Dr. Sidney Darlington's transistor

'FWIW#2: You can also get gain from thie CFP-arrangement by inserting a resistor underneath the PNP, as shown here:"

Yeah, that's exactly the configuration I meant.

Click to expand...

Oops, I goofed, did interchange these two ways of doing gain.
So put another way then, you can also get gain from this by putting a resistor in the emitter-lead of the PNP, as done here with R2 for Q2:

(as from http://www.groupdiy.com/index.php?topic=7793 )

It's amazing how many benefits small variants of that basic pair with a few more Q's can yield. I'm working on a design prompted by the beginning of this thread and getting some pretty interesting performance numbers in simulation at least.

Click to expand...

Please keep us informed, would be interesting to see what you're working on.

BTW, somewhat related and not much noticed when it was posted in Samuel's thread on a simple single-ended mic pre, a design I did for that application when only 5V is available:

http://home.datacomm.ch/gronerfamily/DIY/VoiceLessonPreAmp/simple%205V%20mic%20pre.pdf

This has not been built yet as far as I know but the performance in sim looked quite promising, given the power supply constraints.

[quote author="bcarso"]"because I remember beating reeds into papyrus for my blueprints...."

Haha. Last one into the Nile is a rotten scarab![/quote]

Papyrus? Luxury you had. We had to wait days for the clay tablets to dry out enough to scratch cuneiform schematics.

My friend Chris Shelton always used to talk about how Reader's Digest should have had a archaeological joke page called Humor in Cuneiform.

And the mention of clay reminds me of the late S.J. Perelman's quip that "before they made me, they broke the mold."

<fa dumph>

Just so this isn't totally off topic, I would assert that no matter what the bias is applied to the first transistor of the darlington, structurally the voltage noise generators always r.m.s. sum. So the complementary pair will always outperform unless the configuration has so little gain from the first device that the second's noise becomes important. In many cases within a circuit these noise differences may be insignificant.

Darlingtons are great for simplicity and high input Z though, and the power supply rejection is pretty good owing to the high collector(s) impedance.

Also, for RF work if you make the impedance in the B-E of the second transistor really low, the pair becomes a sort of single-ended f sub t doubler, with the signal now usually taken at the collectors.