L300D PS10AP - no post, power cycling

[spleenharvester]

Got this board today. Got the usual problem with these AMD based L300 boards - no POST. But no NEC Tokin.

Most of the time, it will power on, wait 5 seconds, power off, repeat that three times and then stay off. Occasionally, it will stay on but still with no POST.

I've checked PAD504 and found there was no NB VCC voltage. So I checked the adjacent TPS51117 powering it, and found there is in fact no CPU_PWRGD signal triggering the 51117. And since CPU_PWRGD is linked to the THERMTRIP signal, that's probably what's causing the power cycling.

I've checked, and CPU Vcore is there at 1.1v. Should I replace the tantalum caps incase they've gone high ESR?

Loads of people have this problem so I'd like to get to the bottom of it.

Schematic:

Thankyou very much!

[spleenharvester]

Re: L300D PS10AP - no post, power cycling

On further investigation, CPU_PWRGD IS present (I really need a new DMM). The problem appears to be a short on the VCC_NB rail.

[dycc]

Re: L300D PS10AP - no post, power cycling

first check Q512 if you get V1.8s .then check u18 for PWR_GOOD_3 presence . because PWR_GOOD_3 enable U512 which enable CPU_PWRGD.

i have already repair this type of faulty board and it was Q512 the problem in my case

[dycc]

Re: L300D PS10AP - no post, power cycling

sorry thought CPU_PWRGD was not present .

[spleenharvester]

Re: L300D PS10AP - no post, power cycling

A bit of progress - NB VCC only reads as shorted and 0v when it decides to stay on. When NB voltage is there it power cycles. Maybe something NB-related?

Otherwise, all voltages appear to be present. NB VCC PWRGD is present, CPU PWRGOOD is present, PWR_GOOD_3 is present. The shutdown is not a thermal shutdown, I checked Q550. No idea what this one might be.

[spleenharvester]

Re: L300D PS10AP - no post, power cycling

Actually, just checked Q512 - reads as conducting in both directions from source to drain, source to ground reads short, drain to ground reads infinity.

But I've found three other MOSFETs on the board that also read shorted from source to ground. Is this sometimes normal? Or do I actually have a ton of failed MOSFETs?

[spleenharvester]

Re: L300D PS10AP - no post, power cycling

New discovery, it doesn't power off with no CPU present.

[spleenharvester]

Re: L300D PS10AP - no post, power cycling

Oh crud, fixed it! See https://www.badcaps.net/forum/showpo...&postcount=567