john curl said:
I know, but I said for minimal distortion.
Besides, it was you (post 6075) who ask to make RE = 0.5 * VT / Ie. 😀
I might give a short history of the design and development of this circuit over a decade.
It started as an almost accidental effort using a single mercury battery as the power source in 1971, while working at Alembic Inc. It was found impractical to use a single battery, so it was changed to 2 x 1.5V alkaline batteries and 2 added build-out resistors.
A paralleled form was developed by Mark Levinson Audio in 1973 and first shown at the NY AES in 1973. It achieved 0.4nV/rt. Hz noise and developed as a tranformer replacement for MC cartridges. It was very successful and put Mark Levinson on the map, so to speak. The circuit was developed further in 1974 and 1975, and alternate versions were offered, with emitter input, and non-paralleled (to save cost).
In 1979, an even more improved version with newer devices. removed the need for paralleled operation, but practical problems developed with the best devices due to what appears to be 'early saturation' in some of the better devices. However the Michaelson and Austin, Sota, and several pirated efforts were made around 1980.
In 1981, I decided to develop an all fet alternative and it measured and sounded so much better, that I dropped the original design entirely, and went to fets exclusively. That is the end of the story, and why I can openly discuss it here.
It started as an almost accidental effort using a single mercury battery as the power source in 1971, while working at Alembic Inc. It was found impractical to use a single battery, so it was changed to 2 x 1.5V alkaline batteries and 2 added build-out resistors.
A paralleled form was developed by Mark Levinson Audio in 1973 and first shown at the NY AES in 1973. It achieved 0.4nV/rt. Hz noise and developed as a tranformer replacement for MC cartridges. It was very successful and put Mark Levinson on the map, so to speak. The circuit was developed further in 1974 and 1975, and alternate versions were offered, with emitter input, and non-paralleled (to save cost).
In 1979, an even more improved version with newer devices. removed the need for paralleled operation, but practical problems developed with the best devices due to what appears to be 'early saturation' in some of the better devices. However the Michaelson and Austin, Sota, and several pirated efforts were made around 1980.
In 1981, I decided to develop an all fet alternative and it measured and sounded so much better, that I dropped the original design entirely, and went to fets exclusively. That is the end of the story, and why I can openly discuss it here.
Bob Carver wrote an earlier article 'proving' that crossover distortion in power amplifiers as inaudible as well, with attendant listening tests. You might find it in an archive somewhere.
Edmond Stuart said:
This circuit can be built around Schmook multi-tanh pairs, I leave it as an exercise for the reader. You can get perfectly flat gm for 25mV or so input voltage with no degeneration. But you can't do it with discreets, and as I said before modern high speed processes often don't like to be linear near saturation, too bad.
Mercury batteries have been illegal for years BTW.
Hi John,
As for the pre-amp, thank you for explanation.
As for Carver, apparently these listeners were stone-deaf.
Cheers,
Edmond.
As for the pre-amp, thank you for explanation.
As for Carver, apparently these listeners were stone-deaf.
Cheers,
Edmond.
Schmoock
An input stage transconductance reduction technique for high-slew rate operational amplifiers
Schmoock, J.C.
Solid-State Circuits, IEEE Journal of
Volume 10, Issue 6, Dec 1975 Page(s): 407 - 411
An input stage transconductance reduction technique for high-slew rate operational amplifiers
Schmoock, J.C.
Solid-State Circuits, IEEE Journal of
Volume 10, Issue 6, Dec 1975 Page(s): 407 - 411
scott wurcer said:
Thanks for your input, Scott.
As I see, no option for the DIY folks. So, for ultra low distortion, nothing else remains than the more traditional topologies with...... lots of NFB. 😀
Cheers,
Edmond.
Hitsware,
If you look at the feedback network it has a fairly low impedance so some sort of follower or buffer is needed. The mosfets should get rid of one junction that the signal has to pass through.
Oops, dimitri got here first.........
Jam
If you look at the feedback network it has a fairly low impedance so some sort of follower or buffer is needed. The mosfets should get rid of one junction that the signal has to pass through.
Oops, dimitri got here first.........
Jam
- Status
- Not open for further replies.