It's awful late tonight--- so I will keep my comments brief.
A resource that I wanted to provide a pointer to is located at
http://www.pmillett.com/file_downloads/ultralinear.pdf
it is a pamphlet on the principles of Ultra Linear operation
written by Herbert Keroes.
In reading the pamphlet I was particularly intrigued by the possibility
of UL operation of small signal tubes--- whereas UL in the output stage
has been thoroughly explored and implemented for several decades now.
One of the Acrosound pre-amps that Herb had designed and built was a PP design and used ultra-linear techniques for these small signal tubes. If anyone by chance has a schematic of this pre-amp it would be interesting to see.
So Kevin and I kicked around a bunch of ideas--- and Kevin drew up three schematics (two of which are shown) to illustrate the different conceptions. Interesingly, I think we've covered the major forms of coupling (LC, RC, and transformer coupling) between stages with the exception of direct coupling (I'll have to get Keto or Dan Schmalle on the direct coupled version :=).
Would be interested in any constructive feedback on the two (RC and IT) variations shown.
Tube choices could be expanded to possibly include the following as potential candidates.
EL84, 7788, 6688, 6AQ5, 6AU6, 6005, 6267, 5847, 6AC7, C3M, 6AK5 and 6CL6. Thanks goes to John Levreault for his input in making up this expanded list. He is perhaps the "go to guy" for his knowledge and experience with pentodes.
And geez--- I just remembered the granddaddy of all the audio pentodes for front end duties--- add to the above list the WE 310A as a potential candidate.
Briefly--- and with no real attempt to be exhaustive--- I wanted to extend to the front end tube the theoretical benefits of UL operation (see the Keroes paper for the details). One benefit in particular that attracted me to the UL idea was that you can get a much lowered output impedance while using a pentode or tetrode in an LTP arrangement that has enough gain to do a single stage in front of the output stage (so we keep it a two stager).
And there is much more--- the transformers specified must be optimized amongst a large number of electrical and sonic parameters. One very important consideration in the proposed transformer designs is that they must have a very, very high degree of AC symmetry. For proper UL operation taps off of transformers must be optimized from an AC point of view for this service. Simply pulling taps at random does not achieve this degree of AC balance in every transformer design extant. For a discussion and analysis of the importance of AC symmetry in UL operation I would suggest reading the article authored by Norman Crowhurst which was republished by Glass Audio in the early nineties. Two other references which emphasizes the need for a strict AC balance (as referenced to ground) in UL transformers would be, of course, the patents from Herbert Keroes as well as David Hafler.
Another guiding principle in the topology of the newer schematics is that I wanted all of the feedback (via the UL operation of the tubes) to be absolutely local. I didn't want to pull a UL tap off of the output trans primary and then feed it back to the first stage--- for a number of different technical reasons.
Overall, I think that the two current drawings offer several significant theoretical improvements over the first generation of ideas that had been shown in message number 7472 (MQ forum).
MSL
Builder of MagneQuest™ & Peerless™ transformers since 1989
