• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

Dual PSU How-Tos

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.
think of all the extra posts you would have if you were deleting those bumps... that thread would be like 5 more pages:p

If only it were a sticky...
I used it just the other day when I put an extra PSU in my computer to try to fix my low rails...
 
I'l really like to not have to keep bumping this. You'd think a former sticky with over 7300 views would get some consideration.
 
heresa question...i know not many of us are running dual cpu rigs but i am. my mobo requires a 400watt PSU. i had to get rid of my hardly used 350watt due to a motherobard change. My question is, do the 2 PSUs hafta be the same wattage? basically what your doing here is shorting the correct pins so that the 2nd PSU powers on and doesnt connect into the motherboard correct?

im planning on adding more drives and cooling which is why im interested.
 
No, they don'ty ahve to be the same wattage. And Yes, you are just making it turn on when the other ones is turned on by the motherboard - I use my second to power my HDDs, optical drives and fans and then one to power just the motherboard.
 
Now my other computer is lying on the floor open and ready for dual PSU. I am gonna use a relay to power up the second PSU. Then I'm gonna set it up as CrashOveride.
Does this work right away, or do I have to do something to the ground as mentioned earlier?
Also, do I need a load on 3.3 V to use the 5 V?
 
i was wondering, if you mount your dual psu's in a wooden frame like im thinking, wouldnt you need to find the grounds in the power suplies because there not attatched to ground right?
 
alrighty, seeing this thread has got me curious.

im currently stuck at a max o/c of 2.7 gig for a 2.4 chip, 149 fsb. anythin above that and the system wont boot - freezes at win2k graphical bootscreen (gets past the bootloader) and has to be hard reset.

was thinking about the psu - ive only got a standard aopen 300w psu that came with the case. looking at speedfan readings (i dont have a multimeter, sorry!) the 12v seems to fluctuate more at the o/ced speeds, but the 3 and 5v rails stay put (although one goes up, one goes down under cpu load).

ive got a spare 200w at psu somewhere and was wondering if i can hook my hdd, cdwriter and poss graphics card up to it - its a radeon 9700 pro with its own molex connector. thing is, if the output from the molex doesnt match the output from the mobo, maybe theres a problem?

still, if i was to put the hdd and cdrw on the at psu, how can you syncronise the power on of both - is there an atx pin that turns the power on or something?

thanks.
 
Welcome to the forums.

You'd probably have to use the relay method, which would ensure that both of them kicked on at the same time, also remembering that the AT PSU has an independent power switch(so you may have to power them on one immediately after the other). Take a look at the guides and see if one method seems better than another. Outputs from the molex's should be the same at 5V and 12v no matter which PSU.
 
thanks for that. ill give it a go...does the power on sync really matter thaaaat much - if you power up an hdd, then the mobo the hdd should just sit idle until its detected surely? saves me rigging up a relay for testing purposes, to see if the psu is causing the problem in the first place...
 
As long as you power up the drive(s) before the BIOS runs the POST, you should be OK. Otherwise the mobo won't find the drives and you'll end up rebooting.
 
so far so good - found an old at psu, hard wired the power cables so it just turns on as soon as u plug it in, and have it running hdd as i speak...just not in the system it was intended for (thought it best to test the setup in old crappy pc before blowing up nice new machine...)

the only weird things are that a) the main system psu has to have power (but not be on) otherwise the drive doesnt spin up (probably something to do with the ide signal...) and b) when you shut the system down the drive shutsdown and if you turn the system back on without turning the mains off and on again for it to spin up it gets a bit unhappy (with strange high pitched noises coming from psu) :-S doesnt seem to have done any damage tho...
 
well, my o/cing limit if def due to the power supply - unplugged the usb hub for the system and its now running at 152 fsb no problem! 2.73GHz from a 2.4 cpu, and so far still stock voltages. cant be bad...

thing is, im starting to wonder how the psu is taking this, cos it must be under a pretty heavy load if all it takes is a usb hub with some leds to push it over - 500ma max, at 5 v = 2.5W :eek:

at the higher speeds im getting even bigger voltage drops (well, according to speedfan anyway) but even the bios is reporting biiiig fluctuations. time to move in the second psu methinks...
 
Ok so I have been looking into hooking up dual PSUs to booster my ailing +5 line (4.6-4.68 under load by DMM @ Molex) I thought I had most of the issues ironed out in my head until I was browsing ProCooling's Article database. Yo_Duh's Load Sharing Article made me wonder if things were as easy as I thought.

My plan is to use two unequal wattage PSUs, combine the +5v rail, and use the +12v rail from the weaker PSU to power my perherpials. Like This.

The issue I'm having trouble with comes from Yo_Duh's article where he compares previous methods of using 2 PSUs. Specifically the Cons of the method I'm using: "Possibility of the supplies attempting to charge the other one resulting in some serious trouble."

I just want to make sure that as long as I have pin 14 from both PSUs shared & a common ground, they both should receive the power off and on signals simultaneously, and there should not be a risk of just 1 PSU running at one time.

I also wouldn't mind confirming that with 2 ATX PSUs I do NOT need to isolate the ground of the 2nd PSU?


Thanks to anyone who can help me clairify these points.
 
To be honest, I've never heard of the problem he mentions. All of the options linked in the sticky work just fine, and following the wiring diagrams should produce good results. The article you point out seems to indicate a fix on individual lines, and adds a layer of complexity that seems unnecessary.
 
Back