Trying to Hookup ATX Power Supply Leads in Series

Harry Stoner

Trying to do wacky things that are simple conceptually. I have some old ATX power supplies. I would like a robust 15V or higher output. So I took one of the 4 pin headers and connected it to my breadboard. It has R-BLK-BLK-Y wires corresponding to +5V, GND, GND, +12V. The 5V measures unloaded as 5.28 and the +12V measures 10.86V.

So my simple mind says put them in series like batteries and get ~16V. But it didn't work and I don't understand why.

So, can I actually hook these up in series to get a higher voltage? If so, how? Would this work if the two voltages were the same? I didn't find any definitive answers by googling.

I know there are other ways to get power as I desire, but today I am trying this.

Thanks.

Harry

Chris Savage

Harry,

·· No you can't do that.· Computer Power Supplies aren't designed to operate that way.



▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Chris Savage
Parallax Tech Support
csavage@parallax.com

Paul Baker

I have seen many times people getting not specified voltages from ATX power supplies. The most common voltage is 7V by using +5V as your ground and +12V as your Vdd. By this theory you could use the -5V as your ground·and +12V as your Vdd to get 17V. Now the supplies were not designed to do this, so your mileage may vary. And the regulation of the voltage produced by this is not going to be as great as using it the way it was designed. Above all, don't use this configuration to drive high current loads, or you may hear a popping noise from your supply and smell the magic smoke (my supply burns out yearly, and I haven't done what I describe above except once to supply a fan with 7V, which didnt cause the supply to die)

Harry Stoner

Thanks fellas for the quick replies. Paul, I saw the 7V value when I had the +5V and +12V in parallel. Yes I know this is not good (tried it for a couple of seconds).

I will go back to regulator circuits. I am trying to avoid a voltage sag under load. I know I need a beefy enough supply to begin with. I had a 7812 regulator circuit but an inrush current spike knocked that out. I will try an LM317 regulator next. It is adjustable and I would actually prefer more than 12V in this case, so this could be better.

If that still fails I think it would feasible to use parallel regulators (e.g. a pair of 7812 regulators), as I believe the current spike would be less than 2A. Is there anything evil in doing this?

BTW I also bought an off-the-shelf power supply that advertised (in addition to +5V) +15V/6A. But when I power it up with no load it is only putting out ~12.5V, and it is fluctuating a lot. I tried using it on my project and it did not work at all. Would this be defective or do I need to place dummy loads on it (the doc said .5A min on the +5V circuit and 0A min on the others).

Thanks.

Harry

pvale

If you have several of the power supplies, connect the outputs in series over sveral supplies. I have a 33A, 13.6volts supply to my amateur radio equipment using the 5V outputs of 3 supplies in series, running through 2 silicon diodes to drop the voltage to 13.6 or so. Works fine. I don't think the ouputs of a computer power supply are grounded to the case. My supplies are sitting below my workbench on a shelf. I have enough load that stays on all the time to load the supplies, so haven't needed to use loading resistors. I also have (3) isolated 12V supplies at a few amps each for testing and playing around. Regulation on the main 13.6V output seems pretty tight. The biggest connected load is my HF rig at about 22 amps, and it has no problems putting out it's full 100 watts even on 10 meter FM, which is a power hungry load. I have everything fused. I have overloaded my small 12V outputs before and shut down one of the supplies. I have all three supplies plugged into an outlet strip so that I can shut them all off at the same time. ATX supplies do have a pin that needs to be grounded to get them to operate. The pinout for the motherboard cable is widely available on the internet.

Beau Schwabe

Chris,

I agree with you...The analogy that I have heard before with connecting power supplies in series,
is like a bicycle race. If one guy falls, you have a chain reaction. Just be careful if you choose to
do this and absolutely under no circumstance should you parallel different voltage potentials.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔

---

Beau Schwabe - Mask Designer III

National Semiconductor Corporation
Latest Company News
(Communication Interface Division)
500 Pinnacle Court, Suite 525
Mail Stop GA1
Norcross,GA 30071

---

Harry Stoner

Thanks for the additional input fellas. pvale, I think I came across a web page of yours on your project - great results but I don't think I'd want to tackle it myself. Beau - sounds like good advice. Always stuff to learn.

Harry