Crazy amounts of gain out of Germanium fuzz face

[joelap]

I am getting some crazy amounts of gain out of a simple fuzz face circuit.  I decided to try my hand at making a pcb and getting it made at advanced circuits for their 33 dollar deal, and I cant believe what gains I'm getting.  When the fuzz is on 0, it still sounds like my other fuzz faces when it is halfway up.  Fuzz at midpoint, its the cranked fuzz sound I love.  Turn it up past 2/3rds of the pot travel, and it oscillates like crazy.  I have no idea how this is happening!  Pleasant surprise, but I'd like to get rid of the oscillation...  All resistors are the correct values in the correct locations.  Same with caps.  I'm using a 2kB fuzz pot with 22uf bypassing it.  Transistors are AC128's.  First transistor has the Ge diode in there for temperature stability.  I have kept all wires to pots as short as possible.  What could be causing this insane amount of gain?!  There are some 47uf caps on the board for a buffer... if I had accidentally put a 47uf cap in the fuzz control instead of a 22uf, would it make that much of a difference?  My instinct is telling me no, but I'm not sure.  Poor lead dress could be a cause of oscillation, but that doesn't explain the gains I'm seeing.

Also, I have the circuit in close proximity to a bass fuzz (same enclosure at the moment, and the circuits are side by side) using an audio transformer like described on www.home-wrecker.com  Could this be inducing more current through the AC128's, causing the increase in gain???  Maybe I should cut some metal to sheild the transformer from the AC128's?  My reason for thinking this is because these metal can AC128's have their exterior connected to the collector, there is no sheilding.  But I wouldn't think at such low voltage and current levels, that a 42TM018 would produce enough magnetic flux to cause that much of an impact.

Any ideas?  I'd take some sample clips, but the digirack is at my buddy's place.

Joe

[bean]

Have you tried swapping the 2k for a 1k pot? Are you negative or positive ground?

[joelap]

I'm using Jack Orman's method of using a PNP transistor with a positive 9V supply.  I've basically reversed ground and +9V, connecting 9V in the circuit wherever there is a ground and ground wherever there is a 9V.  I did not try swapping the 2k for a 1k because I think the 2k pot at DC should give less gain at the bottom of the pot travel than the 1k would, so I'd expect it to not remedy the solution.

[R.G.]

...Turn it up past 2/3rds of the pot travel, and it oscillates like crazy.  I have no idea how this is happening!  Pleasant surprise, but I'd like to get rid of the oscillation... 
I'm using Jack Orman's method of using a PNP transistor with a positive 9V supply.  I've basically reversed ground and +9V, connecting 9V in the circuit wherever there is a ground and ground wherever there is a 9V.

Have you read my repeated posts on reversed ground connections in high gain circuits? 

[joelap]

Have you read my repeated posts on reversed ground connections in high gain circuits? 

Nope... can you copy and paste?  I can never seem to find what I'm looking for with this forum's search function.

But I should also add: I have this circuit breadboarded and it does not suffer from these problems in the least bit.

EDIT: Ok I found a few posts.  I just dont understand how it could work perfectly on breadboard, but not in a circuit, where breadboard is inherently noisy and filled with parasitic capacitances, and a circuit board is not as much.  I'm using the same transistors too.  Any theory explanations as to how this could be the problem?

[R.G.]

I post this advice every time someone pops up here with trouble with a reversed-ground circuit. Here's the advice:

-=> Don't do that. <=-

I've tried this a lot. Most of the time it works OK-ish. Sometimes, it cannot be made to work with any circuit tricks I've learned in my entire time in electronics. I don't know the reason or I could tell you how to make it quit. I've tried careful layout, careful grounding, shielding, rerouting wired, star grounding, increasing device bandwidth, lowering device bandwidth, filtering, gain-bandwidth modelling, and tossing salt over my left shoulder at the crossroads at midnight. I've spent a lot of time on this over the years, and the most reliable advice I can come up with is not to do it unless you're willing to take the consequences of having to reverse it.

What keeps this thing going is that lots of times you do get away with it. So people who have tried it an had it work first time feel lucky, and think the advice is silly. Until their turn comes. There are two kind of people with reverse ground circuits that work. Ones who've had their circuits oscillate, and ones who haven't - yet.

[joelap]

Thanks RG.  I guess I'll cut all the traces heading to ground and 9V on the board, and make a small daughterboard on perf with a charge pump to give me -9V.  I just hope this wont adversely effect my switching... I dont want to have to implement that switching thing you have on geofex.com, because I mgiht run out of room.

The way I read about it on muzique I thought wow that sounds great.  It worked on breadboard no problem, and figured it'd be ok.  Doh! 

[mac]

The more I see people asking if it's possible to build a pnp ff neg gnd, the more I think it's not the need of a different power supply but some kind of perception in our minds. Maybe we are so used to neg gnd that when we see a ff pos gnd we think it's not natural to build it that way. Just a thought.

mac

[joelap]

OK.  Status update.

Repairing the board was easy.  I only had to cut a few traces, and was able to turn a ground pad into a -9V pad.  The other pedals in the same enclosure are still working fine off of the +9V supply.  I added the chrage pump circuit on R.G.'s website with the switching method.  The other circuits are still working fine, but the fuzz now gives no sound.  So i instinctively thought something I cut was wrong, or incorrectly cut.  But after measuring voltages, my charge pump is doing something awkward.  I have quadruple checked to make sure everything is connected properly.  Continuity check (no shorts).  On the collector of the 2n3906 I am getting around 1/2 Vss.  Where I should be getting -9V, I am getting milivolts.  My caps are all 25V, 10uF xicon's oriented in the correct direction.  I am using a Max1044 charge pump.  I have connected pins 1 and 8 for the higher frequency operation.  I am using 1k and 6k8 as resistor values.  The entire circuit is on a small piece of perf.  I have offboard wire going to the input ring, ground, +9V, and -9V.  In addition to double and triple checking the orientation of the components, I tried a 2n5087 to make sure that i didnt have an incorrect pinout on the 3906.  I know the 2n5087 is ebc pinout, since i bought it recently and checked the datasheet on mouser.

I'm stumped, but more stumped as to how I'm getting around 4.5V on the collector of the transistor.  I'm trying to figure out how that can be happening, but to no avail.

To Mac:  Actually,  had never even thought of making a PNP negative ground until I read Jack Orman's lab notebook on muzique.  I thought to myself.. "surely it cant be that simple?!" because I hate using batteries in effects.  Well, I breadboarded it up twice and it worked beautifully.  So I thought to myself this is too good to be true!  Well, I was right! 

I didnt even make the connection while I was breadboarding, making a schematic, or making the layout that "hey... is this what I'm always reading about with PNP tranistors?"  it worked, it worked twice, and it worked good. 

Quoth homer simpson: doh!

[joelap]

bump.  maybe i should start a separate thread for the charge pump issue to catch the eye of more people.