[SOLVED] Powering a computer without a standard ATX PSU?

[Myronazz]

Hello, so here is a potentially stupid question:

So, let's say I have a non-standard power supply that can provide the necessary rails (+12V, +5V, etc.) to the computer if I wire it up to an ATX header properly, but it doesn't have the little PWR_GOOD rail. I know that the motherboard expects it to go into a high state of five volts as an indication of power stability, but not too soon, and not too late, it has to be within a certain time range as per the ATX standard.

The question is... Can I fool it? If I have a non-standard power supply, that I know is good, but doesn't have the PWR_GOOD rail, I was thinking of how I could provide a signal manually to the motherboard in order to fool it. I don't think I can straight up send five volts to the PWR_GOOD rail because the motherboard expects some delay, so... how? Maybe with a Time Delay Relay, or something like that?

I am not actually going to try something like this, but I was just curious if this can be done, and what one would use to achieve a fake PWR_GOOD signal.

 

[Joseph57]

bad idea
The PG signal is for protection for the board, if you by-pass you'll risk damaging your board and possibly other components. It might end up costing you more to attempt this than it would to get a good power supply designed for computers. I know the PSU prices have climbed within the last year, but there's not much anybody can do about that right now.

 

[Myronazz]

Yes, well, you are right to say it isn't a good idea. But I was just wondering how (and if) it can be done. I was never going to get a lab power supply and try this, I was just curious.

 

[Joseph57]

The PG signal and other protections are in an Integrated Circuit chip. If you want to make changes you will have to know about IC chips and electronic circuitry.
IC Circuitry Link example

 

[Myronazz]

Right, that was the plan. I said I wouldn't do this... but now I kind of want to. I guess I could get an old no-brand PSU I have and try it with a motherboard that's on its deathbed anyway. Just an experiment, I haven't really seen anyone else try it (probably because there's no need, lol.)

From the datasheet you gave me: "the PC power supply makes the power good signal high after some delay." Which matches what I originally said, the motherboard expects a high signal (at 5 volts) in the PWR_GOOD pin of the ATX header after some delay. So, just straight up throwing five volts there wouldn't work, because there would be no delay.

"T3(PGO Delay) : 100ms < T3 <500ms " Also from the specifications, the IC generates the PGO signal at a high state with this delay, if I understand correctly.

Theoretically, if I can input a 5V signal no sooner than 100ms, but no later than 500ms, then... this should fool the motherboard... unless I'm missing something.

To do this, I believe something like this would work. It's a Time Delay Relay, with a possible range of 0.1 seconds to 999 minutes. 0.1 seconds is 100ms, so it's capable of going higher to something like 200ms if the exact value doesn't please the motherboard.

This is really stupid, but now I want to do it because it is. I'll my old P.O.S. PSU and cut the Power Good wire, and pass through my own 5V signal with the delayed relay to see if I can fool the motherboard.