There is the possible beginnings of a story to this one. So, as we are all ‘friends’ (in wonderful internet-speak that is. And bearing in mind you don’t know me from Adam. Or Eve for that matter) I let you in on an interesting prospect.
A gentleman I knew from a long way back called me out of the blue to ask if I would care to figure out why the amp project he had built does nothing but make nasty noises. This is regardless of what kind of nasty noises are being shoved into the front, you understand. The answer to that enquiry, although the problem I find very interesting, would usually be something like ‘GET THE HELL OUT OF HERE!!!!’ There are cans of worms and bucket loads of them, and sorting out somebody else’s design is definitely a bucket load. But Will is no idiot, a genuine bloke, and a good businessman. I don’t know about the first two but I definitely do not fall into the last category. The idea of somebody else doing business-thingies while leaving me to sort out technical issues might work out alright, thinks I.
The problem he has is with parasitic oscillation. Basically the amp is unstable (and unusable.) Not to give away too much, part of his idea was involving a class A output stage. A push-pull (class AB of some description) output stage is a naturally stable design. The reason that power amps become unstable is often that the power rail (this is a valve amp) carries some of the signal of the output back to the preamp stages and produces a positive feedback loop, not unlike acoustic feedback, which is also a positive loop. In a push -pull stage, one valve conducts through the primary winding of the centre-tapped output transformer, and the opposite valve conducts through the other half of the winding.
This means that the power rail that supplies the centre tap, has no signal content because the two halves of the transformer cancel each other out. It’s not quite like that, but close enough. Class A, for all that is trumpeted about its wonders has three serious failings. Firstly, It has to take a high dc current from the supply. It has to be biased so that the valve(s) will conduct the entire signal. Secondly, there is no cancelling effect from the dc supply rail. So the smoothing has to be very substantial, in order to kill off the inherent signal in the dc rail that is pulled by the output valves. Thirdly, the push-pull arrangement also cancels out smoothing ripple at high current drain. Again, class A relies entirely on major smoothing capacitors, iron-cored chokes etc to eliminate the 100hz ripple generated by high dc current drain.
So, although I’ve not seen the amp design yet, these are the things I’ll be looking for. Class A amp building is not an easy assignment. I’ll keep you posted on the hair loss and macaroon situation.