DIY power supply schematics

[Tabbey]

Hi,

First post here. 
I am trying to get more into electronics, so I am looking for a schematic for a variable voltage 3 supply rail (+0 to 15v, 0v, -0 to 15v) PSU that can be powered with a 26v (26v 13v 0v outputs) 0.66A AC transformer.
I have done quite a bit of searching on here but have so far come up blank. (I have seen references to Dave's PSU design but can't seem to find it anywhere)

Does anyone have any schematics for this that I could have a look at?

[smashedProton]

Do you want linear or switching?  What Is your application?  It seems that you want a +- dual tracking supply.  To look at the lm317 and its equivilant for negative voltages

[Tabbey]

Do you want linear or switching?  What Is your application?  It seems that you want a +- dual tracking supply.  To look at the lm317 and its equivilant for negative voltages

The application will be breadboard projects, most likely audio circuits, so I won't be needing much current. I am aware of the LM317, it has been in quite a few of the designs I looked at.

I am afraid that I have no idea of the difference between switching and linear supplies other than that linear power supplies require larger heatsinks, so I can't answer that yet.

Thanks,
Tom

Edit: After looking up linear and switch mode power supplies, I would probably have to go for a linear supply as I will be making some audio circuits.

[Paul Price]

Get one more transformer and build two of these, one for the negative and one for the positive.

The switcher circuits below require a very few components, no hot hot linear LM317 Oh My God It Burnt My Finger problems.

However, if you are in a linear mood,  one schematic below is for a zero-volt min output LM317 linear approach, but  the LM2678 type circuits  are the more modern switcher approach, so much more efficient, runs cool, few parts.

[Tabbey]

Get one more transformer and build two of these, one for the negative and one for the positive.

The switcher circuits below require a very few components, no hot hot linear LM317 Oh My God It Burnt My Finger problems.

However, if you are in a linear mood,  one schematic below is for a zero-volt min output LM317 linear approach, but  the LM2678 type circuits  are the more modern switcher approach, so much more efficient, runs cool, few parts.

Thanks for those, I shall look into them. I may just make the switching one and see how it goes with audio circuits, if that fails, linear it is.

[Paul Price]

There are many varieties of the LM2678 type of switcher IC, and this is a very new model and maybe hard to find.
There are also the LM2575, LM2676 etc.   The older ones just mention switch at a lower freq. so they might require a more expensive inductor if you don't easily salvage a nice one from a used PC power supply..

[Tabbey]

Just a few more questions:

I watched Dave's video on the diode/capacitor inverter ladder circuit, I assume that a switch mode voltage regulator puts out a square wave signal. Would it be possible to use that instead of a secondary power circuit?

If I do have to build a second power circuit and wire it in parallel, how do I do that exactly? I assume it would be connect every ground/0v point of the +ve circuit to the +v rail of the -ve section, but I always struggle to get my head around +/-v, and I want to check before I do anything stupid. 

Can both halves of the supply be controlled with the same potentiometer? i.e link the 2 feedbacks together.

Is a bigger inductor better for the output smoothing of the LM2678? How does this value affect the circuit? (trying to learn for future reference  )

[Alana]

Example how to do single pot regulation with LM317/337 but it seems its a lot easier to get dual potentiometer. Most of them would be logaritmic for voltage control but i think i had one or two with linear characteristics [letter a after ohm's value of pot].

[Tabbey]

Right, I am now amassing the components I need for this (finally) and want to double check a couple of things.

I have two 21v and one 26v (0.66A) transformers, how should I wire 2 of these to get an output of 0v, 21v and 42v (or -21v 0v 21v)? (trying to work this stuff out always makes me confuse myself  )

Also, for the wiring the supplies in parallel, is it wired like this? (in the attachment)

[Paul Price]

You aren't specifying which of the wires is the ground reference, but it sure looks like you have these connected in series.

[Tabbey]

You aren't specifying which of the wires is the ground reference, but it sure looks like you have these connected in series.

The ground reference is the lowest rail which becomes the -ve voltage when the middle rail is treated as 0v
So how do I go about making them parallel? Connect the ground rails together?

[Paul Price]

yes and + the outputs as well. Is this what you really want and what are you really trying to do..do you want to parallel two supplies to get twice the output current?

[daveatol]

You want the two output voltages connected in series for a dual rail supply.

For the input voltages, you'll want two isolated supplies - this means two secondary coils on the transformer.

[Tabbey]

I have now built the circuits and have a transformer with 2 secondary coils, I am now struggling to figure out how to wire up the circuits to the transformer to get it to work.

How do I go about this? I have tried using 2 bridge rectifiers wired up as shown in the diagram to get 3 rail power out, but the +ve supply really doesn't like this and the regulator gets very hot very quickly, however the bottom power supply works fine.

Also, an amendment to the switching circuit diagram above, you do not connect the on/off to 0v otherwise the regulator provides no output.

[madires]

Remove the connection between both bridge rectfiers and add one at the output of both regulated rails, if there isn't one yet.

[Tabbey]

Remove the connection between both bridge rectfiers and add one at the output of both regulated rails, if there isn't one yet.

Like this?

[madires]

Remove the connection between both bridge rectfiers and add one at the output of both regulated rails, if there isn't one yet.

Like this?

That way you get two positive rails :-) Move to bottom connection from ground of the bottom regulator to +V of the bottom regulator. Then the ground of the bottom regulater becomes -V because it's xV below ground of the top one.

[Tabbey]

That way you get two positive rails :-) Move to bottom connection from ground of the bottom regulator to +V of the bottom regulator. Then the ground of the bottom regulater becomes -V because it's xV below ground of the top one.

I have done this setup, and I still get the very hot very quickly problem:


I have just removed the link wire and still get the problem, looks like I damaged my regulator trying to figure out how to power it earlier.

[madires]

I have done this setup, and I still get the very hot very quickly problem:

I have just removed the link wire and still get the problem, looks like I damaged my regulator trying to figure out how to power it earlier.

You still got the -V output at the wrong location, it's at the negative side of the bottom C1. If the regulator gets very hot whithout any load it might be broken or there might be a wiring error.

[Tabbey]

I have done this setup, and I still get the very hot very quickly problem:

I have just removed the link wire and still get the problem, looks like I damaged my regulator trying to figure out how to power it earlier.

You still got the -V output at the wrong location, it's at the negative side of the bottom C1. If the regulator gets very hot whithout any load it might be broken or there might be a wiring error.

There is no load and I just forgot to change the label on that output, and as far as I am aware, the top and bottom circuits are identical and they worked individually earlier, but I evidently did something very wrong this morning when trying to power it without a clue on what to do because I blew up the input cap on the top circuit (a first for me, and it took me by suprise  )

Well, now I know to ask first if I don't know what I'm doing luckily the regulator wasn't too expensive.