MSI MS-7448 (DELL vostro 430) - fast power on/off, PS_ON circuitry shenanigans

[kotel studios]

Hi,

Recently I had my MSI MS 7448 board from the dell vostro 430 screw up on its own. All I did was leave this PC unattended while it was installing windows and I come back to the shop and now it's powering on/off. It's on for around 1-2 seconds, then turns off for a split second and again and again... I swear the (D)HELL's just like to screw themselves up for no reason than to say "fuck you". This is my 2nd HELL pc I had die on me. It worked fine before, no weird issues at all.

I have ruled out pretty much everything (PSU, dead CMOS battery, CPU etc....). I also tried pulling the PS_ON (which is around 0.78v) signal high by connecting it through an 1k ohm resistor and an 780ohm one (which made it go up to around 3v) , but that only made the board sometimes go to GPU error and it got stuck there sometimes beeping due to the quite fast power on/off behavior. Apart from the inability to shut the PC down and it not wanting to POST 87% of the time by pressing the power button it seems to be okay.

This is my second case where the board goes crazy like that, albeit the first one had boardviews available while this one has nothing.
Any ideas?

[FriedFred]

Is this Error sensitive to temperature (hot air test), or gentle pressing, or gentle benting the board?

[kotel studios]

Is this Error sensitive to temperature (hot air test), or gentle pressing, or gentle benting the board?

Sorry for the late reply.

No, I don't think it is. I pressed hard on the SIO and in the arena near the PS_ON switch but nothing changed. I also tried to gently bend the board but to no avail. It's not possible that somehow an bad transistor does this. It would either be completely on or off if it was broken.

it couldn't be damaged by an scrapper since I got the PC with everything and it died on its own while I was away.

Any ideas?

[DynaxSC]

Such bahavior suggests issue with completing power sequence, some supply voltage in the sequence is not getting up stopping the process. Possible some vrm chip is dead/shorted, check resistances on all coils and look for shorted ones. CPU should be taked out for this.

Another method is looking for the vrm's if they get ENABLE signal, but do not generate any voltage, however you would need a 2-Channel scope for this, as the time the board is powered is very short.

If no BV available you need to look into the vrm data sheets for ENABLE pin.