Possible 10% increase in efficiency in class C tube output stage (easy)

[IN3IEX]

Hi, I recently found a very interesting article on a RCA AM broadcast transmitter.

http://nrcdxas.org/articles/bta5t/

The method for having more than 90% efficiency was discovered after the development of most amateur AM transmitters, therefore you do not find it in boatanchors.
This is a kind of class F output stage.

I verified with SPICE that the claims are true and I found a very simple mod to the pi network of every class C tube transmitters to increase the efficiency.

http://www.ing.unitn.it/~fontana/SuperCeng.pdf

At the moment I cannot test this mod. Let me know if it works....

Thank you
Our tubes must win over class-E MOS regarding efficiency !!

73  Giorgio IN3IEX, a Drake 4 fan.

[k4kyv]

Gates used that same circuit in one of their AM transmitters during the early 60s as well.  The problem with amateur use is having to re-tune the two extra resonant circuits whenever you QSY, unless they could all be ganged together somehow to accurately track using one tuning knob. That should be relatively easy with one band, but very difficult across multiple bands using plug-in coils, and nearly impossible in a band-switching transmitter.

[W2PFY]

There is one station operating as I type running such a circuit. It is a 5 KW RCA transmitter manufactured in 1948. It is the main transmitter backed up by a Gates 1 KW transmitter. The call sign is WXME and is owned by the same guy who owns WBCQ SW. Tim, WA1HLR has often talked with me about doing such a project and he has done it to the above transmitter.The results of the modification have proven an increase in efficiency and power output.

This station is probably the only one in the world operating with the circuit nowadays. I think I'll try to get Tim to comment on this thread. He is not a big poster or email guy

[IN3IEX]

Gates used that same circuit in one of their AM transmitters during the early 60s as well.  The problem with amateur use is having to re-tune the two extra resonant circuits whenever you QSY, unless they could all be ganged together somehow to accurately track using one tuning knob. That should be relatively easy with one band, but very difficult across multiple bands using plug-in coils, and nearly impossible in a band-switching transmitter.

Hi Don, in fact the idea is to use only the anode circuit that can be "integrated" in the existing Pi inductor. The simulation shows that most of the gain in efficiency comes from the anode circuit and that the resonant circuit can be part of the existing Pi inductor. Therefore only an additional variable capacitor is required and for multiband operations an additional ganged band switch.
With the additional capacitor set at minimum capacitance the circuit is "almost" excluded. Maybe dual band operation is possible without the switch.
Anyway field test is essential...

[w4bfs]

Hi Georgio ...thanks for taking time to pass the RCA ad and your simulations along ...73 ... John

[Tim WA1HnyLR]

 I finally stumbled into this post. I modified Allen Weiner's ancient RCA 5-F transmitter to the third harmonic excitation circuit. The original 892-R tubes had been replaced with 5762 tubes. The transmitter had been on 710 Khz. The station was moved to 780 Khz. I rebuilt the tank circuit using very heavy duty components. The icing on the cake was the implementation of the third harmonic resonator circuits . Basically parallel tuned circuits in the plate and cathode of the PA stage. Typical operating parameters: 1.05 amps plate current @5Kv make 5Kw outpoot. I could not believe it. I ran the transmitter for a while then shut it down. I opened up the door to the PA compartment and held my hand over the exhaust stream above the PA tube. The air was piss warm. A testimony to the system. Tim WA1HnyLR

[W2XR]

I finally stumbled into this post. I modified Allen Weiner's ancient RCA 5-F transmitter to the third harmonic excitation circuit. The original 892-R tubes had been replaced with 5762 tubes. The transmitter had been on 710 Khz. The station was moved to 780 Khz. I rebuilt the tank circuit using very heavy duty components. The icing on the cake was the implementation of the third harmonic resonator circuits . Basically parallel tuned circuits in the plate and cathode of the PA stage. Typical operating parameters: 1.05 amps plate current @5Kv make 5Kw outpoot. I could not believe it. I ran the transmitter for a while then shut it down. I opened up the door to the PA compartment and held my hand over the exhaust stream above the PA tube. The air was piss warm. A testimony to the system. Tim WA1HnyLR

Hi Tim,

Unless I am missing something here, this translates to >95% output stage efficiency (DC plate input/RF power output).

This is spectacular efficiency, which I did not believe was achievable in any class of analog amplification (even class C), with the possible exception of the digital modes of power amplification.

Can you verify that this was indeed the level of performance you were able to obtain?

73,

Bruce

[W8IXY]

Back in the 80's, I took care of several 10kw and 50kw Harris MW transmitters that had those 3rd harmonic resonators, that squared off the RF waveform and increased the PA efficiency.   Those were PDM transmitters, but when those resonators were adjusted correctly, we measured a greater than 85% efficiency in the output stage of both the 10kw and 50kw transmitters.

We also found that with the PDM transmitters, you needed to tweak those resonators "just right" to minimize distortion on the modulated waveform.  We'd get the adjustments close by observing the PA current, then feed 1000 Hz near 100% modulation and with the distortion analyzer on the output of the mod mon, we'd set the THD and IMD (actually the best "sound" was achieved adjusting for minimum IMD) for best performance.

We once had the WWWE Harris MW50 (now WTAM) transmitter here in Cleveland, running at >85% PA efficiency, 1.2% IMD and 0.9% THD.   That was a really sweet sounding transmitter after the adjustments.

Nowadays, the high power AM stuff is all solid state, digitally modulated.   In the old days, for 50kw, we'd read 9500 volts at ~6Amps on the PA.  With the solid state transmitters, PA voltage on a Harris DX50 is about 230 volts and PA current idles about 230 amps, and kicks up to over 400 amps with 100% mod on a sine wave.

73
Ted W8IXY

[Bill, KD0HG]

Years ago, I tried that 3rd harmonic trick on my 304-TL rig.
Driving the snot out of the tube in a conventional pi-net gave me around 85% measured efficiency, the 3rd harmonic trap added, at best, two or three percent.

The thing is, when you drive a triode hard, the plate waveform does approach a square wave, being switched hard between "on" and "off".

I suspect this 3rd harmonic wave shaping scheme works the best when drive to the Class C stage is fairly low.

Bill

[W2PFY]

The transmitter that Tim modified has a 5 KW plate dissipation triode tube in it and it was driver with a pair of 810's. That has been changed to a 304 tube.  I think the tube requires about 200 watts of drive.