Reverse Active Mode...But Why?

[turdadactyl]

I was checking out the schematic online of the DBA Supersonic Fuzz Gun and I noticed a funny configuration on the first two BJT's.



Googling a bit, I discovered these are configured in "reverse active mode."  I believe the setup makes them common collector amplifiers.  BUT WHY?  All I can figure is this was one way to get a high input impedance.  Anyone have other, more satisfying, explanations?

[PRR]

Emitter and Collector junctions are similar. In a few Ge parts for switching they are interchangeable; most amplifier and large switch parts dope the junctions differently. Emitter is doped for better current gain. Collector is doped for higher breakdown voltage. (The default doping used on Emitter gives ~~7V breakdown.)

In this plan (assuming no typos!!) the "collector" junction is the forward bias junction, so indeed it is "inverted", swapped C/E. The 7V breakdown of the reverse junction may not matter in a 9V circuit. The very low current gain of the forward ("Collector" used as Emitter) junction means current gain is 0.5 to 5, instead of 50-500 when we use the designated Emitter.

If my analysis is right: the input impedance is LOW! Maybe 1k plus C1 C2 Dens pot. On naked guitar this will roll-off highs, which with new highs from distortion, may be "a sound". But it will be different for every source device (guitar, guitar VOL setting, or boosters etc in front.)

There is also the "backward connected" pot at the output. This *can* be a fine design when the load impedance is known (and usually low). It is not clear why it is used here.

Understanding ALL ways circuits can be built is overwhelming. Designers usually use "common" connections which can be shown to be "best" for specific cases. I suspect that here we have a designer who tried all the "wrong" connections and found one which was interesting.

[Fancy Lime]

IIRC, the most common explanation I have read for this is that you get very low hfe out of otherwise high hfe transistors if you use them the wrong way round (or reverse biased, as it is sometimes called). I have done this in the past and it can sound pretty good. The point is that many of the classic fuzz designs are actually just a a box of EE tricks to squeeze the maximum possible gain out of terrible, low gain transistors that are not remotely up to the task. And it turns out that running really "bad" transistors way past their comfort zone sounds pretty good. That's why we like germanium fuzzes. If you stick a pair of 2N5089s in an Orpheum or a fuzz face type design, that sounds rather ...meh. But if you reverse bias these technically quite good transistors, you get a pair of technically very bad ones, which you can run at maximum gain, because that maximum gain isn't all that high any more. And the amplification in general is not as linear as it would normally be, which may also contribute to the interesting sound. I seem to remember Jack Orman had an article about this somewhere, I just can't find it at the moment.

TL;DR: Looks weird but can sound surprisingly good, worth giving a try, especially for designs with no emitter resistor.

Andy

[iainpunk]

there are two things at play here,

the Hfe and the saturation current

the Hfe:
in my experience, most transistors have 1/3 to 1/4 their Hfe when reverse active.

the saturation current:
the saturation current of a reverse forward transistor is remarkably low, thus it will clip WAY earlier, also, it will clip differently, instead of clipping against the power rails and/or ground which is quite hard, it creates its own harmonic pattern which is different.

[Rob Strand]

At the end of the day reverse active mode looks like a very low hFE transistor.  Perhaps anything from 1 to 20.   Once you know that all the stuff PRR mentioned follows.   The reverse BE breakdown does need to be considered which is say 8 to 12V.   The other thing is once you reverse them and get the low hFE you need to change the bias resistors considerably - well, unless by fluke it sounds good.

[turdadactyl]

I suspect that here we have a designer who tried all the "wrong" connections and found one which was interesting.

This seems like a pretty satisfying conclusion to me (FWIW).