I have been working on a DC dual-output power supply, and have received some great help from the forum on a couple things. I now have this idea to switch +5V into the negative output, which would allow me to convert the negative output into a fixed 5V source (when the negative output is already adjusted down to 0V). My question is, will this work with only a SPST switch?

Here is my logic as to why it will work: Once the negative output is adjusted to 0V (bottom 5K pot down to 0Ω), shorting 5V to that point will not interfere with anything because 1) the LM337 holds 1.2V across the 208Ω resistor no matter what, and 2) the 1uF capacitor to ground will stop the 5V from shorting to ground. I might be going wrong when it comes to the fact that I am connecting 5V to the output of the 337. Will current be sent to or away from the chip in a way it shouldn't be sent?

Circled in red is what I was thinking of doing. Something about it doesn't seem right but I can't quite put my finger on it. Does anyone know if this is a good/bad idea, or what techniques I can use to analyze this to find out if there is a problem?

 

A positive regulator into a negative regulator will cause the fight of their (short) lives.
If you're lucky, they will just shut down on overheat. If one of them wins, it will shove current backwards through the other regulator and smoke it.

 

Just the answer I needed. So connecting positive voltage to the output of a negative regulator is a bad idea?

I'll need a SPDT switch to disconnect the negative regulator to the output first?

 

Use a dpdt if you want to, just don't connect 2 unlimited (no resistor) outputs together.

 

Just a note.
You don't really need two pots unless you want independent control of the + and - supplies.

317 / 337 based dual supplies will happily track from one pot and can be connected to provide the opposite polarity reference to permit operation down to zero volts. Elektor did an excellent project psu based on this a few years back.

 

I went with two pots for independent control. I may find myself adjusting them both to the same value quite a bit but eh, oh well! That's one reason I wanted the ability to convert the negative output to +5V, for versatility. The current draw won't be be able to go up to much, but for small circuits it should do alright.

 

Just a note.
You don't really need two pots unless you want independent control of the + and - supplies.

317 / 337 based dual supplies will happily track from one pot and can be connected to provide the opposite polarity reference to permit operation down to zero volts. Elektor did an excellent project psu based on this a few years back.

I believe you need to add an op amp to make a tracking supply with only one pot, for example as shown here.

 

Interesting.

On second thought, if I want to use this circuit as a power supply with two + outputs, can I do that by making the -V output "ground"? i.e. reference +V and GND to -V? For example if I want +10V and +5V can I adjust -V to -5V, and adjust +V to 5V?

 

The mantle is adiabatic, impermeable to heat transfer

With proper isolation, yes.
Professional units have this characterestic as standard.

Note that the supply is then floating and a separate earth teminal is provided which can be connected to any terminal, including the zero.

I cannot access the Elektor art tonight to comment further on control of twin regs, but they may well have used and op amp - they delight in them. However their circuit has appeared several times in several guises and provides true zero out, ie below the reg's minimum. This is, of course, done by returning the regs control terminals to an opposite polarity voltage.

 

Why not just add another output terminal, it will cost less than a switch.

 

Before you try some other else, you can try the way is similar as yours.

crutschow mentioned that the linking page, it's the auto tracking method, that's a good way to do.