Positive Ground question

[Astronaurt]

Hey so I'm building and experimenting with a Ram's head style Big Muff circuit like this one: http://www.kitrae.net/music/BIG%20MUFF%20CIRCUIT%20GUIDE.jpg

and I'm also about to Custom build a regulated Pedal Power supply. Now the power supply I'm building is designed of course to make a positive voltage in reference to ground like usual, but this circuit is PNP and requires a negative Voltage ref. to ground. SO my question is this: Can I simply change this circuit so that the Yellow Power supply rail becomes Ground and the ground connections become V+, effectively switching power and ground? Upon first thought, the Voltage drop is the exact same as if it was positive ground, Pretty much like switching the ends of a battery. But considering that this circuit will be connected to the Pedal Power supply and to other pedals, and that the same regulated voltage in the power supply will be connected to other pedals, what would go wrong if I did this? Or is there something about electron flow that I'm fundamentally misunderstanding here? Help is greatly appreciated!

[Astronaurt]

Am I on the right track, or just totally off the beaten path here?

[petemoore]

  It   For PNP to run with NPN a floating ground for all circuits of one of the types assign a floating voltage potential to ground.
  To use NPN and PNP on the same signal ground would mean the ground has +/-...no good for a common power supply because of ''shorted''.
  The DC blocking caps block DC and so the outside circuits can see only AC, and don't otherwise care what previously [or after]  processed/processes the AC signal it is input/work/outputting.
  Floating ground is battery, but a DC9.?V can be calculated/made from a floating AC voltage conversion [high voltage step down transformer.
  The related inferences and 'the Spyder' article [GEO] describe it better, but it essentially boils down to _/+ on the same rail is short, floating power supply rails [ie PPII, Spyder or other form of this] float all connections so no possible +/- sneaking' across a rail, nor can a ground loop be created through no ground connection, with proper filtering is bees knees of ripple and ground loop free  PS's.
 

[Astronaurt]

Allrighty then, makes enough sense. Thanks Pete! I guess I'll try and put in a another Bridge rectifier in the other direction in there so I can make a Negative voltage out of the power supply too then! You can't regulate negative voltage with an LM317 can you, or is there a Negative version of that chip?

[merlinb]

my question is this: Can I simply change this circuit so that the Yellow Power supply rail becomes Ground and the ground connections become V+, effectively switching power and ground?

Yes you can do that no problem. Apparently the Fuzz Face can get a bit funny if you try the same thing, but less archaic circuits like this one don't mind.

[Astronaurt]

f'realz? awesome, thanks Merlin! knowing this simplifies things immensely. I might just finish 'em both tonight!

[R.G.]

Yes you can do that no problem. Apparently the Fuzz Face can get a bit funny if you try the same thing, but less archaic circuits like this one don't mind.

I personally would qualify that. You can do that with no problem - until you hit the circuit(s) where you can't. The reversed-ground connection has some quirks in any circuit, the obvious ones being that the impedance between power and ground is non-zero, depends on the frequency oddities of the filtering caps, and the shift of reference for circuits which have little or no common mode rejection is somewhere between negligible to catastrophic.

Many circuits, even many implementations of the same circuit, have no problems. Sometimes there will be oscillation and noise, which can sometimes be cured. Sometimes not, though. And not for lack of informed trying. I've done the ground/power reverse many times; sometimes had problems which could be cured with careful wiring, bypassing, regulation, or shielding, and a very few times hit circuits that were quite stable with the normal grounding, and oscillated with reversed grounding; and nothing could make the reversed grounding work.

What I carried away from that is that it's a bad idea unless you really, really, absolutely have to have reversed ground and are willing to roll the dice. People who don't have the experience to rip it out and revert it shouldn't try it; or at least should be ready to discard it as a bad idea.

The higher the gain of the circuit, the more complex, and the higher the impedances involved, the more opportunities for this to go wrong.

[merlinb]

I personally would qualify that. You can do that with no problem - until you hit the circuit(s) where you can't. The reversed-ground connection has some quirks in any circuit, the obvious ones being that the impedance between power and ground is non-zero, depends on the frequency oddities of the filtering caps, and the shift of reference for circuits which have little or no common mode rejection is somewhere between negligible to catastrophic.

