Biploar to single rail conversion question

[lars-musik]

Hi folks,

my last post already revolved around the Vox Pathfinder 15 amplifier. I'd like to convert the frontend to a pedal and have concerns about my understanding of voltage reference points. The original utilises a +/-15V supply. Now I'd like to transfer that to a voltage quadruplet single supply circuit.
y basic understanding would be that I simply use a voltage divider to create a half-supply reference point (let's call it VB) and link everything that was ground in the original circuit to that new reference. All former +15V become Vcc (here 36V) and all former -15V become ground.

Two questions:

(1) I looked at a lot of schematics since I became a member here (and honestly, a bit addicted) and it just didn't feel right

(2) The charge pump circuit itself references to ground (which would also be the enclosure/shielding of the pedal -to-be), so I surely couldn't mix this up with VB being the new GRD....

Here's the original schematic



and here's what my gut feeling tells me to try

[antonis]

R22 has to be connected at VB..!!
In general, ALL non-inverting inputs have to be connected to VB, either directly (in case of inverting configuration, like U1.2) or via a big resistor (in case of non-inverting configuration, like U1.1 & U2.1)

C3, C4 and R34 can go either to VB or GND..
(I strongly recommed the later, for less garbage going to VB, but it could be inconvenient for board design..)

You might save a cap by deleting C9 and connecting clipping diode pairs to VB, but let it be for the time being..

Can't follow charge pump circuitry but C25 must be deleted or a 100R resistor be placed between U2.2 output and C25/VB junction..
You realise that R9 sets VB about 10% lower than "ideally" half Vcc..

P.S.
Beware of elctro caps voltage rating..

[merlinb]

KISS

[antonis]

KISS

So Merlin, you like U2.1 to be biased at 0 Volts..
(it should exhibit a very interesting asymmetrical clipping..)

[lars-musik]

Thanks a lot, antonis and merlin! It's comforting to have two masters looking over my drawing.
Plus, there's a high level of agreement !

You both think, that the capacitor to "stabilise" VB is unnecessary? Something in the the back of my head wants it there. With the resistor, you'd let me get by with it?

[antonis]

IMHO, better "stabilize" VB with a capacitor between R10/R11 junction and GND..
(op-amps don't like capacitive loads on their outputs..)

See Merlin's proposal above..

[Ben N]

The original utilises a +/-15V supply. Now I'd like to transfer that to a voltage quadruplet single supply circuit.

I'm curious as to why. Why not use the charge pump to generate a bipolar supply, and leave reference at 0V?

[merlinb]

In fact, as a standard dirt box, why boost the voltage at all? 9V is usually plenty...

[merlinb]

You both think, that the capacitor to "stabilise" VB is unnecessary?

The opamp will likely oscillate when driving a naked cap. Smooth Vb before the buffer, then let the buffer do its thing, that's what it's for; it will hold Vb rock solid. Adding stuff after the buffer un-does the hard work of the buffer!

[FSFX]

In fact, as a standard dirt box, why boost the voltage at all? 9V is usually plenty...

Very true - especially when two of the op amp stages are slugged in performance by the Miller capacitors C7 (270p) and C16 (470p).
So for those who will mention things like 'it improves the slew-rate' or other mythical reasons, it is just not true.
I fact it is far better for noise and stability to run with a dual power supply if you feel you must use a higher voltage for that other mythical creature of pedals and guitar level signals, headroom.   

[lars-musik]

I'm curious as to why. Why not use the charge pump to generate a bipolar supply, and leave reference at 0V?

I am the trial and error type without the knack for breadboarding. So I usually design "PCB modules", etch them and then play around. So here, I'd like to be able to run the thing off standard 9V (hence the bi to single PSU conversion), but I'd also like to hook this up to more voltage to see if the original designers of the crcuit had an audible reason for using +/-15V.

[lars-musik]

I fact it is far better for noise and stability to run with a dual power supply if you feel you must use a higher voltage for that other mythical creature of pedals and guitar level signals, headroom.   

With a dual power supply, you mean mains power, AC transformer with a centre tap - bridge rectifier and all that? Or a DC-DC converter? First is not really practicable for a pedal, second hard to find and if, expensive.

