I've just realised that one bit I wrote above is misleading: -
(e.g. the changing load presented by the current dumpers as they turn on and potentially present a widely varying load to the class A stage).
This is a bit misleading, as one may interpret that the load presented by the base current into the dumpers could cause non-linearity, but this actually isn't the case, as it's also cancelled by the bridge, but the large changes in gain, with varying current in the dumpers,
can give rise to non-linearity due to large gain variations in the o/p devices and the non-zero o/p Z of the class A stage. Even this though is theoretically at a very low level in the Quad circuit.
We can address this in several ways though, better choices of output devices now available will help, but the clever way is to see if there is a further modification to the bridge circuit that can remove this problem.
I have some papers somewhere that present just such an idea, hence my keen interest in the circuit, I wonder what Linn have done.
I even have some info on a Class D amp with feedforward correction too!
Andy.
P.S. Paul, I've just found that Linn patent on the PSU, written by none other than Bill Miller - looks interesting!
P.P.S. the Troika body is patented too! (GB2188718)