Four for SuSy?

[Nelson Pass]

Yes, those were 389's and 109's.

I think the "diamond" topology is more like this diagram, where
you see the diff pairs operating push pull without limitation to
the potential current they can deliver - they are biased at
the Gates, not single-ended, and could operate class AB
beyond the bias point. This allows for the current to deliver
enormous slew rates.

[Nelson Pass]

Actually, now that I think about it, you can get the Curl
circuit to do that too.

[acaudio]

Originally posted by Onvinyl
Is this Idea a valid option worth trying?
Improved linearity through complementary differential further tamed with sort of double-x?
Or would the circuit blow and explode due to the different charges meeting at the gates?



Just a silly morning-idea before coffee & work?
yours,
Rüdiger

Have a look here:

http://www.diyaudio.com/forums/showt...823#post613823

Regards
Adam

[GRollins]

Nelson,
I see the schematic you drew as being the same as John's, just with Rs=0; i.e. each device running at Idss, which ain't that far a stretch, given that the biasing resistor is usually pretty small, anyway.
Note that all bets are off if I'm mis-remembering the diamond circuit topology. For some reason, I'm thinking that I saw it drawn with the connections criss-crossed between the two differentials...probably just to confuse the issue.
As I said, I regard it as being "John's circuit" regardless of other versions. One of these days I'm gonna buy me a JC-2 (regardless of phono EQ errors). I think the circuit's elegant and I like the way they look. So sue me.
Your and John's pinky fingers know more than I will ever know about electronics, but that doesn't stop me from thinking the circuits are cool. Actually, come to think of it, I've got a complementary differential in the tweeter section of my current active crossover. Thanks, John.

Grey

[mlloyd1]

i really like complimentary jfet input stages too, just seems too elegant to not use! the thing that really burns my britches is the matching. i don't sleep well now because i measured the gains of my upper and lower diff amps of my Servo 50 years ago when I was building it. even though the jfets IDSS matched, transconductance clearly didn't, so the gains were slighty different (yeah, yeah, my wife calls me a nerd, too). that can't do anything nice to my distortion spectrum once I wrap global NFB around the amp.

what can i say, it's a complementary input stage, i want complements, d@mmit



mlloyd1

[Nelson Pass]

Originally posted by GRollins
I see the schematic you drew as being the same as John's, just with Rs=0; i.e. each device running at Idss, which ain't that far a stretch, given that the biasing resistor is usually pretty small, anyway.

First you will note that I drew a bipolar circuit (upside down).
It seems to me that John's contribution consisted of the self-
biasing of JFET dual differentials, not the notion of dual
differentials as such. I can only go by the apparent 1974 release
of the JC-2, and unless I am mistaken, the notion of dual
differentials precedes that. If I am mistaken, perhaps John will
comment.

Such amplifiers as the Daniel Myers 'saurus amps and Jim
Bongiorno's Ampzilla used dual differentials also, but you will
note that in those cases the circuits were CCS biased. This
means that they did not slip into a Class AB mode so as to
provide more than the bias current.

John did not intend such operation, as it is highly non-linear,
and Sansui may have tarted it up for the occasion, but you'll
recall that this was during the "slew-rate wars" of the 70's.

Regardless of the details, the basic concept of the diamond
differential (the not-the-Curl circuit) remains the same, but it's
only value seems to have been for high slew rate numbers -
running an input device in Class AB is not a popular approach.

[Steven]

If I remember correctly also Yamaha made an amplifier with a dual differential input, each differential having its own current source like all the others do. So far nothing special.
The special thing is that they connected a big capacitor between the common emitters of both differentials (could have been common sources too, I don't know). Now you get the same class AB characteristics as mentioned above, to get really high slew rate specs. The charging current of the next Miller cap is not limited anymore to the value of the current sources. The capacitor allows for shifted DC biasing.

Steven

[GRollins]

Nelson,
Mea culpa on the bipolars above. What with Olivia wanting to sit in my lap and my wife requesting that I fetch a bottle of wine for supper, my mind just interpreted them as active devices in the generic sense. My focus was on the hookup between the two differentials, not the differentials themselves.
I already had the John/JFET connection, just didn't look at your schematic closely enough.

Grey

[Nelson Pass]

Well of course it was a test to see if you were paying attention.

[GRollins]

Clearly, I wasn't...not enough, anyway. (Note my reference to Idss etc. above. Sigh.)
Like the Shadow, wine has the power to cloud mens' minds. The mere mention of it is sufficient to flatten my brainwaves. At the moment I am drinking Le Grand Montmirail '98, a Gigondas (i.e. Cotes-du-Rhone, France--not my usual, since I lean towards Bordeaux, but hey, it pours and is tasty and that's sufficient unto the moment). Hence, anything I say in this post should be taken with a grain of salt.

Grey