[quote author="TomWaterman"] Gus, I would love to see a few more discussions around here about the correct implementation of ICs, in particular things like grounding, compensation etc....
I'd love to learn what the 'small details' are and should be...how to do it right like the guys at Benchmark etc[/quote]
There was an interview with the head honcho of Benchmark in, I believe, the 1996:2 issue of Audio Amateur. He emphasized the need to always use a cap in parallel with the feedback resistor so the amplifier isn't running flat out to the point where it intersects the open-loop gain curve. I've found really good supply regulation, and film decoupling on each IC package, crucial. Also keeping the output in Class-A wherever possible; it's normally biased at only about 0.2mA, so either you have to operate into highish impedances or bias the output up into single-ended Class-A using a resistor or, preferably, a current source.
I also like to include a resistor between the output pin and the place where you take the output and feedback (see the circuit I drew for the -10>+4 converter thread) for maximum stability (Walt Jung trick) and another in series with the output if it's going to look at the outside world. Oh, and lots of RFI-proofing on the inputs.
Peace,
Paul