Lab Power Supply From a Hacked ATX Computer Power Supply

Joined
Jan 9, 2007
Messages
772
Location
Up where the air is fresh and clean
Stumbled across one of these devices while walking around Maker Faire with Fechter last weekend..

ATX-to-Lab-PSU-8.JPG


..and it got me thinking, "I have a couple of these power supplies -pulled from otherwise dead computers- just sitting in my garage.. hmmm."
Anyway, I found a cool 'How To' @ http://www.wikihow.com/Convert-a-Computer-ATX-Power-Supply-to-a-Lab-Power-Supply ..thought other spherites might find this useful.
 
I started a thread called "Power Supplies" a while ago. We talked about computer power supplies and they do look excellent when it comes to price per energy delivered.

http://endless-sphere.com/forums/viewtopic.php?f=14&t=3444&p=50364&hilit=power+supplies#p50364

The problems (we learned) were:

:arrow: Many computer supplies can't really deliver their stated wattage without burning out in a short period of time.

:arrow: Since there is no built in current limiting mechanism there's no way to guarantee that you get what you are looking for as far as current.

...most computers have very low power requirements and so the power supply simply sits and does very little. A charging situation tends to absorb all you can give it and that's not well adapted to computer power supplies.

Some computer power supplies have built in current regulation and those are the one's to use if you go this way...
 
a lot of what "safe" says may be true of the power supplies bought in the bargain bin. you know the ones that go new for $15.00 or $20.00.

but many of the better power supplies will produce a heck of a lot more power reliably for a long period of time. and sometimes you can find these in the used parts bin at the local Salvation army or St. Vincent DePaul thrift shops. just look for recognizable name brands... Enerpac, Thermaltake USA etc. look for the ones rated for 500W plus.

i bought a used Thermaltake 650W powersupply for $5.00 for a friend who uses it to fast charge his RC truck packs. on race day it will happily supply 20+ amps for hours on end to charge all of the clubs packs during the meeting.

rick
 
:?: Somebody crapped on my thread!! Hmmm.. :| ACTIVATE -

:!: 'Safe' Filter

http://endless-sphere.com/forums/viewtopic.php?f=1&t=4343

BZZRrrp..
:shock: 8)
 
I've been pondering on how to get a reasonably high power low cost power supply/charger for my A123 pack for some time.

I have copies of an article from several years back on how to tear down and rewind a PC PSU for any arbitrary voltage. I modified a 360W one to produce 24V at 20A quite successfully, but it was a lot of work - it was also an older, simpler generation of PSU.
Current PSUs are smarter and probably more difficult to hack - but, here's an idea:

Take a high power standard PSU, chop all the wires off and apply a 5W load resistor on the 5V supply to ensure the PSU starts reliably. Now, identify the 'ends' of the 12V windings from the transformer (this is usually a single centre-tapped winding for both 5v and 12V - ie 12-5-0-5-12, there are then two centre-tapped pairs of Schottky rectifiers for the respective supplies) There is no separate adjustment for 12V, it is locked to the +5V rail.

So, we now have a 24V AC supply at a few tens of KHz.

We could:

1) Put a schottky bridge rectifier across the ends and get 28-30V DC after rectification and filtering (no current limit or control, but reasonably efficient) And on top of that, you can often get at least 50% change in voltage by tweaking the 5V adjust pot on the board, on old models anyway. So this simple approach might get you 45VDC or even more.

Warning, since the centre-tap of the winding is earthed, the output voltage is no longer referenced to earth, due to the bridge rec.

2) Put a Cockcroft-Walton voltage multiplier (chain of diodes and capacitors) to get any multiple of 24V that we want (not very efficient, but can generate quite a high voltage, perhaps if you have a 72V or higher but low capacity battery pack, still need some current control) By using two identical multiplier chains - one for each output winding of the transformer - and summing the outputs, better efficiency may be achieved and still retain a supply referenced to ground. Capacitors for this are a bit special and may not be cheap - unless scrapping old PSUs...


3) get a transformer core similar to the one in the PSU (by scrapping an old PSU) and rewind it - feed it 24V from the main winding and, say, dual centre-tapped 48-0-48V on the output side. Use a schottky split rectifier as in the original PSU - you now have 55-60VDC out, after rectification and filtering - depending on load. now use that raw DC to feed a constant current charger circuit.

A few guesstimated numbers: take a 500W PC PSU, 450W of that is probably available from the 5/12V windings. Lets assume 80% efficiency on our home-wound additional transformer. That gives well in excess of 300W for charging use, but it might get a bit warm in continuous operation, so derate a bit or add more airflow if that gets to be a problem.

While this has the disadvantage of having to rewind a transformer, at least you do not have to dig into the main PSU much at all and you can easily get any voltage you need.

There are so many models of PSU (and they change so rapidly) that trying to produce detailed methods for modifying a particular PSU are doomed to failure, hence my stab at a generic method. I don't know if the wire thickness changes between the 5V and 12V parts of the winding - it did not on the old one I rewound, if that has changed, output power would be limited somewhat.
 
Back
Top