gain stage only works with transistor inserted ''backwards''

[Eddododo]

simple gain stages from a larger preamp circuit I'm working on.
Here's the thing- I only get sound (it is clean, too) when the 3906 (pnp) is inserted backwards on the breadboard- that is to say, with the emitter and collector reversed..... Am I missing something?

I have checked the pinout several times, and I have breadboarded this and similar versions of the stages...

anyone?

[GGBB]

No clue really, but the circuit looks like you should be using NPN BJTs, not PNP.  What is the ">" symbol? +18V?

[Eddododo]

yes, its a 'V' sideways... I forget that when i Play in spice that I don't leave things very readable to the outsider 



so i  just found this, http://www.reprise.com/host/circuits/images/to-92.gif

thinking perhaps I misunderstood the nature of bipolars (done very little with them...)
am i misunderstanding the nature of the emitter/collector, such that for PNP the emitter expects a positive ground?



http://www.jensen-transformers.com/wp-content/uploads/2014/08/as098.pdf
thats the reference for the topology, though i think you have to sign in (free)

[GGBB]

for PNP the emitter expects a positive ground?

That is also my understanding - a PNP BJT's emitter should be positive relative to its collector.

But I will defer to an expert.

[R.G.]

Bipolar transistors are two junctions diffused to either side of a middle section. Originally, the middle section was one chunk of N-doped germanium and the other junctions were made by attaching pieces of indium (a P-type donor) to either side of this base (...!!...) and cooking the whole sandwich in an oven at high temperatures until the indium diffused into the germanium base.

In the originals, both junctions were as symmetrical as this clumsy manual setup could make them, so the choice of collector and emitter was a bit arbitrary. Both junctions had similar-ish characteristics, and the gain in one direction (alpha) was about the same as the gain in the other direction (beta... !!...). Then they found out that by making the junctions asymmetrical, and putting more voltage resistance on one of them and more current flow on the other, and making the base region thinner and smaller, you got more gain and higher frequency and...

Well, 50-60 years later we know a lot more about them. Bipolars work both ways. They work best for somethings the normal way, with the collector-base junction reverse biased and the base-emitter forward biased, and the junctions are no longer symmetrical. Most modern silicon devices have a sustaining voltage of 5-7V for the base-emitter, and the BVcbo on the base-collector, and this last may be hundreds of volts.

But they still work backwards, kind of. The gain is asymmetrical too. The reverse gain (alpha) is now quite low, down in the units of times, a consequence of getting the forward gain (beta) up into the hundreds.

And now we're getting close to an answer to the original post... 

I happen to have been working on a circuit that is remarkably similar to that recently. It's the basis of the preamps in the Thomas Organ Vox amps' last models. The JFET provides input transconductance ( i.e. voltage change at the gate to current change at the drain) and the transistor offers current gain, then the overall gain is determined by the loading resistor from the PNP collector position to ground. This is stabilized by feedback from the nominal collector of the PNP to the source of the JFET. It turns out to be possible to make this thing have parasitic oscillations at very (very!) high frequencies. Some low-end oscilloscopes won't see it. But it sure won't amplify audio while that happens.

The Vox preamps have a couple of things you don't. There is a 470pf cap from the collector of the PNP to its base to lower the frequency response. Even then, they used a 2N4303 JFET, which has a lower transconductance than the 2N5485. I tamed this by lowering the base-emitter resistor on the PNP from the 1.2K Thomas used down to 680 ohms. The 10K you have there is letting the thing run wild, all the gain it can scream with.

Flipping the PNP around gets you much the same thing by forcing the PNP to work with the reverse gain (alpha) instead of the high beta. I would guess that this is what's going on. Works with lower gain. Try it with a 560 or 680 instead of that 10K, and flip the PNP to the normal direction. Also try the feedback cap from collector to base of the PNP.

[Joe]

That is also my understanding - a PNP BJT's emitter should be positive relative to its collector.

This is correct. Probably need to be using an NPN in there.

[Eddododo]

OP update:
let me start with a question- the 'TOP' of the bipolar,connected to v+ ; would that be called emitter or collector?
If emitter, then I was wrong with my layout, and the 'backwards' that was working was, instead, correct


R.G, I actually do have a 100pf in there as you said..


Sticking with the 8k or  so that i trimmed in... working very cleanly as far as i can tell, now

[R.G.]

The schematic symbol is that the transistor lead with an arrow on it is the emitter. The no-arrow lead at an angle is the collector. The lead at right angles to the bar is the base.

If it works with an 8K, great!

[Eddododo]

thanks, as always, R.G.