New 36W build, guidance needed

[Henrik76]

Greetings all!
I just joined the community, after searching for different designs of 4xEL84 PP amps. Please correct me if I'm posting in the wrong forum.

A little introduction is probably in order:
I'm a mechanical engineer, based in Bergen, Norway. I have built and modded guitar amps as a hobby since the late 90s, although not continuously. I class myself as a medium-experienced hobbyist.

Now, to the project at hand: A smelly beast just fell into my hands. It's a Peavey Classic 50/410 which belongs to a friend of mine. I have been given carte blanche to turn it into whatever I please. The transformers check out OK, but the boards and some valves are toast.
My initial plan was to rip out the boards and do a PTP version of the original power section, coupled to a completely different pre-amp. My reasoning being that the transformers were designed for the original circuit and that they probably wouldn't handle the idle current of something like an AC30 power section and the supply voltage would most likely be excessive.
I have been studying the schematic for the Peavey, which can be found here: https://www.thetubestore.com/lib/thetub ... ematic.pdf
It is a fixed bias design, which I would probably make adjustable. I could go with cathode biasing, of course, biasing slightly cold to protect the valves and transformers, but since the bias supply is already here and the voltages probably suits that approach better, I'm leaning in that direction.
However, and this is what I cannot get my head around, in the Peavey schematic, only one out of two paralleled EL84s has a screen resistor on it! This seems very dubious to me and I cannot see why that would be, unless it's a misprint.
For this reason, and the fact that it burnt itself to a crisp, I'm looking for a different 4xEL84 fixed bias design, which is where this forum hopefully proves helpful.

If anyone knows of a circuit which would probably work with the transformers we have, I would be thankful for any pointers.
Also, if someone knows why Peavey did what they did with the screen supply, I'm eager to know!

Thanks and regards,
Henrik

[Bieworm]

Check out craigs build on the 36W. You might find some useful stuff there. Lots of nice features and an example for great layout and building skills.

viewtopic.php?f=11&t=25648&hilit=36+watt+turretboard

[geoff 1965]

welcome to the forum Henrik,
have a look at the "36W Variants" in the downloads section as well.

[Henrik76]

Thank you both!
I'm soaking up information by the minute. What a great site this is!

Cheers!

[Bieworm]

This is only stage 1. Wait until you are an addict like many here among us

[Henrik76]

This is only stage 1. Wait until you are an addict like many here among us

Hahaha! I was already there mate, but this certainly won't help!

Here's another strange one: Mesa have been doing quite a few fixed bias quad el84 amps, like this one: https://el34world.com/charts/Schematics ... omad45.pdf
See anything amiss? Only the outer pair of power valves have grid stopper resistors. Why would that be? Is that another means to the same end as the Peavey's missing screen resistors? I'm confused here!

Cheers!
Henrik

[Bieworm]

This is only stage 1. Wait until you are an addict like many here among us

Hahaha! I was already there mate, but this certainly won't help!

Here's another strange one: Mesa have been doing quite a few fixed bias quad el84 amps, like this one: https://el34world.com/charts/Schematics ... omad45.pdf
See anything amiss? Only the outer pair of power valves have grid stopper resistors. Why would that be? Is that another means to the same end as the Peavey's missing screen resistors? I'm confused here!

Cheers!
Henrik

Maybe the outer tubes have longer wires to the grids? The inner tubes are 7 and 8, but the outer tubes are both number V9???

[JMPGuitars]

The grid resistors are shared as the tubes are functioning in pairs. If you look again you can see that the grids are connected to each other via a 2k7 resistor.

[Henrik76]

Hi!
I haven't got the schem in front of me at the minute, as I'm on my phone, but what you're saying is exactly what I was looking for, but couldn't find. From where I was sitting, I could see the signal from the PI hitting the innermost set of grids directly, then bleed off to the outer pair through 2k7 grid stoppers. The way I read it, the signal should be split between the pair after that 2k7 resistor, not before. Am I missing something?

