200 watt A M transmitter

[W4RFM]

My apologies everyone, I am just trying to finally get it right, and I know you folks know a lot more about this than I do. 

Okay, here is a plan, using the Apache mod iron rated for 100 watts, I build a modulator with 2x 807's using a plate voltage of 750 at 240mA for 180 watts which becomes more like 117 according to the tube values, coupled to that 4-125a R F stage running at 1750 volts with .175 plate current for 306 watts which should shake out to around 200 real watts "at the wheels", and I have a mod impedance of 3125 ohms, and a R F Plate impedance of 10,000, so my Apache iron backwards at 3K to 11K, would be a good fit.  Yes? No?  Also I would still use the mod reactor of minimum 14HY and a 4Mfd cap.  Your thoughts?

[AB2EZ]

A few clarifications (just terminology)

1. The ratio of: B+ on the plates of the rf tubes / average plate current in the rf tubes (without modulation) is usually called the modulation resistance of the rf output stage. [The rf output impedance is something else]. As you calculated, the modulation resistance of your proposed rf output stage is 1750 Volts / 175 mA = 10,000 Ohms.

2. To get the maximum amount of power out of you modulator's output stage (with reasonably low distortion), you need to present a load to the modulator whose value depends on a number of details. It is approximately true that this optimal value of load resistance (for a Class B or Class AB push pull modulator) is equal to (2 x the B+ applied to the modulator output tubes) / (the current that will flow through one of the tubes on an audio peak - the current that will flow through one of the tubes with no audio signal applied at its input). In any event, you can look up what this optimal load resistance is in the tube's specification sheet that describes the "typical"/ or recommended operating conditions (B+, biasing, etc.) of the tube when used in a class B or Class AB push-pull power amplifier. According to the RCA specification sheet, a pair of 807's operating in Class AB2, with 750 volts B+ will have an optimal load resistance of 7300 Ohms; and will deliver 120 Watts of audio power (maximum) into that load resistance.

Therefore, your modulation transformer (if you want to get the maximum power out of your modulator) must match a 7300 Ohm source (i.e. the optimal modulator load resistance) to a 10,000 Ohm load (the r.f. output stage modulation resistance).

Note that a 10,000 Ohm rf output stage modulation resistance implies some tough requirements on both the modulation transformer and the Heising reactor.

For example, you want the impedance of the Heising reactor, at the lowest audio frequency of interest, to be equal to (or greater than) the rf output stage modulation resistance. If you want your audio response to extend down to 50Hz, then you require: 2pi x 50 Hz x L = 10,000 Ohms. Therefore, the Heising reactor must have a value of at least: 31 Henries. If you use a 14 Henry Heising reactor, then your frequency response will fall off rapidly below 114 Hz.

Similarly, you will need to use a modulation transformer that has a magnetizing inductance that is comparable to or greater than the inductance of the Heising reactor you use. I don't know if the magnetizing inductance (as seen from the output side) of the modulation transformer you propose to use is high enough. [Note that an Apache's rf output stage operates at a much lower plate voltage and a much higher plate current than your proposed rf output stage. Therefore, the modulation resistance of the Apache's rf output stage is much lower... perhaps around 2500 Ohms - 3000 Ohms, depending upon how it is operated].

Your Heising capacitor should be chosen so that its impedance is equal to 10,000 Ohms at the same frequency at which the impedance of the Heising reactor is equal to 10000 Ohms. If you use a 14 Henry Heising reactor, then the impedance of your Heising capacitor should be 10,000 Ohms at 114 Hz. Therefore, your Heising capacitor should have a value of 0.14uF (much lower than 4uF).

You might want to consider a design for your rf output stage that has a lower modulation resistance (less voltage and more current).

Stu

[KE6DF]

Also  you can only use the mod transformer backwards if the 3000 ohm winding is center tapped. Otherwise you can't use two 807's in PP as modulators.

[Steve - K4HX]

It would seem to me that you really should try to find a 250 (or more) watt modulation transformer.

[AB2EZ]

If you use tetrodes as modulator output tubes... then the impedance, looking back toward the modulator, is extremely high. Therefore, if anything interrupts the plate current in the rf stage (like a negative modulation peak that drives the plate voltage close to zero)... then the current in the Heising choke will have no place to go. It can't flow back through the modulation transformer into the modulator, because (unlike a triode) each modulator tube's output current is almost entirely controlled by its grid-to-cathode voltage and its screen-to-cathode voltage (not the tube's plate-to-cathode voltage).

With 1700 Volts of B+ on the rf stage ... you are going to get some huge voltage spikes across the Heising reactor, and the primary and secondary of the mod transformer if (for example) you over-modulate.

Therefore, you will need components that can handle those high voltages.

With a tetrode-based modulator (essentially a current source), things like spark gaps, negative peak limiters, and proper transmit-standby sequencing are essential.