Well that people's own fault for using noisy crappy power supplies. With a battery or regulated DC the noise on either rail will be essentially nil, so there is nothing to couple into the emitters of the PNPs. (I disagree with your impedance comment; the impedance between each rail remains the same whichever way round you connect your PSU caps. The have no significant directional qualities AFAIK).

[petemoore]

   In an attempt to sum up the thread segue'...
    I didn't catch anything related to any of the approaches that couldn't work.
   Instead of PNP Neg. Gnd FF experiments, I tried one using a MAX1044 to flip/invert the +/- supply line.
  Flipping the polarity inside the circuit box [small daughter board] was believed to be an excellent solution, which works, every time, as well as a PNP Neg. Gnd FF should work..
  It was quite easy to convert to a FF schematic with a ground rail that has a trace to Q1E. 

[R.G.]

Well that people's own fault for using noisy crappy power supplies. With a battery or regulated DC the noise on either rail will be essentially nil, so there is nothing to couple into the emitters of the PNPs.

I'd agree wholeheartedly in 90+% of cases. However, I've run into situations myself where nothing done in cleaning up power, ground, decoupling, regulation, whatever, would make it stop oscillating. Re-verting ground to the original side instantly fixed them.

You are completely right - it *should not* be that way. However, having run into the case where theory didn't work as expected, I decided that there is some bit of theory that I have not yet learned. Not being able to explain and fix it is not for lack of informed trying. As I note, people who happen never to have run into one of these are absolutely certain that my caution is nonsense. Until they do.

(I disagree with your impedance comment; the impedance between each rail remains the same whichever way round you connect your PSU caps. The have no significant directional qualities AFAIK).

The rail impedance item is speculation, as I note above. No, rail impedance is not directional. Actually, the PSU caps *will* always be the same way round, and can't be reversed because they're polarized, but I don't think that means their impedance is directional.

However, the dual roles of ground as both electrical power return and a reference voltage mean that ground impedance does matter, and what currents get mixed into what conductors does matter in circuits where the power supply rejection ratio is essentially zero.

But all this is just sniffing the trail of whatever beast(s) are hidden there, as you recognize. Yep, it should not matter. And many times it does not.

Until it does.

[merlinb]

However, having run into the case where theory didn't work as expected, I decided that there is some bit of theory that I have not yet learned.

I wish I had a circuit that exhibited this, it sounds like a wonderful puzzle to solve! Have you seen it in silicon circuits as well as germanium then?

[R.G.]

I wish I had a circuit that exhibited this, it sounds like a wonderful puzzle to solve! Have you seen it in silicon circuits as well as germanium then?

Me too. I destroyed the first one in a fit of pique. The second was a silicon device I was repairing for a guy. He had converted it to negative ground from positive, and it sang. I spent about a week of evenings on it, and finally converted it back to positive ground, upon which it behaved, and I mailed it back, telling him "don't do that". Yes, in silicon.

My best guess is that it's related to high gain and the exact conductor the ground current is flowing in. I'll test that the next time I happen on one of these beasts, especially since I now have a clip-on high-bandwidth Hall effect current probe to go sniffing the current paths.

[merlinb]

Me too. I destroyed the first one in a fit of pique.

Of course, there's nothing special about PNPs. Astronaut could just as easily make a BigMuff using NPNs and a conventional PSU.

[petemoore]

  I wish I had a circuit that exhibited this, it sounds like a wonderful puzzle to solve!
   I resolved to consider it resolved enough.
  Maybe that cuts the newbs out of newballs of waxy re-wires/attempts to re-work a harder 'lucky puzzle'...considering themselves lucky if seemingly intractable problems surface.
   I went through it for a little while, it took only slight encouragement for me to give up on the ground scheme since the alternatives consumed less time and offered operational circuit every time.
    Max1044 or similar, NPN transistors, battery operation, suitable power supply rails [like Spyder]..fine option IMO since then the PS's are all ironed out, albeit less risky and challenging, seem to be perfectly workable options from simple as can be to not very difficult.
 

[R.G.]

Of course, there's nothing special about PNPs. Astronaut could just as easily make a BigMuff using NPNs and a conventional PSU.

Yep, correct. The only reason this comes up with PNPs is that the vast majority of effects which have positive ground to begin with is a PNP effect from the Dark Ages, or a clone of same. So the junior assistant Einstein pedal modders only go for inverting ground on positive ground pedals to get them to match the ground schemeof most other pedals.