Is there really a Holy Grail cap in our Sansuis...

I've heard it before - A cap in the negative feedback loop, being the most critical in the entire amp. Lore, legend myth ?

Then Reading through "Designing Power amplifiers" by Bob Cordell, page 266

"The biggest offender in most power amplifiers is the electrolytic capacitor commonly found in the shunt leg of the negative feedback network... - ...If Electrolytic caps must be used in this location, large value non-polarized types should be employed... - ... These capacitors should be bypassed with a film capacitor of at least 1uf"

Then he goes on to show a THD vs Frequency graph of a 100V, 100uf NP, vs a 16V, 100uf, P - with the NP 100V exhibiting "25db" less distortion.

Whats your take ?

This should be interesting.

IMO, the components that make up the feedback divider are outside the feedback look. IOW, any non-linearity or frequency dependent error in those components will be impressed on the overall performance of the amp. What cap solution is best is probably open to debate, but yes, it's the place where I'd apply the most thought and the best components. I avoid 6.3V parts in that location. The appeal of a servo circuit is that you can replace a large electrolytic with a nice film cap and avoid most of the issues with caps. Then you just have the issues of building a stable well performing servo circuit.:scratch2: