[AMRadio] plate modulating an 809 experiments and numbers

Patrick Jankowiak recycler at swbell.net
Thu Jul 8 00:21:42 EDT 2004

I had this e-mail conversation with a friend about plate 
modulating an 809 tube. It sort of turned from trying to find out 
why some fuses blew, to trying to determine the easiest way to 
match the 809 to a despicable 60W solid state amp for plate 

I thought it might be of interest, especially the calculatons 
show how to match a 70V line output of an amplifier to a variety 
of load resistances based on 'book' operating conditions for the 
809. Finally, a novel method of continuously variable impedance 
matching is described.

names witheld to protect the guilty.

 > Pat,
 > Hope you had a good holiday weekend, and managed to stay out 
of the
 > heat...
 > I managed to get some time to experiment and try to cheat the 
gods of AM,
 > but I think they forgave just this one time..
 > I took another power transformer that was 120V to 1200V and 
hooked the
 > primary to the 70 volt line of my despicable sand state audio 
amp. The
 > 1200V secondary was put in series with the plate supply of the 
 > which runs 750 volts at 70 ma..
 > The modulated waveform was looking good, plenty of up and down 
 > Suddenly, on a voice peak, the RF went dead.. Both the primary 
AC line
 > fuse and the secondary rectifiers-return--to-ground fuse (3 
amp and 1 amp
 > respectively) had vaporized..
 > Never heard any arcing in the transmitter part.. Replaced the 
fuses and all
 > was well again.. Played a bit more, but with less audio and 
looks like
 > nothing
 > fried, but I'm wondering what happened... Only thing I can 
think of is that
 > the audio peaks, added to the transformer voltage exceeded the 
diode string
 > PIV maybe ??  Do I actually need 6 X PIV rectifiers in a plate 
 > supply ?? ( I've got 3 X PIV now..)
 > Anyhow, I think the 809 power supply needs some work / 
revision to stand
 > up to plate modulation.. Any opinions ??

Hmm. possibly a flashover in the rectifier during a negative 
modulation peak, or a flash to ground (or primary) in the 
transformer? If PIV is exceeded, it usually damages the 
rectifiers. Maybe put a string of 4-5 rectifiers in series with 
the B+ to the mod xfmr to help keep negative peaks (negative at 
the tube, positive at the B+ HV) rectifier from over-volting the 
B+ rectifiers. Could have been also a capacitor arc. what's the 
cap. volt rating in the power supply?

Try putting a power resistor (same value as 809's impedance) in 
place of the transmitter, and use a 10:1 divider to measure the 
audio on the scope, see how much voltage you are getting. Should 
be 1500V peak to peak, so you get 0 to +1500 at the tube. If you 
grossly exceed it on a peak, maybe something's flashing over. A 
transient in he amp could cause an inductive kick, especially 
when the mod. peak goes negative and there is little plate 
current. Also, the transformer was designed for 600-0-600, so 
putting the CT at +750 could make a flashover hazard to GND or 
the primary.

What is the audio amp rated, was it 60 watts?
Using the 70V output on the amp,
60W/70V=0.86A through 81.4 ohms.

so the amp's output there at the 70V terminals is 80 ohms.

Hmm.. You know, the 8K CT to 2800 ohm 120W modulation xfmr we 
thought about using, would stand the voltage well. Let's look at 
the RCA tube manual TT-3. (use fixed width font to display such 
as courier new)


809 plate modulated class B amp.
plate______________________500.0_____750.0 V
bias________________________-5.0_____-10.0 V
peak RF grid volts__________35.0______40.0 V
DC plate current____________50.0______50.0 mA
DC grid current______________6.0_______5.0 mA
drive wattage________________1.4_______1.5 watts
power out (carrier)__________7.5______12.5 watts


809 plate modulated class C amp.
plate______________________500.0_____600.0 V
bias______________________-160.0____-160.0 V
peak RF grid volts_________250.0_____250.0 V
DC plate current____________83.0______83.0 mA
DC grid current_____________32.0______32.0 mA
grid resistor_____________5000.0____5000.0 ohms
drive wattage________________7.2_______7.2 watts
power out (carrier)_________30.0______38.0 watts

note this is a grid leak circuit, therefore the 5K resistor. 
Otherwise you use bias of the value shown.

