• WARNING: Tube/Valve amplifiers use potentially LETHAL HIGH VOLTAGES.
    Building, troubleshooting and testing of these amplifiers should only be
    performed by someone who is thoroughly familiar with
    the safety precautions around high voltages.

EL84 rebuild suggestions

This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Hello All,

As I have completed my 6L6 monoblocks, I have decided to play rebuild/modify my stereo EL84 amp. I have a chassis with 1 6N2P driver and 2 6P14P tubes per channel. Power supply is around 275 volts. I have 6.6K output transformers.

I am looking to modify what I have without having to adjust the power supply too much (could probably get a few more volts out of it by reducing some resistance) or re-drill the chassis to take more tubes. As the driver tube is wired for 6 volts, small bonus for keeping heater wiring, but modding to take a 12AT(X) is no big deal. I have ECC81/12AT7, 6N1P-EV, 6N2P-EV, 6N23P-EV small signal tubes on hand. Do not want to purchase any more tubes at this point if I can help it.

I have CCS circuit already in chassis, easily modded for variable current. My recent success with the tubelab designed diff amp driver for my 6L6 amp has me interested in pursuing a simplified version for my EL84 amp.

The -15 volt supply is in place for CCS, so easy to rig up for fixed bias.

This is roughly what I have in mind. Feedback value just a rough pluck, have not calculated yet.

Worth pursuing?



An externally hosted image should be here but it was not working when we last tested it.

Uploaded with ImageShack.us
Another option is use 1 triode for a gain stage with a CCS anode load and LED in the cathode with a small (100 ohm) resistor below the LED for feedback. Then use the second half of the dual triode as a concertina phase splitter.

That said, what you have looks interesting as well.
Do you really mean to have 100% feedback (unity gain)? If so, stability could be a big issue.

If you're running EL84 as pentode, you can make the grid stoppers MUCH bigger (30-50k), which will help clipping and recovery performance.

Thanks Sy,

Are you referring to the value of the feedback resistor R15? I should have left that value blank, as it is not calculated. Will consult the 'Good Book' (MJ) to work my way through that calculation.

I have not built a pentode amp before, only triode and U/L. Will take your advice re grid stoppers for the pentode. Should also re-read your Red Light district pages, as I recall it is pentode, but different drive and bias setup.

Point taken re concertina splitter, not ruling it out, just that CCS circuit is there already set up for cathode, and I have been impressed with the diff amp in my 6L6. From what I have seen there appears to be Baby Huey and El Cheapo using diff amp and MJ Bevuois Valley, Simple PP and RLD using concertina, so either look like good options?



Concertina or diff amp, both work well. If you have the CCS sitting there and like the diff amp then go for it. The schem looks good to me, just need a couple minor changes. R15 and R14 make a voltage divider for the feedback. R14 needs to be really small. This is what SY was hinting at. Maybe R14 500 ohms and R15 10K for a WAG, to start with. Consider with say 10V at the output transformer, the divider as shown would be feeding around 7V to the 2nd grid, way too much. Important thing is that R14 is reasonably small as it is trying to hold the grid at ground potential, otherwise the diff amp will not work so well.

The resistors R7 and R8 might need to be made smaller, check the datasheet, often for fixed bias the recommended values will be lower than those for cathode bias.

I would make R1 100K instead of 1M, no big deal though. R4 and R5 might need to be made bigger in value to get enough gain, I would start with around twice the value shown, but you could start there and see how it goes.

Its fairly difficult to get a bad sound out of a 6N1P-EV IME, I like them a lot. Don't be too fussed about GNFB at first, you may find it sounds great with none at all. Or maybe not. With B+ 275V the odds are in favour of no feedback IMO.

30K to 50K grid stoppers? Never heard that before, maybe SY you meant 3 to 5K?

Your ears are your best friend with a design like this :) I don't think you'll be disappointed.
Thanks Ian,

Your comments make sense regarding feedback with R14 and R15. What is the general consensus for feedback required for EL84 in pentode?

Regarding R4 and R5, the plate load resistors, I was worried about ending up with too few volts at the plate. I might up them to 47K and see how that looks...

Bingo on the grid leak resistors, 300K max on the datasheet, missed that, thanks.

I have also added a decoupling capacitor prior to the plate load resistors.

Turns out I had a 5U4GB in the power supply. With a GZ34, it is 300 volts from the last filter section, 328 volts after the choke at the second cap (have CLCRC).

Thanks for the assistance, much appreciated!

Last edited:
30K to 50K grid stoppers? Never heard that before, maybe SY you meant 3 to 5K?

Nope, I really meant 30-50k. The RLD article goes into some detail about this and the feedback calculation, both of which are applicable here. Agreed that an LTP splitter with CCS loading will work perfectly well- it's not what I did, but it's another perfectly valid design choice.

SYclotron Audio The Red Light District
Thanks for the help Sy and Ian. I went through the feedback calculation in MJ. I am assuming that as the feedback resistor does not form part of the cathode bias, it is simply the ratio of feedback resistor and the resistor going from grid to ground (R15:R14). Assuming 500R for R14, using Jones I get about 3K. Erring on the side of too sensitive rather than not sensitive enough, round up to about 3.6K. Comes close to Kentucky Windage method of RLD, though the input stage will be less sensitive in this design...

Depending how busy things are around the house this weekend, might get a few hours to wire this up.

Cheers, and thanks again,


An externally hosted image should be here but it was not working when we last tested it.

Uploaded with ImageShack.us
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.