class A amp, but a little unusual...

Well, this one is not actually a question, but more like a punch of my ideas and perhaps someone will have comments on these ideas.

I'm working on my own class A push pull amp design. I wan't to experiment with some less common input and VA stage designs, so at first I tought I'd do a pcb design according to that but with possibility to turn the schematic to more common if the ideas will not work as well as expected.

Some time back though I saw a page where a diy-er had some connectors added to the amp pcb which let him change the CCS-s for example and then compare the different designs more easily.

This led me to idea to try to make my amp layout so that there is a very solid power supply, heatsink and the output stage on one pcb. The output stage would include two parallel push pull stages and a predriver stage + a bias voltage circuit with npn-pnp transistors determing the output stage current by their vbe voltages. Then there would be a 10 pin connector for sepparate pcb with input and VA stage. As the bias circuit is such that the different currects trough it do not need part changes it should make quite good platform for experiments.

Now assuming that a properly made class A output stage is sonically very good and I would make several different boards for two first stages it could be very interesting to be able to listen to one version for an hour or a week and then very quickly be able to change the additional pcb to some other topology.

It is of course a bit tricky and different boards would first need to be bugfree and have the stability needed, but still it would be interesting experiment. It would ideally provide an opportunity to listen to different topologies during one day and in same conditions and as such degrading the possible psychological effects.....

I have also developed a pc based spectral analyser for measurements. So I would be able to measure the THD and IMD and also provide a graphical images of different the spectral content of these measurements.

I at least haven't seen any such information on web and I think if I'd do a web page with the results it would interst some of you....

Anyway, any comments on my ideas are more than welcome.
 
Ergo,

I have experimented with amps that do not use a differential front end with good results ie. better bandwidth and lower noise but there was no free lunch, to direct couple the amplifier I had to use a servo which has its own set of problems.

I am now experimenting with current feedback amplifiers which seem to offer the best of both worlds.

Jam