so.. modulator load impedance for the operating parameters:

1----10,000 ohms
2----15,000 ohms
3-----6,024 ohms
4-----7,228 ohms

the 80 ohm output of the modulator, transformed through a 120V to 
1200V transformer, is 8000 ohms:
voltage ratio________________1200/120=10
impedance ratio______________10*10=100
transformation of 80 ohms____80*100=8000

pretty close. possibly two xfmrs could be used, to end up with a 
high voltage rated MOD xfmr at the tube end. Using the available 
2800 ohm to 8K ohm transformer, you would need to step up 80 ohms 
to 2800 ohms.
impedance ratio_____2800/80 =35
voltage ratio_______SQRT 35 =5.916
power transformer of 120V,
709V sec. would be 2800 ohms, if the pri. were 80 ohms.

To crosscheck:
60 watts/2800 ohms= 0.0214 (I squared).
I= SQRT 0.0214 = 0.14638A
2800 ohms * 0.14638A = 409.87V
Since we are using 70V and not 120V, the secondary of the 
intermediate transformer would be
(close enough to 409.87V considering not using consistent decimal 

So, this is a possible solution, if the power transformer you are 
using is arcing over.

Taking the need to match the parameter list for the tube 
operating conditions shown in the TT-3 manual,
1----10,000 ohms
2----15,000 ohms
3-----6,024 ohms
4-----7,228 ohms

and using the said actual 2.8K-8K modulation transformer with an 
impedance ratio of 1:2.857, the intermediate (power) transformer 
impedance ratios would be:

parameter_____pri / sec_______pri / sec
______________Z mod XFMR______Z pwr xfmr

1:____________3500/10000______80 / 3500

2:____________5250/15000______80 / 5250

3:____________2108 /6024______80 / 2108

4:____________2529 /7228______80 / 2529

Taking the Z ratio, and doing a SQRT on it gives the voltage 
ratio. The power transformer voltages to match the 80 ohm 
modulator  to the modulated load, through a modulation 
transformer having a 1:2.857 impedance ratio would be:

parameter__pri / sec___________pri / sec
___________Z ratio____SQRT_____Volts

1:_________.02285_____.1512____120 / 793V

2:_________.01524_____.1234____120 / 972V

3:_________.03795_____.1948____120 / 616V

4:_________.03163_____.1778____120 / 675V

This is all speculation. A fun exercise. I don't remember how big 
that audio amp was, but if it is a 60W amp and you use the 70V 
terminals, here be the numerals. All just in case the plate 
transformer you are using is flashing over..

Now, here is a fine and mostly unknown trick to adjusting 
modulation transformation impedances with the twist of a dial:

Use a 120V variac rated 3A with the modulator's 70V, 100V, or 
120V audio outputs (many amps have such outputs available), and 
there you have a good autotransformer which can be adjusted to 
match almost any impedance which is -lower- than the output 
impedance of the modulator. The tap feeds the modulation 
transformer's 120V primary.

So, if you make your low Z limit to be at the halfway point on 
the variac (using all these wild calculations for voltage and Z 
ratios etc), your high Z limit will be 4 times the low Z setting, 
since the variac output voltage will be 100% instead of 50%.

I have used this method before in public address situations and 
it does work very well provided the variac is oversized by 2-3 
times the audio power level. Who needs fixed taps?

With this lash up, feeding the amp through the variac to a 
120-to-1200V power transformer, then feeding the 1200V side to 
the 2800 ohm winding of the 2800-to-8K modulation transformer, 
you match perfectly at some point between and 51% and 81% on the 
variac dial, for each of the 809 conditions from the TT-3 manual.

1: 10K   66%
2: 15K   81%
3: 6K    51%
4: 7.2K  56%
X: ~24K  100%

Well, have good modulation scheme thoughts.. I have TXSG duty 
this weekend.

More information about the AMRadio mailing list

This page last updated 16 Oct 2017.