Self Split Push Pull - Kicking this dead horse one more time.

[Ronquest]

I've been looking up circuits and reading past posts.  Has everyone pretty much given up on the self split (SS)?  Is there a "better" SS circuit to use?
Looking at 6v6 or EL84.  Not concerned about max power output.  Does the distortion on the SS sound bad for guitar?  If used below clipping is there much difference from a cathodyne PI?

Circuit picture for reference.

[shooter]

take it to the glue factory already
I quit wondering and started BB'n because "what sounds good to me, don't sound good to you"
Not my board, but worth stealing

[66Strat]

Magnatone 109 Melodier

https://www.youtube.com/watch?v=tA7q0FWY4Q0

The schematic is documented in the Magnatone schematics archived on this web site.

[pullshocks]

Self-split came up several times on the old AX-84 forum.  Many of the posts are available on the Guitar Amp Builders forum which is now hosting the old content:


https://perpetualinnovation.us/MyBBDir/search.php?action=results&sid=d280959e974775ae66829523830c923d&sortby=&order=desc

[HotBluePlates]

... the self split (SS)?  Is there a "better" SS circuit to use? ... Does the distortion on the SS sound bad for guitar?  If used below clipping is there much difference from a cathodyne PI?

Harder to drive than it needs to be.

This guy built a "fet-a-dyne."

[Jonas]

Ronquest - I will be experimenting with adding a 12AU7 (or 6SN7) self split to an amp I have on the bench. I am going to start with this schematic.

[Williamblake]

I have tried the firefly self-split 12AU7 output and i think it works very well. I am using it as a reverb driver where you could also use parallel 12AT7 triodes and i like the selfsplit version very much. When i connected it to a speaker i was surprised how loud it is. But i am still breadboarding with it.

[Jonas]

@Williamblake I'm planning on getting parts to experiment with self split 12AU7 soon. Curious to know what your voltage on the 12AU7 plates? Are you finding the snubbing cap to be necessary?

@Ronquest sorry for the post hijack

[SnickSound]

I've been looking up circuits and reading past posts.  Has everyone pretty much given up on the self split (SS)?  Is there a "better" SS circuit to use?
Looking at 6v6 or EL84.  Not concerned about max power output.  Does the distortion on the SS sound bad for guitar?  If used below clipping is there much difference from a cathodyne PI?

Circuit picture for reference.

One issue is you can't bypass Rk, so gain is pretty low. It only makes sense when you're counting pennies and every saved part helps.

I built a Firefly a while back, like many people did, and found it so boring and sterile (fun little project, but a curiosity at best once finished). Recently went back to an ECC99 build and threw a cathodyne at it instead... Oh my, so much better.

[Williamblake]

When i last had time to dwell on this it was 295V DC on the 12AU7. I must add that i liked the power it  gave on the speaker to check things and thought it could be a good reverb driver. I did not tweak it as an output stage and it was too clean as an output driver to my ears. Maybe i was spoiled by PRR saying rock'n roll likes pentodes. The attached documents is my current state of affairs and i am still going to give this a fair try, didn't have the time to take it further the last year but it worked fine without caps across the output transformer. I did add a light bulb in series to the reverb spring for limiting and i must admit for fancy looks, too. 

[sluckey]

One issue is you can't bypass Rk, so gain is pretty low.

The Melodier circuit will allow you to use a bypass cap.

     https://el34world.com/charts/Schematics/files/Magnatone/Magnatone_109_melodier.pdf

[SnickSound]

One issue is you can't bypass Rk, so gain is pretty low.

The Melodier circuit will allow you to use a bypass cap.

     https://el34world.com/charts/Schematics/files/Magnatone/Magnatone_109_melodier.pdf

Oh I missed that detail, so the screens are used instead to modulate the 2nd tube.

I'll have to breadboard this when my prototyping platform is ready. I'm curious.

[sluckey]

Oh I missed that detail, so the screens are used instead to modulate the 2nd tube.

I'll have to breadboard this when my prototyping platform is ready. I'm curious.

I'm curious too. Please let us know the results.

[SnickSound]

Oh I missed that detail, so the screens are used instead to modulate the 2nd tube.

I'll have to breadboard this when my prototyping platform is ready. I'm curious.

I'm curious too. Please let us know the results.

This will be a great first project to test this massive thing when it's ready :)

[passaloutre]

Where did you get those tube PCBs?

[SnickSound]

Where did you get those tube PCBs?

Designed them myself in Kicad and had them made from JLCPCB

This whole project started with wanting to get into custom PCB builds. Wanted to start with something small, and also wanted to test my circuits carefully before committing them to print.

If there's interest I'll post a separate thread to discuss it. Still waiting for some hardware to finish it up.

[SnickSound]

I've been looking up circuits and reading past posts.  Has everyone pretty much given up on the self split (SS)?  Is there a "better" SS circuit to use?
Looking at 6v6 or EL84.  Not concerned about max power output.  Does the distortion on the SS sound bad for guitar?  If used below clipping is there much difference from a cathodyne PI?

Circuit picture for reference.

So this was a fun way to test my prototype bench! Very simple circuit to start with.

I've built the simplest form of self-splitting right now, where I simply connect the 2nd grid to ground (cathode resistor unbypassed). It sounds clean and full at low volume, gets fuzzy as you crank it up but there's no odd artefacts.

It's weak though! Best I got was 6VAC into an 8ohm load, which is not even 5W (don't ask me about THD% though haha, had to "earball" it due to not having an oscilloscope)

That's with over 400V on the plates (working with a set of Deluxe Reverb transformers, it's all I had on hand).

I don't have the tools to measure if it's the preamp running out of headroom too quick, but I have over 230VDC on the preamp plates (with 100K plate load), should be enough.

I'll play around with it some more and then move on to this variation with the 2nd grid connected to the first screen. I'm working with way more than 300V though so hopefully this works even if I'm running the screens off a 2nd power node.

Until then, the conclusion is: it sounds perfectly fine when running clean, it's just very inefficient (single ended could get just as loud... so far)

[SnickSound]

The explanation for the low power was right there under my nose: self split cannot operate correctly in class A/B. Once the first tube goes into cutoff, the 2nd tube just hangs at max current until the 1st tube comes back in.

Quick math says this explains it.

My set of transformers isn't appropriate for Class A operation (voltage too high).

So there's another limitation with self split: limited to Class A operation which reduces available power dramatically (I'm assuming the max would be about twice what you get from single ended, so about 10W from a pair of 6V6)

You also get lopsided distortion characteristics, with the positive signal side going into saturation but negative always clipping as the first tube hits cutoff. That would explain what I perceive as a fuzzier sound, quite reminiscent of single ended actually. Unlike "proper" push pull which cancels even harmonics, this is full of them.

[HotBluePlates]

The explanation for the low power was right there under my nose: self split cannot operate correctly in class A/B. ...

So there's another limitation with self split: limited to Class A operation which reduces available power dramatically ...

And because you're using one tube to develop the drive for the other tube instead of giving a push-pull power section 2 drives signals, the whole thing is:

Harder to drive than it needs to be.