N times 9V PSU, complete galvanic isolation possible for each section?

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

Kingston

Well-known member
Joined
Nov 1, 2005
Messages
3,716
Location
Helsinki, Finland
I have a big bunch of stomp boxes and would like a complete PSU solution for powering these. I've have seen how the big boys and all guitar effects DIY sites are doing it and it's simple, no problem. At its most complex just a preregulated ~12VDC going to a bunch of separate 9VDC regulators fanning out to each stomp. Still very cheap.

I foresee one problem: The common ground reference shared between all effects finally entering either guitar amp or some preamp DI input.

Which finally brings me to the question: is it possible to build a PSU that has a battery-like complete "galvanic isolation" from all shared grounds. I realise a battery doesn't actually have any isolation from the grounds of a shared system, but at least it in no way propagates ground loops, and allows an isolated floating section.

This is purely theoretical musing, it might be expensive, unnecessary and overkill considering the problem. But inquiring minds want to know.

Also there might be some incredibly easy hack I've overlooked to solve this. Feel free to point out.

Thanks,
Mike
 
Kingston said:
I foresee one problem: The common ground reference shared between all effects finally entering either guitar amp or some preamp DI input.

Which finally brings me to the question: is it possible to build a PSU that has a battery-like complete "galvanic isolation" from all shared grounds.
That's the way the most expensive (Eventide, Voodoo Lab, T-Rex,...) do. They have multiple secondaries with separate bridge rectifier, smoothing cap and regulator.
I've had the opportunity to compare these with a much simpler type that uses  a single secondary, single regulator and I couldn't detect any difference, even shaking the connections made no difference, which means the ground loop was not detectable.
My pedal board has:
Wireless receiver
Tuner
Clean boost
Dirt
Delay
Line 6 rotary speaker
Volume pedal
Amp simulator
Graphic EQ
Output A/B switch
The Line 6 Roto Machine is some kind of a pig; it draws a lot of current and sends back spikes. If the ground connections (audio and PSU) are not perfect, it pollutes the whole system.
I have tried  the aforementioned Voodoo Lab, I have tried a separate PSU for it, I have tried a non-separate feed, there is no difference.
Which is not very surprising, since the ground loop is quite contained.
This is purely theoretical musing, it might be expensive, unnecessary and overkill considering the problem.
  that's more or less the conclusion I draw.
Now it may be different if I had FX loops with several sends & returns...
 
abbey road d enfer said:
They have multiple secondaries with separate bridge rectifier, smoothing cap and regulator.

I thought of this, but ground ref is still the same. All these sections must have a star ground point somewhere. A plain multiple regulator based solution is easy to design in a very similar way.

Or do you mean the sections are floating? Isn't that kind of dangerous, and what is the ground return path?
 
Kingston said:
abbey road d enfer said:
They have multiple secondaries with separate bridge rectifier, smoothing cap and regulator.
I thought of this, but ground ref is still the same. All these sections must have a star ground point somewhere.
No, they don't.
Or do you mean the sections are floating?
Indeed.
Isn't that kind of dangerous, and what is the ground return path?
It could be dangerous if the transformer developed a leakage between primary and secondaries. That's why the xfmr must be spec'd for double insulation.
 
abbey road d enfer said:
abbey road d enfer said:
They have multiple secondaries with separate bridge rectifier, smoothing cap and regulator.
 
I thought of this, but ground ref is still the same. All these sections must have a star ground point somewhere.
No, they don't.

Interesting. I didn't know I could do that. But from my experience from bigger projects that have balanced audio I/O, a fully floating system is always significantly more noisy than the same configuration with a single well placed ground ref. Even if audio I/O has complete galvanic isolation. Dirt has nowhere to go, and ends up somewhere in the audio path eventually.

I don't understand how this type of floating would work any better for system that is unbalanced throughout. Or does it just work "well enough"?
 
