I am trying to tune/tweak/repair this transmitter for a friend of mine. It is oscillating, and it has rf power on the antenna, but it just doesn't seem as "strong" as it should be.

The diagram says 35-40 watt carrier. Now its a class A final amp, so I should be expecting half (or less) of that carrier wattage, correct? It came to me with a VFO, Hartley oscillator which was very unstable, so I replaced it with a Colpitts crystal for testing purposes. (I changed the diagram as well.) The only other thing I have added is a 40uH RF choke in the plate supply line to the 807s. I could not get a clean carrier without this choke.

As far as I can tell, everything is working. It oscillates, has RF at the buffer EL84 stage, and has RF at the 807s.

I tune for the "dip" in 807 plate current, which then has the highest RF on my diy field strength meter, and here are the 807 measurements: Plate voltage: 500-525v Plate current: 90mA No signal Screen voltage: 115v No signal <-- this seems low to me. Grid voltage: -14.7v No signal <-- also seems low. (not negative enough)

The antenna is a 1 wave long wire. 25AWG magnet wire, about 10ft off the ground, no counterpoise. RF ground is a foundation stake pounded into the earth about 4 ft. (and it's raining here. ground is wet.) I was using an antenna matching large variable inductor with multiple taps and fine tuning, but it seems to be "robbing" RF. It has a large neon bulb that glows brightly when it's hooked up though.

Is there anything fundamentally wrong with this circuit? Any red flags looking at the schematic? Lack of RF chokes? RF feeding back into the power transformer? Does the EL84 oscillator buffer need a tank circuit on the plate to tune to resonance?

Any help would be appreciated! Thanks.

enter image description here Here's a link to the 807 class A specs I've been using:

enter image description here

Dwayne Reid: Thanks for your help. When I pull the audio EL84, the 807 plate current shoots up to 160mA, screen voltage increases to 204v, grid voltage stays the same, -14ish. The field strength of the antenna increases as well.

  • 3
    \$\begingroup\$ Dropbox wants me to have an account before I can look at the picture. And yes, I know it's "free", but it's free as in "have some white powder to stick up your nose". \$\endgroup\$
    – TimWescott
    Mar 6, 2019 at 17:48
  • \$\begingroup\$ Haha, sorry Tim. I think I fixed it now. I forgot how to add pictures to the question, but I think I figured it out. \$\endgroup\$
    – McMurdo
    Mar 6, 2019 at 18:22
  • \$\begingroup\$ What happens to your output power and the 807 screen grid & control grid voltages if you pull or disconnect the EL84 modulation amplifier? Please edit / update your question if you can make those measurements. \$\endgroup\$ Mar 6, 2019 at 18:51
  • \$\begingroup\$ Thanks Dwayne, I added the results in the question. \$\endgroup\$
    – McMurdo
    Mar 6, 2019 at 19:07

2 Answers 2


I think this is normal operation for a screen modulated AM linear amplifier. The output of the amp for an unmodulated carrier should be 1/2 power. Applying full amplitude audio should cause the 807s power to cover the range of zero power to full power. The average power is still half the maximum power.

By being a class-A amplifier, a resonant cutcuit on the output isn't needed.

  • \$\begingroup\$ Thanks for the reply. The output signal/range just seems so weak. I recently worked on a two tube, single 6L6GC final class A transmitter and I think it had better range than this one. Could be just the way it is though. \$\endgroup\$
    – McMurdo
    Mar 7, 2019 at 2:03
  • \$\begingroup\$ You said the antenna was a full wave wire. Doesn't that present a very high impedance at the feedpoint? Can you substitute a 50 ohm dummy load? \$\endgroup\$
    – cmm
    Mar 7, 2019 at 2:23
  • \$\begingroup\$ I have read that a full wave can present a very high impedance to the transmitter, unfortunately I have no way of measuring that. I have tried an antenna matching “L” network at the feedpoint, but it seems to have less RF on my field strength meter when the L network is connected. \$\endgroup\$
    – McMurdo
    Mar 7, 2019 at 2:52
  • \$\begingroup\$ I'm not sure the field strength meter is useful in this case since there are many components of the field, some helpful, some not. Do you have an oscilloscope available? Thst will let you measure (through a voltage divider) the actual plate voltage change in the idle and modulated conditions. It may be that you need more grid drive. \$\endgroup\$
    – cmm
    Mar 7, 2019 at 5:35
  • \$\begingroup\$ I agree that from the spec the grid should be more negative. In the circuit, it looks like it depends on the a \$\endgroup\$
    – cmm
    Mar 7, 2019 at 5:37

