Using an El84 as a driver

[alexmoose]

Hello DIY Audio Community!

I am currently working the pipe dream of designing my own amplifier. You must realize of course that by 'designing' I mean butchering together pieces I like of other people's designs, and tweaking them to all work together. Currently I am toying with the idea of a Push-Pull El34 (yes, I intended to say EL34, NOT EL84) amplifier that is Self Split, and Driven by a pair of EL84s.

The inspiration for this amplifier is Yeo's Simple El84, driven by an EL84
http://www.diyparadise.com/simpleel84.html
http://diyparadise.com/variation.html

I am Reasonable sure that the EL84 (Triode Strapped) can be used to drive it's Larger cousin the EL34, its all just a Matter of Re-biasing the driver stage (Something I am quite new to). However I still face a few problems.

Is it okay to CCS bias both the Driver and the Output stage? I am pretty sure this is okay, but I don't remember.

I am pretty sure I need to find a different CCS to Bias the Self-Splitting Output stage, as I know the LM317 doesn't like Class AB operation. could someone explain this to me in more detail, and if possible offer a solution?

Please feel free to tear down this idea, I present to the DIY Audio Community for constructive criticism.

-Moose

[Tubes4e4]

Hi moose,

Is it okay to CCS bias both the Driver and the Output stage? I am pretty sure this is okay, but I don't remember.

You certainly mean plate CCS for the driver? Not really needed in this case, except you go for very low distorten - but if so, you shouldn´t use a self-splitting power stage running in class AB in the first place since a class AB stage (its "B" operation part, to be more precise) is not suitable to do LTP-style phase splitting because it is not a constant current stage

I am pretty sure I need to find a different CCS to Bias the Self-Splitting Output stage, as I know the LM317 doesn't like Class AB operation

Please view it the other way round: Since a class AB push pull stage inherently is not constant current drawing in total/sum when leaving class A and entering class B, no CCS source whatsoever can really help - this is not the fault of the LM317 IC type.

Any differential/LTP-type splitter must run in class A, hence you are limited to strict class A operation with your self-splitting PP final stage idea. And in that case (pure class A), a LM317 very likely will do - given it has some reserves by an additional negative supply voltage, or you will lose quite some power, because the possible grid swing of the self splitting power stage is limited by the CCS source stalling (quite regardless of type).

All in all I think your idea is not really viable, except you limit the self-splitting PP powerstage to class A (sacrificing Po) and give its CCS an additional negative voltage supply (or you would sacrifice even more power).

Regards,

Tom

[DougL]

Moose;

Sounds like a fun project.

I think that a CSS is only used in a class A stage.
My understanding is the CSS actually prevents class B operation and provides what Lynn Olson referred to as "forced balance"
ETF presentation Once the stage goes into class B operation, there is no AC current path with a CSS and the stage clips hard.

Fixed Bias is what is used for class AB output stages.

I see no reason that a pair of EL84 could not be used as a phase splitter / driver in a 3 stage amp, or Input/Phase splitter/Driver in a 2 stage. As long as you have enough voltage gain to drive the outputs.

I would recommend The tube Cad Jurnal tools to set bias, etc.
The Push-pull calculator for the output stage TCJ P-P Calc
and Tube Cad for voltage gain stages.

For CSS, I would either use a variation of the one SY used on Heretical from the Morgan Jones book, or the IXYS chip CSS. The LM317 would work, too. YMMV.

HTH

Doug