Kingston said:
But from my experience from bigger projects that have balanced audio I/O, a fully floating system is always significantly more noisy than the same configuration with a single well placed ground ref. Even if audio I/O has complete galvanic isolation. Dirt has nowhere to go, and ends up somewhere in the audio path eventually.
There is nothing like a "fully floating system that have balanced audio I/O". All balanced audio equipment has enormous leakage from mains to chassis ground; that's why (and for safety reasons) it is necessary to connect the chassis to technical earth. Floating a piece of gear that is meant to be earthed is a receipe for injecting all sorts of crap in the audio ground.
I don't understand how this type of floating would work any better for system that is unbalanced throughout. Or does it just work "well enough"?
The idea is that there is no ground loop because the PSU grounds are floating respective to each other and as a consequence there is no current flowing through the audio cable shields. A perfectly valid concept, but in practice it doesn't make much difference, because the dominant source of noise in a pedal board is not ground loop current.
 
Back when I worked at Peavey we made a rack mount lump eliminator. Since I and other engineers used our standard 16VAc 1A wall wart differently (sometimes FW rectified other time HW rectified, or whatever). So you couldn't use a common ground.

I tooled up a custom transformer with 6 floating 1A secondary windings. Added some lights to the rack package and I had the product that everybody was asking for, except for one little detail. The cost to make and sell this monster was far more than the customers were willing to pay. because of the low volume these sold in the custom transformer was expensive and the product price reflected that.

In hindsight, a few decades of hindsight, I suspect an electronic solution could be cheaper.  Basically you need to deliver a modest power/current AC signal that can be floating wrt ground say +/- peak AC voltage.  I hypothetically visualized a pass implementation, but these days I'd be tempted make it switching. 

Realistically to power up a bunch of 9V pedals, especially if they are operating from DC wall warts, this should be much simpler. They don't need all that much current so you can put a 10 or 20 ohm resistor in series with the ground, and deliver 9V on top of whatever that ground is. If you want to get fancy, I would sense how much current is coming out of the +9V lead, and suck that same exact current into the ground lead, so there shouldn't be stray PS current leaking between connected units if their ground is common. Another approach is to servo this 9v ground to be 0V, which may be a little easier and more accurate. So 0v but still 10 or 20 ohm impedance between it and common PS ground.

or not...

JR
 
abbey road d enfer said:
The idea is that there is no ground loop because the PSU grounds are floating respective to each other and as a consequence there is no current flowing through the audio cable shields. A perfectly valid concept, but in practice it doesn't make much difference, because the dominant source of noise in a pedal board is not ground loop current.

no ground loops, but what about ground dirt current in general? How does it know where to go? There's a bridge rectifier involved, and at least two filter caps (around the regulator). It somehow doesn't fit in my head the system will just happily float perfectly still with the filter caps and a three-term regulator. Surely there can be situations where something in the system being powered decides to become a ground reference, and now something devious like rectifier crud current happily shoots in that direction and leaks into audio circuits. We got rid of ground loops and replaced it with another possibly worse problem?
 
Kingston said:
no ground loops, but what about ground dirt current in general?
What "ground dirt current"? If you mean the dirty current resulting from rectification, it is perfectly localised by making it floating. If you're talking about EMI current, first they are minimised by the double insulation, second they go to the audio ground. Their consequence there is most certainly less than those resulting from ground loops.
 
ok this localisation is new to me. I assumed currents would follow the path of least resistance and that in bad designs this might lead to surprises. I have been under the impression nearly all common audio design, even simple amps, would generate enough dirty ground currents they always need to have a direct path of least resistance route to chassis/safety ground.
 
For a 9v DC distribution system, there shouldn't be diode or PS rectification ground noise from the pedals to deal with.

You may be able to make it work with simple approach you first described, with say 10-20 ohms series resistors in each ground leg.

If this causes a DC voltage build up on the ground that can click between units, add a servo to correct them back to 0V (note this means a similar negative PS current) Or perhaps source the 9V regulated supply from the middle of a larger PS voltage supply, so the current out the 9v roughly balances the current into each ground return. )

Try it first the simple way. Sometimes the simple answer is the right one.

JR
 
adding some dummy resistance. of course, that makes sense, at least for the very lightweight stomps. For the current drinkers I'll experiment with the servo perhaps.

Thanks everyone. have plenty of information to satisfy my curiosity for now.
 

Latest posts

Back
Top