around 1500 V 
$$$$$$ and I get very nervous when I read some posts on here.  There is NO 
SUCH THING  AS AN ELECTRICAL SHOCK when dealing with any significant DC 
plate supply.  It will simply tear the Hell out of you if you survive. There 
will be deep burns and scaring, and probably nerve damage.   also prudent to 
have someone with you when working on this stuff.

Now after all that, the difference between your DX100 and the  so called 
legal AM transmitter will not be detectable on the receiving end under 
normal band conditions, less than one S unit, what ever that really is. 
normal fading probably runs three or four S units.  However the " coolness
 factor"  has to be taken into account and for a lot of people like me that 
is the main issue.  To that end consider using a one tube crystal oscillator 
driving the 813s  on 80 and 40 meters.  does not get any simpler, a 6L6 and 
a few crystals will do the job and be very "cool" Be sure to get a crystal 
for around 3705  that is where I usually am.  Use the ugly round meters and 
black paint on the panel.  Have fun. Bernie