bipolar power supply using positive regulators only ?

[StephenGiles]

Interesting thread over at electro-music:

http://electro-music.com/forum/viewtopic.php?t=69183

[anotherjim]

Commonly done, also with psu modules. What catches many out is that the "negative" supply must have 0v to heatsink/chassis isolation, since it's 0v is now the negative supply rail. Some psu builds isolate heatsinks anyway, even if they could be common to 0v, which makes it easier to use them as negative regulators.

[amptramp]

The heat sink on a regulator chip is usually the most negative terminal in the circuit.  For the LM317/LT1083 this does not mean much since it is the adjustment pin on a 317 and the input power on a 337 negative regulator and neither one should go to ground.  The advantage is being able to use one line item type of regulator.  As you can see from the schematic, there are disadvantages: two bridges required (which reduces efficiency by adding one more diode drop to each output) and a transformer with two isolated output windings.

[Djentronio]

Commonly done, also with psu modules. What catches many out is that the "negative" supply must have 0v to heatsink/chassis isolation, since it's 0v is now the negative supply rail. Some psu builds isolate heatsinks anyway, even if they could be common to 0v, which makes it easier to use them as negative regulators.

I thought that whatever was common voltage should be linked to the chassis. So you have a +9/-9v supply and the common 0v is not the same thing you wire to the chassis?

[PRR]

> 0v is not the same thing you wire to the chassis?

Ron has slipped into Positive-centric speech habits.

*IF* you take the negative side alone, and label it "+9V" and "0V", then "+9V" goes to whole-circuit "common" and "0V" (of the negative sub-supply) is now "-9V".

Which is why we have Negative regulators. It is simpler. Less likely to confuse us. (And avoids a separate winding and a rectifier.)

[anotherjim]

I think Negative regulator devices came later, so Positive regulators had to be used once upon a time, certainly in the 1970's. I don't think there was ever a negative version of the uA723. It is hard to find more recent bi-polar designs like that because complementary negative devices are commonly available. But, many psu modules don't come in either negative or bi-polar versions, you are expected to wire them accordingly.

I do think it's easier to make assembly errors is you only have positive devices. One of my most common build errors is to put a zener diode in the wrong way - brain auto-pilot sees a diode & wants cathode to negative side of the circuit. Brain auto-pilot would want to do the same thing with the 0v of a regulator.

A good PSU read here...
http://sound.whsites.net/articles/power-supplies2.htm
...goes above and beyond what we usually need, but it's all worth a look.  Specifically, go to section 9.
Section 7 is interesting in respect of other recent discussions

[ElectricDruid]

My brain auto-pilot always screws me up with negative regulators anyway...the number of times I've put the electrolytic cap following the regulator the wrong way around is getting embarrassing. My brain sees "power rail" and thinks "positive" despite many schematic markings to the contrary.
It's become one of those things where I just have to double-check my double-checking.

T.

[Rob Strand]

think Negative regulator devices came later, so Positive regulators had to be used once upon a time, certainly in the 1970's. I don't think there was ever a negative version of the uA723.

I seemed to remember the Fairchild datasheet showing a way to connected it like a negative regulator.  I saw it used in a commercial multi-rail PSU once.