Best way of powering a -3V circuit on a multi power supply

[BluffChill]

One of my customers has ordered a custom job of my 3V fuzz (think FZ-1 or Selmer Buzz) but wants to be able to daisy chain it with his other effects. I'm pretty sure I can do it, but what are my options? I've thought of a few.

1. 3.3V regulator, into LT1054/7660S/MAX1054 to get +/- 3.3V
2. LT1054/7660S/MAX1054 to get +/- 9V, then voltage dividier on the -9V rail to get it down to around 3V.

Does anyone have an opinion on the most efficient way?

[Rob Strand]

The obvious solution would be,

9V  --->  LT1054 wired as a negative voltage generator ---> -3V0 or -3V3 regulator.
So the down sides are:
- you are introducing a switcher which can introduce noise but it usually OK with some care
- need to find a negative regulator which can also handle -12V input.
  You could use an LM337 or roll you own with transistors.


Perhaps another way to spin it is to wire the circuit as a negative ground:

- What is currently  -3V gets wired to the new ground.
- What is currently 0V gets wire to +3V
- Add a 220uF cap across the 3V rail.
- power the circuit from a +3V or +3V3 regulator (which can handle +12V input)

3V3 regulator needs to be stable with large caps otherwise you need to follow the data sheet or application
notes about how to drive capacitive loads.

Optional but better/recommended:
- move the input socket ground connection from old ground (now +3V) to the new ground.
- flip the polarity of the input cap
- move the level pot from old ground (now +3V) to the new ground.
- flip the polarity of the output cap
- move the output socket ground connection from old ground (now +3V) to the new ground.

I can't guarantee it to be problem free but it does simplify things circuit-wise and connection-wise since after that it is essentially a negative ground effect.   The 220uF cap certainly helps reduce problems.

[Eb7+9]

If we assume that your -3v fuzz circuit uses PNP germanium devices, like the FZ-1 does, then it most likely matters (ie., both for maintaining original breakup quality and also providing circuit stability) to ground the Emitters directly rather than use a frequency selective capacitive ground point ... If we also assume that your fuzz doesn't draw too much current to begin with and the power inverter can handle one extra mA of current to drive a grounded 10k pot then you can go first to -9v from a power inverter circuit and then from the said 10k pot whose sweeper goes to a modern Si PNP (2n3906) follower to power your fuzz circuit ... the advantage here is you now have the ability to set your -3v supply exactly to where you want it (either internally "fixed" or externally "user") and the Zout of your follower now becomes 50ohms AC or better assuming the follower device is running with a Beta of 100 A/A or (likely) more / closer to what fresh batteries would provide than a big cap would, and without the frequency selective nature of a big cap ... of course you would have to compare both approaches to see which breakup you prefer in the end ... in this later approach you preserve the grounded Fuzz Emitter requirements and obtain an adjustable final supply voltage at once ... with no other caps needed other than in the power inverter ckt

[BluffChill]

Thanks both but I'm having a little difficulty visualising these! Any illustrative assistance?

[diffeq]

The obvious solution would be,

9V  --->  LT1054 wired as a negative voltage generator ---> -3V0 or -3V3 regulator.
So the down sides are:
- you are introducing a switcher which can introduce noise but it usually OK with some care
- need to find a negative regulator which can also handle -12V input.
  You could use an LM337 or roll you own with transistors.

LT1054 has feedback pin to program the output voltage. I suspect it is more efficient way to regulate than doing: 9V going -> linear regulator. Datasheet has an example for -5V regulation. Even that can be rigged into 4 series diodes to get the desired -3.3V.

[PRR]

Simple old fuzzes, two C-cells will last for years.

[Eb7+9]

BluffChill, see if this works for you ...

[ElectricDruid]

How about adding a 6V regulator to the 9V input, and then running the -3V effect from 6V to 9V upside down?

Since the circuit's "ground" side would be connected to the 9V supply, there's no problem with capacitance, and the circuit won't draw much from the 6V regulator, so that shouldn't get hot or anything. And the input/output jacks can be connected together to proper ground just like normal.

Just a thought. Seems easier than messing about with switchers. Although I like Paul's plan too. Just put some batteries in it.

[Rob Strand]

How about adding a 6V regulator to the 9V input, and then running the -3V effect from 6V to 9V upside down?

The "ground conversion" method with the 3V regulator is less problematic.  The supply impedance is as low as it can be (helped by the 220uF cap).  The impedance of the 6V method is the regulator in series with 9V rail.  Also noise on the 9V gets injected into the signal path - in the 3V reg case the supply is as clean as it can be.

[PRR]

> running the -3V effect from 6V to 9V upside down?

If that came to my bench for troubleshooting, after the pain of figuring WHAT was done, I'd bless it with my 20 pound hammer and put us all out of misery.

No doubt it works in concept, and can probably be built, but diagnosis is just too painful to think about.

[antonis]

If that came to my bench for troubleshooting, after the pain of figuring WHAT was done, I'd bless it with my 20 pound hammer and put us all out of misery.

(you owe me about half cup of coffe and a brand new keyboard, Paul..)


P.S.
An old fashion negative charge pump might work well..
(for light load should exhibit  acceptable ripple..)



R3 value should be set according to current need & unregulated voltage..
C2 & C3 according to intended ripple..
D5 & D6     //        //   //       current..

[Rob Strand]

An old fashion negative charge pump might work well..

It only works for AC input.  For DC input you need an LT1054, which pretty much the same idea as that negative rail generator except the LT1054 has to make the AC.

[antonis]

An old fashion negative charge pump might work well..

It only works for AC input.

But of course, Rob..
(the idea is to hang it up to wall-wart PS further feeding with rectified/regulated supply the rest of daisy chain..)

[Rob Strand]

(the idea is to hang it up to wall-wart PS further feeding with rectified/regulated supply the rest of daisy chain..)

That can be done for sure.  I got the impression the OP wanted to run off the +9VDC rail.

[ElectricDruid]

> running the -3V effect from 6V to 9V upside down?

If that came to my bench for troubleshooting, after the pain of figuring WHAT was done, I'd bless it with my 20 pound hammer and put us all out of misery.

No doubt it works in concept, and can probably be built, but diagnosis is just too painful to think about.

hohoho! Fair enough! I agree it would confuse the hell out of a person if you had to debug it without knowing what it was. But it's a one-off custom job, right? And we want to leave interesting challenges for the pedal archeologists of the future, right?!

I still think the best idea so far is your "Batteries", Paul.

[PRR]

you owe me about half cup of coffee

[BluffChill]

> running the -3V effect from 6V to 9V upside down?

If that came to my bench for troubleshooting, after the pain of figuring WHAT was done, I'd bless it with my 20 pound hammer and put us all out of misery.

No doubt it works in concept, and can probably be built, but diagnosis is just too painful to think about.

hohoho! Fair enough! I agree it would confuse the hell out of a person if you had to debug it without knowing what it was. But it's a one-off custom job, right? And we want to leave interesting challenges for the pedal archeologists of the future, right?!

I still think the best idea so far is your "Batteries", Paul.

Haha, indeed it was a one-off! In the end, I asked the guy if he minded the battery option, and he said yes, but I'll remember all this for future daisy-chainable FZ-1s and Selmer Buzztones!!

I wonder how Land Devices did it with their No Masters? That was a Buzztone replica with selectable 3v, 6v and 9v supply options. Anyone have one they can crack open!?

Thanks all for your suggestions!