Noisy transistors in my FuzzFace got me thinking

[brett]

Built another FuzzFace and got a waterfall in the background.
Standard circuit, swapped for neg ground.  The PNP transistors leak about 10 and 40uA, hFE of 50 and 60 (GT308B, 1977).
Q1.  What generates the "white noise" in a low leakage transistor?
Q2.  Would transistors in a simple high gain test rig (no input) reveal noisier vs quieter devices?  ie rank transistors in a simular order to which they would perform in a FF.
(test rig could be 9V, Rc=10k, Re=0, collector feedback bias (Rf~220k), output tapped at collector via 0.1uF cap)

[mozz]

Have you tried to use them in a regular pos gnd circuit? I've yet to build one with the swapped arrangement but have read about problems doing that.

[brett]

In my case it's not the old motorboating problem.  My power supply is decoupled wrt ac (fancy words for a 22uF cap across the supply).
See Jack Orman's page for more info.
I've realised that a fair bit of this noise is "antenna-ing" where the guitar pickups are the antennae for RF noise, amplified by the FF.
In the meantime I'll replace the Ge bjts with Si pnp power bjts (BD140s or TIP32s). As a test for potentially noisey vs almost silent transistors.  Or is there something I'm doing differently with the pcb this time?
cheers

[brett]

I tried quiet, extremely low leakage Si transistors without luck.
Then tried 100pF caps from collector to base of Q1 and Q2, no luck.
Swapped 0.1uF MKT input cap to standard 2.2uF RB electro.  No luck.
My 22uF decoupling cap for the power rails might be too small or too high ESR.  I'll add a 0.1 poly cap  and 470uF in parallel.
But.... I was wonderibg about the "grounding" of Q2's emitter.  The output signal is wrt 0V, but the emitter is connected to -9V.  Although the decoupling cap should make them both AC of 0 volts.... any residual AC is amplified enormously. 
Looking to see if there were any "tricks" wrt the 22uF bypass cap, I found this in an old post by Fenton B, and will "swap" the grounding of Q2's emitter to the common DC and AC grounds used for in and out.
Finally, I feel that success is just around the corner.
"I could never get this circuit to work properly without a big a$$ cap across the power rails, but then I came across a slightly different solution to yours.

When building a negative ground PNP Fuzz Face (still only on the breadboard), I have had great success bypassing Q2s emitter resistance to "ground" ground rather than to the positive rail. So with the positive side of the cap on the emitter and the negative to the lower rail. Seems to solve all the issues I was having."

[PRR]

Battery or power supply?

The reversed rails will bring-out any crap in a power supply. Don't blame the build until you try a dead-quiet battery.

[brett]

Update: I tried both wall-wart and battery in Neg ground, still no luck. 
There's clearly an antenna effect going on.

So I've gone Pos ground with a battery.
Quiet.  Vastly different. 

Something about literally tying all of the AC and DC grounds together, and having signal shielding continuous from the guitar to the box has worked.
What I didn't try: a big filter cap (I only used a 22uF in parallel with a 0.1uF film cap).

In other news, I'm using 2 x 60 hFE Ge transistors (Russian GT308B).  Q1 has a 10pF cap from collector to resistor to suppress RFI.  There's no pull-down resistor or any other mods.  A 4.7 k resistor on the collector of Q2 biased it to 4.8 V.
Fairly much what would be expdeted: low hFE, smaller resistor.

In my opinion it sounds great.  Maybe not the extreme fuzz of a fully cranked FF with Q2 of hFE = 120, but lots of usable settings with ghe fuzz control set at 7, 8 and 9.

A fantastic result, even if it wasn't what I set out to do.