Cheers,
Henrik

[JMPGuitars]

Hi!
I haven't got the schem in front of me at the minute, as I'm on my phone, but what you're saying is exactly what I was looking for, but couldn't find. From where I was sitting, I could see the signal from the PI hitting the innermost set of grids directly, then bleed off to the outer pair through 2k7 grid stoppers. The way I read it, the signal should be split between the pair after that 2k7 resistor, not before. Am I missing something?

You're missing the question: "Why?"

The circuit is different than others you might be comparing it to like the 36 watt schematics on this site. If you're going to try and compare them, then compare everything between the PI, the power tube grids, and where / how that goes to ground. Don't focus on a single resistor.

[Henrik76]

You're missing the question: "Why?"

The circuit is different than others you might be comparing it to like the 36 watt schematics on this site. If you're going to try and compare them, then compare everything between the PI, the power tube grids, and where / how that goes to ground. Don't focus on a single resistor.

Hi,
Ok, I'll try to explain my logic and then hopefully, you can tell me where I'm mistaken. I'll focus on the upper pair of output tubes, labeled V7 and V9:
The signal leaves the plate of V5b through the .047 coupling cap. At the downstream side of the cap, it is joined by the grid bias, through a 220k resistor. Signal and bias voltage pass through another 220k resistor and connect to the junction between V7 grid and the 2M2 grid leak resistor. V9 grid is connected to this same junction, through a 2k7 grid stopper resistor.
Although there's probably a sound reason for this, the signal going into the grid of V9 is facing 2700 ohms of resistance, which the counterpart at V7 isn't seeing. We're also adding these 2700 ohms to the 2M2 grid leak resistor between V9g and ground.
To me, it would make more sense if the two grids were tied together through two 1k5 resistors in series and then have the signal, bias voltage and ground reference all connect at the junction between these two resistors, for perfect symmetry.

If you can tell me where my perception is wrong, I'll be a wiser man for it!

Cheers!
Henrik

[JMPGuitars]

To me, it would make more sense if the two grids were tied together through two 1k5 resistors in series and then have the signal, bias voltage and ground reference all connect at the junction between these two resistors, for perfect symmetry.

If you can tell me where my perception is wrong, I'll be a wiser man for it!

Cheers!
Henrik

Why 1k5 in series? How would that make more sense?

I don't know this circuit (other than having owned one in the past), but my guess is that the 2k7 is more to isolate the grids than function as a grid stopper. I could be wrong, but that's what makes more sense to me.

If you want to put grid stoppers on, you can. Besides isolating the tubes, my guess is that they may have wanted a slight imbalance between the pairs. That's also just a guess though. If you know how to work on an amp, you can experiment with that however you like.

I've owned the Nomad 45 and a plethora of other Mesa amps in the past. None of them sounded anywhere near as good as the circuits on this site.

[Henrik76]

Hi!
I was just trying to understand why they would have the paralleled output tubes wired differently. An intentional imbalance is probably correct, which is most likely what Peavey was shooting for as well, with their screen resistors on only two out of four tubes.

In my humble experience, cathode biased EL84s are typically operated at lower plate voltages than what I'm expecting from this power transformer. The board was toast, so I'm unable to get the exact voltages until I've wired it up, but it's a non-centre-tapped secondary, intended for a bridge rectifier and it gives out close to 340vac unloaded. If you think this would work safely with cathode bias, I'm definitely going to try both to see what I prefer.

My idea about the two 1k5 resistors in series was that it would "share" that 2k7 (OK, total would be 3k) between the two grids, splitting the signal evenly between the two. Perhaps, as you suggest, a little out-of-equal is actually preferable.

Thanks for all your help!

Henrik

[JMPGuitars]

My idea about the two 1k5 resistors in series was that it would "share" that 2k7 (OK, total would be 3k) between the two grids, splitting the signal evenly between the two. Perhaps, as you suggest, a little out-of-equal is actually preferable.

Thanks for all your help!

Henrik

No problem.

The two resistors wouldn't actually be in series if they're done that way. In series would be 1k5 + 1k5 going from source to 1 grid. If you're splitting 1k5 resistors with both connected to the source, and then independently going to each grid, they're not series or parallel, they're separate.

Also, think about this: if the 220K resistor is connected close to or directly to the tube socket, why would you need an additional grid stopper?