I don't know why you think your parallel 807's are a class A amplifier. In my ham radio opinion, they are a class C amplifier. From your stated plate current and voltage, the power is about 45 watts, so you should be able to deliver about 35 watts to a well-matched load (because nobody is 100% efficient, not even me).

However, you also have screen grid modulation. That is a cheap form of modulation that does not require a modulation transformer. It works by cutting the power way down, not by adding sideband power the way plate modulation does and is probably not capable of 100% modulation. So the 35 watts is your PEAK power when modulating, not your average power. That's OK, that's what my Heath DX-35 had in 1958 for voice (but I only did CW in those days).

Another issue is your measurement of the 807 grid voltage. You need to be aware that this requires a very high impedance voltmeter. The value of the grid leak resistor is not given in the schematic, but I am guessing it is high. In fact, if it is dirty or has leakage paths, its lowered resistance could impair the performance of the final amp. Your grid voltage may be higher than you think. You need a high impedance voltmeter to correctly measure it.

With your rig on my workbench of 60 years ago, I would also check that the RF from the EL84 buffer is strong enough to switch the 807's on and off. Better transmitters of that era provided a way to measure grid current to verify that sufficient drive was there. You are not doing any frequency multiplication here, so I don't think there is any need for a resonant circuit or filter between the EL84 and the parallel 807's.

Good luck finding rocks and replacement jugs for that beast!

  • \$\begingroup\$ Hi Richard. Thanks for the clarification on the amp. I just assumed it was class A. I see now that you're correct. How would I go about checking the RF from the EL84 buffer? The grid resistor that is missing the label, I'm currently using a 3300 ohm. The meter I'm using to measure the grids is a Fluke 187. Probably not a high impedance meter. What do you think about the low screen voltage on the 807s? Even in class C, I see they're supposed to be 300v. Thanks for your help! Much appreciated!! \$\endgroup\$
    – McMurdo
    Mar 8, 2019 at 6:40
  • \$\begingroup\$ If this were a class C amplifier, wouldn't the grid need to be more negative so that the tube would be in cutoff most of the time? The Class-A theory seems reasonable. \$\endgroup\$
    – cmm
    Mar 8, 2019 at 19:50
  • \$\begingroup\$ The screen grid voltage modulates the output. Most likely the PEAK screen voltage is near the 300 volts you expect. If you want to play with grid bias, look at old transmitter schematics in QST from the 1950’s. The 807 was the most common power amp jug until the vastly superior 6146 and the TV sweep tubes (for SSB) came along. I used to get rocks from C&H in Pasadena and grind them into the 40 meter band. If you are into retro tech, I have a slide rule you might like, and I am working on an improved log table. \$\endgroup\$ Mar 9, 2019 at 12:50
  • \$\begingroup\$ Richard, I played with the grid resistors yesterday, added a 12k (output stayed the same) swapped for a 20k (output went down) swapped for a 10k (output went up slightly from the 12k), the specs for a GL-807 call for a 12.8k in class C. No real change in the range of the transmitter. The crystals I bought were the old HC-6 style from Surplus Sales of Nebraska. \$\endgroup\$
    – McMurdo
    Mar 9, 2019 at 14:49
  • \$\begingroup\$ cmm, I think the large pi network in the output is the telltale that it's a class C? I wish I could get a hold of the author of the diagram. They don't answer my emails though. \$\endgroup\$
    – McMurdo
    Mar 9, 2019 at 14:52

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.