EL34 Nostalgia

[Cressy Snr]

You try tellin’ that to kids today?
They’ll not believe yer!

[ed]

don't you really envy those yanks, they've got 60!

[andrew Ivimey]

Well they would wouldn't they!

[Mike H]

Yer daft lot .....

[Cressy Snr]

After several hours of work, I've managed to produce a proper diagram of the whole amplifier.
I actually rebuilt the audio circuit and modded the power supply of the E-Linear amplifier to make this work. The E-linear was a great sounding amplifier, don't get me wrong, but I am unconvinced about the triode driver stage. Because of the nature of the feedback, I reckon that for best results it needs a hi-gm pentode as the VA/driver stage. Two of the D3a I got from Nick, look to be the ideal candidates for the job. So in the medium term that is what I will look at for E-Linear operation.
Meanwhile, the 'Nostalgia' circuit is sounding beautiful, so will be my 'go-to' valve amp when I want a good helping of tone.
I found Mullard, Brimar and Philips Miniwatt ECC83s in my boxes, so that's useful. I also found two Psvane and an Electro-Harmonix, so not a bad stash.
Here is the schematic, with the PSU included on the same piece of paper:

EL34NostalgiaAudiocct&PSU.png (107.51 KiB) Viewed 3050 times

[Nick]

You are putting big (150R) resistors in the power supply again. Just saying.

[Cressy Snr]

You are putting big (150R) resistors in the power supply again. Just saying.

I don’t see what else I can do. I am severely constrained by the size of the chassis, and the chokes I have, have an even higher DCR than the resistors and are only specced for 50mA so would be saturating at the current the EL34s are pulling. The corner frequency for that stage (150R with 100uF) is 10.6Hz which may or may not be a problem; I don’t know, and that is perhaps the real problem.

It is highly probable, nay almost certain, that I misunderstood something fundamental, twenty years ago about power supplies and have carried it with me, like a lead backpack for all that time, making the same mistakes time and again and not understanding why I’m mistaken, or what I should be doing differently to resolve such issues.

This I’m afraid is what is so disheartening about this game, or to put it more accurately, the maths involved with it. I don’t know. Maybe I learn something from a book, apply it diligently, carefully working out the answers, but the maths doesn’t fit the situation, so that’s not what you need. I’ve worked hard, but not smart. This is one reason why I decided to stop with all this, get the NVA thing out, tweak it up and be satisfied with that. I’d just had enough of making no progress.

For example, Morgan Jones goes deeply into cascaded RC section filters to remove noise efficiently from power supplies, referring to ‘Cathode Ray’ and his tables giving suggestions for the minimum number of sections that will achieve a noise level of -ndB. Is this what I’m doing here? latching onto something, then using it in the wrong context?

All I know is that the supply is essentially silent. I cannot hear any hum whatsoever at the speaker, even with an ear inside the bass driver.

It was the same at school. Teacher would teach, I would think I was diligently carrying out the instructions, be pleased with a job well done, then the work would come back covered in red ink. When, quite naturally, I asked what I’d done wrong, more often than not, the answer would be, ‘you’re at a Grammar school lad! If you don’t get something this simple, I suggest you take yourself off to the Secondary Modern tout suite!’ usually said in front of a class full of sniggering kids.

So….I need to understand what is wrong with the supply I’ve built and why? because clearly I’m using learned information that is not applicable to the situation and not understanding why it doesn't apply at that point in the circuit, but somewhere else it does.

I could do with getting this valve amp power supply business sorted out in my head once and for all, or I’ll simply give up again, which would be a shame, as this particular amplifier really does stir up deep emotions. In that respect at least, nothing else has ever come close.

[Nick]

Low noise is important, but it matters to the driver stage where the gain is higher and the PSRR. The output stage needs power, it needs access to current, the resistor gets in the way of that. The thread that wandered into regulators and sag is a example of that and why it matters. It may be fine with 150R, but it may be better with less, that's all.

[Cressy Snr]

So a simple C filter at that point might be a better way to go?

[Nick]

Maybe, just I would try.

[Cressy Snr]

Maybe, just I would try.

So as to make sure I understand this: the 10K/47uF filter is perfectly in order for the ECC83, because in conjunction with all the previous filtering, you get low noise and there is little or no penalty from the 10K resistor because of the very low current being drawn off at that point. However, the 150R/100uF in the higher current environment of the power stage section, is basically raising the supply impedance, which puts an obstacle in the way of the power stage doing what it needs to do?

If the above is right, I can't believe I'd not cottoned on before now. (stop laughing)

[IslandPink]

Yes, exactly.

[Cressy Snr]

Yes, exactly.

I'd better go stand in the corner and face the wall.
This is fecking embarrassing.

[IslandPink]

Just try a simple PSUD model, and see how it reacts to a step change in current load.
First case, use your c,L then a 100u cap to ground.
Second case, as your circuit, with 40u and the extra 150R,100u.

[Cressy Snr]

I’ll do that later Mark.
I always suspected I’d missed something fundamental about power supplies. The wonder of it is, I got this far down the line without it.