le Chargeur (yet another EF86 AC4 kind of amp)

[tictac]

Hope you can see this....

TT

[ChopSauce]

Well, as its list of revisions grows, so does my temptation to rebuild this amp in a "normal" chassis - with an improved layout

I have a 50HZ oscillation, which develops on idle, with anything plugged in & no signal on input. Not sure what it is

I'm pretty sure any harshness issue - as limited as it is - comes from the speaker, though. The EL84 grid voltage seems to be well below the limits, around -7 Volts.

[ChopSauce]

For the sole sake of completion:

I solved the above mentionned noise issue using an "anode mixer" wiring scheme for the inputs - instead of the resistance network "à la" Fender Greta.
(now I suspect that the noise caused by this kind of resistor network might be a main responsible for a choke being used on the Greta)

Hope you can see this....

Yes thanks!
(I added it to my library of schematics )

[ChopSauce]

While waiting for the pair of 100R resistor to complete the build of a stock AC4 - from the sixities - I powered that thing again and found it all "whooly-muddy"...

I went through these advices (*)
(here: https://ampgarage.com/forum/viewtopic.php?f=6&t=19130 )

... A mini Vox mated to a Celestion Blue would be killer. Done right, you'd get the chime and grind in one package. To the design I might add:

1) ...

4) You might need to reduce the size of the bypass caps on V1a and V1b. 25uF might be too woofy and overload the next stages. It won't take much to drive that EL84 into the zone. 0.68uF - 2.2uF (or maybe 4.7uF) should be fine. Do not worry about not having enough bass. It will be plenty of low end. You want clarity and punch not woof, blat, and mud.

5) If you go with adding the 1/2 phase inverter, you show the 220k in series with 50-something on the input (not sure what value that is meant to be). Consider adding another 220k to ground to create a voltage divider to simulate the load of the 'other' channel and bleed some signal going to the fake PI. There is going to be loads of gain, no need to overdo it.

6) The 220k grid leak on the output could be made into a 250kA pot for a master volume if desired. You might find that 100k is plenty of drive for the EL84.

7) Consider changing the 220k plate load on the fake 1/2- PI to 47k (more headroom). You don't need to crush the EL84, it takes very little to drive it into ugly blocking distortion (add that grid resistor).

8 ) Consider 7-9k primary for an OT!

9) Use a HIGH quality 1M metal film resistor on the input (1-3W). Any noise going in gets louder.

10) Consider adjustable cathode bias on the EL84 for more flexibility.
...

And decided to lower all the cap values on the early stages as well as the ridiculously high values that I have been recommanded elsewhere for the grid resistors. That and the volume pot but it didn't helped much. I even swapped the tubes - but the EF86. I'll try also that tomorrow, as everybody's asleep now.

Anyway, here is the outcome to date

z4.jpg

I look forward to read your comments possibly but I am currently tempted to gut it and build something else with these 3 tubes, and possibly the Mosfet - whose main interest is to drain the caps in a few seconds - if it can be used as a normal triode.
_____

(*) Plus - as a note to self:

So you would do away with the ground bar for the pots and instead ground them to a corresponding position on the ground bus?

Yes, I would do away with a pot ground bus bar. It 'works' but it's also a possible way to introduce a ground loop IMO and there's better grounding methods. I would route your pot grounds (one for Volume and one for the Tone Stack) back to the respective nodes that they serve.

By node, I mean B+. Each section of an amp has an associated B+ filter cap. Route grounds within each of those sections back to their respective B+ filter cap negative/ground, "closing the loop".