With a triode-based modulator, there things are somewhat less critical

With a modern audio amplifier (essentially a voltage source) as the modulator, an interruption of the plate current in the rf stage is much less critical... provided you do not disconnect the audio amplifier when you switch to standby.

Stu

[W4RFM]

Wow, thanks Stewart, you have re-aquainted me with some principles I have not thought about for 40 years.  I follow most of what you have stated, and I also appreciate the additional input from HX and IYH.  I agree a different transformer would be easier than trying to make 11K to 3 K work in this situation. 
It appears to me that using the Heising reactor and capacitor senario, that any driver transformer of the appropriate power handling cpacity would work just as well as a true modulation transformer, since it only will have to match impedances at an audio rate, without the burden of d c voltage.  Am I right?

[AB2EZ]

Bob

I'm not sure whether or not the experts in the historical terminology would say that an audio transformer, designed to work at these impedance levels and these power levels would be (by definition) a modulation transformer.

What you need to have is a transformer that

a) Can handle these voltages without arcing
b) Can handle these currents at these frequencies without saturating (even if there is no DC... for a given output voltage: the lower the audio frequency, the larger the B field and the associated H-field). I.e. Bmax ~ 1/the frequency, which implies Hmax ~ 1/frequency.
c) Has low enough leakage inductance and low enough capacitance between windings (and between turns/layers of the same winding) to handle the highest frequencies of interest
d) Has a high enough magnetizing inductance to handle these impedance levels
e) Won't produce "microphonics" due to the high audio frequency magnetic fields

That's somewhat of a "tall order" with traditional transformer core materials. For that reason, I would be inclined to use a professionally designed/engineered modulation transformer to build a traditional plate modulated transmitter.

Note that, if you were going to manufacture these transmitters, and sell them into a competitive marketplace (circa 1940 - 1960)... then you would design the modulator to be capable of producing (into its optimal load resistance) just enough audio output power to modulate the rf output tubes; and you would pick a transformer that comes close to producing a perfect match between the modulator and the rf output stage.

However, since you are only going to make one, and since you are going to use whatever parts that you can get your hands on now, you might want to consider using a modulator with 50-100% more output power capability than you need to modulate the rf output tubes. That way, it is not critical to achieve a match between the modulator and the rf output stage.

Of note, I use a modern audio output amplifier (rated a 700 Watts) to modulate my legal limit triode rf amplifier (less than 500 Watts of DC input power). Thus I have almost 200% of excess audio output power to "play" with. I use a modern ferrite toroidal power transformer (1:26 turns ratio, rated at 1000 Watts and 2300 volts peak on the secondary) that has an astoundingly good audio frequency response, and an astoundingly high magnetizing inductance. I use this in a Heising configuration with a 50H Heising choke to achieve a frequency response (audio input-to-demodulated off air monitor output) that is flat from below 30Hz to above 10kHz. The magnetizing inductance of the transformer is so high, that the total inductance is controlled entirely by the 50H Heising choke. This is a consequence of the astoundingly high permeability of the modern ferrite material. The leakage inductance is astoundingly small, so the H-field is very low. [i.e. the H-field in the core produced by the current on the primary side is almost perfectly canceled by the H-field in the core produced by the current on the secondary side. So... as long as there is no DC in either winding, the transformer behaves essentially as an "ideal" 26:1 step-up transformer]

Naturally, the solid state modulator is not "traditional"... but it sure has worked well over the last 3 years.

Stu

[W4RFM]

When I spoke of driver transformers, I was referring to the S-21 and S-22 UTC designs from yesterday that were rated for 115 and 250 watts respectively. In an old catalog I have, UTC even spells out which plate transformer (of theirs) to use to make it all work.  Oh, if we could only got to the parts store in 1960 again!

I have a working Apache, with the front end audio stage bypassed, to take line level audio from a broadcast processor straight to the 12BY7 driver.  I also have a very nice Viking II, which is still stock, both of these in the 120 watt class. My goal was to build something with my hands that would produce 200 watts or more.  I guess I need to wait for a larger mod transfoer to show itself and build it by the book instead of trying to mix and match.  Thanks again.

[KE6DF]

Keep watching ebay. Mod transformers in the 250 watt+ range come up every month or two. Right now there is a transformer from a Viking Valiant listed, for example. But that would probably not be enough of a step up from what you have now.

Last month I bought a nice condition CVM-4 on ebay (300W).

You will find what you need within three months max is my prediction.

Dave

[W4RFM]

Thanks Dave, this is encouraging. I have not been shopping for a bigger xfmr lately, so I will turn my attention to the "bay" and see what shakes. This is not to say I wouldnt mind buying from a forum member as well.
I thank everyone for their input, and hopefully someday we will communicate thru the air.