ad797 vs. ad8610

hey, I hear that the AD86xx's are pretty damn good bang for the buck, in fact, I am going to be putting them in my Rotel CD player. They are known for their low grain and their smooth high's.

The 797 is used in some preamps, and very highly regarded in such an application IIRC.

I have both; I prefer the 797. I don't like the AD8610/8620, although with the BUF634 it is tolerable in the PIMETA, I guess. I like AD's op-amps quite a bit, in general, but the 8610/20 is one I never cared for.

However, if you use the 797 in a PIMETA, you'll need to balance the bias currents which is a bit complicated in that sort of topology. A chip I think might be interesting to try in the PIMETA is the OPA655, though it's a bit difficult to obtain. It would be interesting to see how well the OPA2827 will work.

Intersil's (5002) don't match the pin layouts for the BUF634's in the PIMETA, so they will present a number of issues in a PIMETA. Even if you were able to modify the pin layouts, they have no output protection and need resistors on the outputs. Further, they are not stackable without intermediate resistors between buffers, or individual series resistors prior to connecting to the load. The Intersil-version PPA used 1K resistors prior to each buffer inside the stacking.

The OPA551 is a good buffer choice instead of the BUF634 - you only need jumper pins 2 and 7. This won't affect the use of the BUF634, either.

I sort of agree with Filburt about the AD8620/8610 - they are "too lean" to my tastes. Frequency response seems outstanding, but they just seem to lack body.

The PIMETA is able to take many opamp combinations, but not as many as I expected. For instance, I cannot get the AD843 or AD845 to work without huge offset. I haven't tried them in all three positions, and that may work. It seems that if you don't have a Ground channel that can "synergize" with the LR chip, you have a problem. At first, I thought it had to do with the bandwidth capability - that the Ground channel needed an equivalent speed capability as the LR - but that's not strictly true, either. The Gnd chip must be unity-gain stable, too - which limits your choices a bit.

The absolute best combinations I've tried are the OPA637-LR/OPA627-Gnd, the OPA627 all around, and the AD8066/AD744 (my absolute favorite). The OPA2107 is good with AD744 or AD8610 in Gnd, but it sounds too much like 627-637 - just not as good. Be sure you don't have the other chips if you try it. You'll find yourself not using it as a result. Of course, the AD8620/8610 works very, very well - assuming that you like the sound.

P.S. Off subject slightly, but since mentioning the OPA2107, I've found it to be the perfect match for the tube in a SOHA. It has just enough detail to reveal the best parts of the tube, but not enough detail to reveal the bad parts.

Are you building this PIMETA from Tangent's PCB or on your own?

If you are using Tangent's board, it's not set up for discrete buffers. There were a couple of experiments with Steinchen's Millett DB's, but it wasn't something that could be done on a regular basis.

If you are building without a PCB, then anything is possible. The Diamond Buffers are highly regarded:
That's the schematic that was used for the DB's on the Millett Hybrid. Bias is dependent on the heat sink method used, and you may not have to parallel the output transistors. There is also a wide variety of output transistors to pick from, each with their own sound characteristics. You might check Steinchen's web site for the Diamond Buffers for more info:
dDB for revMH