Maximum weirdness amplifier project.

OK. So I now have three amplifier projects on the go. When I get sick of one I just start on the other. SO then, this is my latest effort at turning unwanted junk into amplifiers. I got the original idea from an article in Tubecad
It seems to be a cross between a conventional drain loaded push pull, a source loaded push pull (long tailed pair?), a circlotron with a single supply rail, and a pair of oppositely driven paraphase phase-splitters!

The impetus for the project was that I recently inherited a 5kva :happy1: C-core transformer with four independent 240v windings. The way I see it, the transformer doesn't have to be anything special in the audio department because for the most part it only supplies current to the cross-coupled capacitors. These do most of the work supplying signal to the load at medium to high frequencies, the transformer only digging in at the lower frequencies. As a result the caps need not be huge, perhaps only 1000uF or so.

Another reason for starting this amp is that I have a number of 48v (actually 54v) 12.5 amp switchmode supplies that because of their HF filter caps to earth, won't float the output at audio as needed in a circlotron. With this amp I think I'll just bump up the supply rail at 50v per time and see how we go. heh heh. :mischiev: Better keep an eye on the fet safe operating area. I havent actually started doing anything yet, just mapping out the general idea. This could be fun.



  • cross coupled amp.gif
    cross coupled amp.gif
    5 KB · Views: 874
Vacuum tubes

Now that I have been pondering it a little, I think this kind of output stage has a lot to offer for tube amps with a proper double wound o/p tranny. A normal class AB tube amp only drives more or less one half of the o/p tranny primary winding at any one time. Makes for greater leakage reactance between primary and secondary so worse HF response and also slightly lower winding utilisation i.e. greater primary winding heating for a given power output.

This setup though, drives the entire length of the primary always. In fact seeing there are two complete primaries there is greater scope for interleaving primary and secondary windings for tighter coupling. I suppose you could try it out using 2 separate conventional transformers with the secondaries paralleled. Not quite as good but it might give you some indication.


Joe Berry

2001-03-15 6:15 pm

I don't have a good answer for why the idea didn't catch on in a big way. My guess is that it just didn't (appear to) offer a compelling advantage in cost or performance over the status quo at the time.

At least one aspect of the design -- the idea of making a floating bridge from a single decoupled power supply -- does survive in the form of a preamplifier circuit used by Atma-Sphere. The circuit they use is covered by US Patent 6,242,977. In case you want to look it up, here's a better link for the USPTO search page:
Yeah Joe, I do remember coming across that patent but I never noticed it was single supply at the time. I sort of re-invented the wheel on that one now that I look at it closely; it is the same as the first and last ccts on the page in this thread.
It oscillated at me so I didn't go any further, thinking it was fundamentally flawed. Perhaps not? It's funny how that when a circuit is drawn differently it can seem to be entirely different to another. The Karsten patent is a pair of cross coupled paraphase phase splitters, although at first glance it looks more like something from outer space.