How to isolate DC power outputs

[Fancy Lime]

Hi there,

I have a technical question. There is an increasing number of DC bricks on the market, which have several galvanically isolated outputs. Some, maybe even most of these are constructed as "distributor bricks" that are fed DC by a single wall-wart, which is the actual AC-DC converter (e.g. MXR iso-brick). Meaning the isolation takes place on DC, not AC. I am wondering how that works. For AC, isolation is easy. But for DC? They don't convert it back to AC, isolate, and convert it back-back to DC once again, do they? Not saying they shouldn't. Just wondering.

Thanks and cheers,
Andy

[amz-fx]

They don't convert it back to AC, isolate, and convert it back-back to DC once again, do they?

That's one common way.

There are small modules that have a driver and a tiny transformer inside the chip - bigger than an op amp but still fairly small.

http://www.muzique.com/lab/9v_iso.htm

Best regards, Jack

[Fancy Lime]

Thanks, Jack!

Another thing I have problems with wrapping my head around is: why? Sure, if we have both positive and negative ground pedals we want to supply from the same unit, we need to isolate those two groups. And we can bridge two isolated power outs to get double the voltage. But the main argumentin most application cases is noise. But why would several linked pedals fed with isolated outputs create less noise. They still share a ground via the audio connection. And if that is so, is it really a problem if the "isolated" power outs have connected negative pins and only the positive lines are really isolated, as seems to be the case for some of the cheaper "iso-brick" -style supplies from the land of cheap electronics (only for the "negative ground only, no bridging" case of course)?

BTW, has anyone ever used an isolation module like Jack's inside a pedal to then bridge it back to the original for voltage doubling? It's more complicated and probably more expensive than even the fancier of the usual CMOS voltage converters (LT1054 ...) but I would expect it to potentially be quieter and it would certainly deliver more current. There are not a whole lot of applications, where this would be important but I can think of a few.

Cheers,
Andy

[amz-fx]

The dc-dc converters are available in versions that step up the output as well as providing isolation:

http://www.muzique.com/lab/24v_iso.htm

regards, Jack

[anotherjim]

I use pseudo isolation - put some resistance in series with the 0v output for each DC outlet. 4R7 to 10R will do. This really does help reduce hum loop problems.

[mozz]

I would think that a diode on each output would isolate it. Or each output would have a separate 7809. But parts is money, is profit, so whichever way is cheaper. Weber used to sell a transformer with a bunch of the same secondaries but I doubt any manufacturer is using something like that.

[Fancy Lime]

The dc-dc converters are available in versions that step up the output as well as providing isolation:

http://www.muzique.com/lab/24v_iso.htm

regards, Jack

Oooh, nice! Now that you say they exist I feel a bit silly for not knowing or at least assuming they would. Have to go hunting the right type on the interwebs now...

Thanks,
Andy

[PRR]

> They don't convert it back to AC, isolate, and convert it back-back to DC once again, do they?

In 1954, that was awfully awkward. Mechanical buzzer, transformer, rectifier tube. Nevertheless all car radios did it to get 250VDC from 6VDC. (At higher power, police/taxi radio, a DC motor turned a DC generator wound for high voltage.)

By 1964 this was less awkward. Transistor inverters and solid diodes. Still not cheap.

Less expensive modules proliferated in stray corners of the field such as telephony racks.

A turning point came around 1990. Millions of network cards needed an isolated DC supply. Millions of isolated 5V:9V(?) converter modules made cheap.

Going beyond what we do for pedals: every cellphone takes a 3.6V battry and buzzes it up or down for multiple outputs, 1.7V 2.6V 5.0V so each stage gets the exact voltage to do its job efficiently.

So yeah, it seems stupid to transformer DC, but it is so useful that it has become a commodity with overwhelming choice.

[Fancy Lime]

In 1954, that was awfully awkward. Mechanical buzzer, transformer, rectifier tube. Nevertheless all car radios did it to get 250VDC from 6VDC. (At higher power, police/taxi radio, a DC motor turned a DC generator wound for high voltage.)

I must say, I love the old "this is crazy but we have no other way of doing it so here goes" solutions. Thinking outside a box, which wouldn't be invented for another bunch of decades...

So yeah, it seems stupid to transformer DC, but it is so useful that it has become a commodity with overwhelming choice.

Well, I wouldn't say it seems stupid. It seems amazingly useful. It is just one of those things that shows how for some things doing everything AC would be so much easier. That is, if it weren't so completely impossible for other things. Quite a teachable bit of circuitry, is all I'm saying.

Cheers,
Andy