[Henrik76]

My idea about the two 1k5 resistors in series was that it would "share" that 2k7 (OK, total would be 3k) between the two grids, splitting the signal evenly between the two. Perhaps, as you suggest, a little out-of-equal is actually preferable.

Thanks for all your help!

Henrik

No problem.

The two resistors wouldn't actually be in series if they're done that way. In series would be 1k5 + 1k5 going from source to 1 grid. If you're splitting 1k5 resistors with both connected to the source, and then independently going to each grid, they're not series or parallel, they're separate.

Also, think about this: if the 220K resistor is connected close to or directly to the tube socket, why would you need an additional grid stopper?

Absolutely. They wouldn't be in series as far as the signal source is concerned, but they would make up a series between the grids. I simply thought it would make more sense than having 2k7 on one grid and nothing on the other. The 220k could be connected at one tube socket, but certainly not on both, which may explain why the outer tube has that additional grid stopper and the inner tube doesn't. Hah! Maybe that's why they did it!

[crgfrench]

Mesa has some quirky designs. The Nomad 45 has "Dyna-Watt" power section. The odd setup is probably due to that. The Nomad 55 and Nomad 100 models also have selectable bias so you can switch between 6L6 and one or two pairs of EL34s. That adds to the circuit complexity but is not included in the Nomad 45.

[Henrik76]

Mesa has some quirky designs. The Nomad 45 has "Dyna-Watt" power section. The odd setup is probably due to that. The Nomad 55 and Nomad 100 models also have selectable bias so you can switch between 6L6 and one or two pairs of EL34s. That adds to the circuit complexity but is not included in the Nomad 45.

Thanks for the tip! From what I'm reading, the Dynawatt scheme is mainly a power supply thing, although it might possibly extend to some parts of the power section.
Anyway, I have no intention of building a Mesa clone. I was simply looking for general ideas and existing circuits for a 4 x EL84 PP power amp with fixed bias, since I expect the supply voltage will be significantly higher than ideal for most of the tried and tested VOX/Matchless-type circuits. I will not rebuild the original Peavey circuit, so I was looking for alternatives. I think what I'll do is wire it up so that it sort of makes sense and then tweak it from there to get the tubes running reliably cool. I might make it switchable cathode bias, but I think some of the magic of a cathode biased EL84 amp is that they tend to run the tubes quite hot (this may well be a misconception, but then it's widespread!). This PT is on the small side, so I'm not so sure it will handle the high idle current of that type of circuit. It's certainly nowhere near as hefty as a 30W Westlabs PT I have in my stash, or any of the AC30 trannies I've seen, but then again, these amps have tube rectifiers and maybe twice as many preamp tubes.
I'll wire it up, bias it sensibly and keep an eye on the transformer temperature and how much the voltage is pulled down by the circuit as I increase the idle current.
I'm in the middle of moving houses, so I won't get properly started for a while yet, but I'll be sure to post when I do.

Cheers,
Henrik

[Henrik76]

20201213_015103.jpg

20201213_015050.jpg

Gutted chassis getting ready. PCB tube boards replaced by folded sheetmetal with mounted sockets. I'm putting the two first-stage tubes on a separate shock-mounted plate up near the inputs.
Stand by!

Henrik

[Henrik76]

20201213_204355.jpg

20201213_204349.jpg

20201213_204108.jpg

Details of the shock-mounted pre-amp sockets. One will be a pentode, which is why I did it.
I'll try to get the power supply wired up tomorrow.

Henrik

[TriodeLuvr]

I just want to mention that the design represented in the schematic regarding the grid stoppers is incorrect. I suspect a misprint in the schematic, but regardless, there's no valid reason for the existence of stoppers on only two tubes. Also, to be effective at eliminating oscillation, stoppers need to be soldered directly to the grid pins with the shortest possible leads.

Don't be tempted to use cathode bias as mentioned in your first post. Cathode bias applied to amplifiers operating in other than Class A causes an increase in DC bias with increasing output. This isn't always obvious when a large signal is present, because the loudness of the material masks lower level artifacts. However, when the signal suddenly goes quiet, the output tubes are left in an under-biased state for some brief period of time. The result can be a short term presence of audible crossover distortion.