I absolutely do not subscribe to pedal mysticism or cork sniffing, so I'll probaly be fine with my Harley-Benton powered 9V. However, you might call me a pedal agnostic: because in the end - who knows.....? 

[FSFX]

to see if the original designers of the crcuit had an audible reason for using +/-15V.

In the world of modular synthesisers and pro-audio supplies of  +/-12, +/-15 and +/-18 are very common.
The chances are that this was done by someone working with those type of things.
For pro audio, there are good reasons to use it.
For guitar dirt pedals it is normally not necessary unless the op amps are those that really can't run properly on 9 volts.
It is therefore unlikely have any audible benefits.

[duck_arse]

(2) The charge pump circuit itself references to ground (which would also be the enclosure/shielding of the pedal -to-be), so I surely couldn't mix this up with VB being the new GRD....

your new VB will not be ground, it will be the mid-point between the + supply and 0V. with a plus and minus supply, 0V is the supply mid-point, used to bias the opamp. it happens to connect to ground via the supply for reasons.

[ElectricDruid]

to see if the original designers of the crcuit had an audible reason for using +/-15V.

In the world of modular synthesisers and pro-audio supplies of  +/-12, +/-15 and +/-18 are very common.
The chances are that this was done by someone working with those type of things.

+1 agree with FSFX, the reason will absolutely have been this. You've already got a ton of gear that uses +/-15V supplies, so what do you design for? +/-15V! It was the default supply for audio for years.

That said, if you've got a 36V supply handy, then why not? Go for it.

[Ben N]

My one & only (partly) SS guitar amp is a Fender Champ 25SE. The preamp, the clean channel of which looks more or less like an opamp version of a BF/SF preamp if you squint at it right, runs on +/-16V. I don't think this has anything to do with a synth background any more than the Vox does. It's about headroom, and bipolar is easy to do when you have a dedicated center-tapped winding on your custom-spec'd PT. That said, the same charge pump will do +/-9v as easily as it will give you +17v, and with fewer parts than any higher single voltage, and less fiddle-dee-doo with circuit part values. FSFX's point about dirt boxes is well taken, but this is a preamp that can be run clean, too, if that is what is desired, so maybe the headroom does matter, but does +/-9v or +17v possibly give enough?

[lars-musik]

I agree, Ben. Possibly a distinction without a difference: 'preamp vs. overdrive' BUT for the headroom thing (examples for the confusing terms are the "Benson Preamp" that is more or less a "Wampler Plexi Drive" or the "BB Preamp" that is basically a "Tubescreamer"). I never liked my overdrives at higher voltages (except the Klon possibly), because they should be dirty and compress.
This here is just an experiment if I can use the Pathfinder15 as a pedal or maybe in combination with an Harley Benton pedal poweramp and a cabinet as a very portable amp solution.

The 36V charge pump circuit I stole from the Fortin33, just some three diodes and 10µF caps more than the standard 7660S/1054 charge pump.

By the way, I just etched the board. Hopfully there'll be time for this on the weekend.

[antonis]

The 36V charge pump circuit I stole from the Fortin33, just some three diodes and 10µF caps more than the standard 7660S/1054 charge pump.

I'd suggest 10μF and 100nF(ceramic) power supply decoupling caps on each IC Vcc, as close as possible to pin(s) 8, but you've already etched the board..

[lars-musik]

The 36V charge pump circuit I stole from the Fortin33, just some three diodes and 10µF caps more than the standard 7660S/1054 charge pump.

I'd suggest 10μF and 100nF(ceramic) power supply decoupling caps on each IC Vcc, as close as possible to pin(s) 8, but you've already etched the board..

Good hint, thanks. I've soldered such decoupling caps directly to the pads of the ICs before. When I run into noise problems, I'll do that.
Maybe I should have posted a final schematic before etching . Now I'd rather build it and worry later.

[antonis]

Maybe I should have posted a final schematic before etching . Now I'd rather build it and worry later.

You might not worry at all..
(such a voltage multiplication usually involves relatively high ripple but particular circuit low current draw should countervail it..)