SMPS mod

[Mach5]

I have a rather embarrasing dumb question so bear with me.

I have four matching powerman 300w SMPS supplies with the goal of a +/-24v gainclone supply. Anybody familiar with ATX supplies would know -12v amp limit of many of them. I've been trying to mod mine to gain the extra -12v side.

Step number one, schematic. 2. find the weak link 3. upgrade
Problem is this is a very tightly packed board. Not very easy to read the P/N. After 6 hours of desoldering I've copied down about half of the board and have yet to trace the -12v supply!

It just now dawned on me that someone may have tried this before. You know who you are(if you see a guy laughing at me sitting next to you, that's him)

So after a long explanation here is a short question(s). In the attached photo is the circled tranny the -12v supply?(probably yes)
If I wind a new tranny from some enamel wire and replace the depicted will there be many adverse effects?(e.g. to much current draw in the wrong place)

Since where on dumb questions maybe someone would also like to answer another that's been nagging me: what are the advantages of a DC servo?

[DigitalJunkie]

Ah-Ha!

I can't see the attachment (yet?) but,here's a few possible pointers,from memory..
The -12V rail is usually supplied from the same winding as the +12V rail,it just has smaller rectifier diodes/filter caps/etc.,and they all have their polarity reversed,since it's a -V supply.
Some PSU's have a separate smaller winding for the -V,that is usually then fed to the rectifier/filter/and a set of negative regulators (7905 and 7912) for the -V outputs. You can't use this winding,as it's too small for any real current..BUT..you can still use this type of supply and get +/-12V

The way I (and a friend) have done it in the past is to just clone the main +12V rail,but reverse the rectifier diode/cap polarity,so you get a negative voltage instead of positive.
So,You end up with the original +12V rail,and you add a second -V rail,complete with the diodes/caps/choke,etc. The 'new' -12V rail will track with the +12V rail,so you still get regulation and stuff too.

I'll see if I can draw up a rough pic of what I mean,if it isn't clear.

Edit: there are a few PC SMPS schematics here:
http://www.smps.us/computer-power-supply.html
Most of them contain fairly similar circuitry.

[DigitalJunkie]

Okay,here's an edited up schematic for ya:
http://i199.photobucket.com/albums/a...0/atxpshax.png

You'll need to add the parts in red.
The first coil after the rectifier diodes should ideally be a winding on the big main toroidal choke on the output (usually yellow),mind the polarity of the winding. In the past,I have just unsoldered the choke,and used the +5V windings (usually 2 wires in parallel) one wire for the +12V rail,and one for the -12V rail..and wired the old +12V winding back onto the +5v output..a bit of lead-juggling. That trick only works on the toroidal cores/chokes without the plastic 'foot' on them,with all the lead wires. The ones that are just soldered straight in on the magnet wire leads are easiest..

I'm not entirely sure this procedure is needed,but it can't hurt. You can probably get away without the first choke at all. (regulation may suffer?)

For the second choke,you'll need to find one that's similar in size/current handling as the original one on the +12V rail.. (Old dead PC PSU's are a great source for the extra parts you'll need.)

You're basically looking to make a clone of the existing +12V rail,but with the polarity of everything swapped,so you get a high current -12V rail.

NOTE: Power output will be halved on each rail. IE;if you have a SMPS that says 20A on the 12V rail,it will now be 20A *TOTAL* for both rails,or 10A per rail. You can draw up to 10A on the +V rail,and 10A on the -V rail,at the same time. The TOTAL power stays the same.

Anyways,Hope this helps!

[Mach5]

Good website link
This schematic seems closer to mine,it's hard to know since I haven't yet copied it all down yet, transistors aren't the same of course...but the output is right, -12v supplies the -5 via 7905.
http://www.smpspowersupply.com/ATX_p..._schematic.pdf

I guess I haven't got my "training wheels" off just yet. You can see the picture of the tranny in question off my site here:
http://bngreer.googlepages.com/induc...full;init:.JPG

Through a bit of wishful thinking I managed to convince myself that the I could easily clip the lead and install a new transformer in it's place, bypass the rest of the board and come out the top(no soldering!)
A look at the schematic makes this seem unlikely

The weird thing about the trannys on this board is 1. they are all clearly numbered and 2. the P/Ns apparently don't exist!(on the web)
Unfortunetaly, they all feature terminals soldered directly to the board(and very tight working space). I think I could isolate some of terminals with a bit of dremel work to the board.

Wait....did you mean remove the tranny and rewind it? Or just add another second stage and attach to the +5v? Hmm.
It'll be obvious after I look at in the morning

[Mach5]

Thanks for the help by the way. Quality.