New CPU Mobo combo only works sometimes.

[Andrew Tubbiolo]

Hey All:

I'm trying to get a new mobo-CPU combo working. It's a AMD Phenom
8400 (3core) sitting on a ECS A770M-A motherboard with 2 gig of ram.

When the system starts up I can run and change BIOS, run a live
linux distro off USB, and it'll run for hours. Then If I cycle power
with the case pushbutton, it's as if the CPU stops working. The fans
come on, but no POST no anything. Of course the push button still
works to power down. However, if I use the power switch on the power
supply power reboots work fine.

What is having the problem here? The CPU, or the Motherboard? When
I purchased the system, the CPU has a huge glop of thermal paste
between the heat sink and CPU. It was about 2 mm thick. After fixing
that, I have this problem. Could there be thermal damage to the CPU?
Maybe even the Mobo?

Anyone have similar problems where the cause was finally
understood. Any help would be appreciated.

Me

 

[Paul]

Hey All:

I'm trying to get a new mobo-CPU combo working. It's a AMD Phenom
8400 (3core) sitting on a ECS A770M-A motherboard with 2 gig of ram.

When the system starts up I can run and change BIOS, run a live
linux distro off USB, and it'll run for hours. Then If I cycle power
with the case pushbutton, it's as if the CPU stops working. The fans
come on, but no POST no anything. Of course the push button still
works to power down. However, if I use the power switch on the power
supply power reboots work fine.

What is having the problem here? The CPU, or the Motherboard? When
I purchased the system, the CPU has a huge glop of thermal paste
between the heat sink and CPU. It was about 2 mm thick. After fixing
that, I have this problem. Could there be thermal damage to the CPU?
Maybe even the Mobo?

Anyone have similar problems where the cause was finally
understood. Any help would be appreciated.

Me

It could be the Vcore regulator on the motherboard, going
into "latch-off" to protect itself. Or any number of
motherboard problems. I would suspect the motherboard.
I hope you didn't get thermal paste in the socket.

To check for latch-off, would require detailed knowledge
about the Vcore regulator chip, and whether it has any
status outputs or not. So it would not be a trivial
matter to check. (You could also try measuring Vcore
with a multimeter, when the computer won't start.)

It could even bw a problem with the power_good logic,
preventing the system from coming out of reset. Again,
that would be a motherboard issue.

There will not be thermal damage to the CPU, because the
CPU has THERMTRIP. That turns off the power on the
computer, in the event that the CPU gets too hot.
So the CPU can protect itself.

Paul

 

[Rarius]

Hey All:

I'm trying to get a new mobo-CPU combo working. It's a AMD Phenom
8400 (3core) sitting on a ECS A770M-A motherboard with 2 gig of ram.

When the system starts up I can run and change BIOS, run a live
linux distro off USB, and it'll run for hours. Then If I cycle power
with the case pushbutton, it's as if the CPU stops working. The fans
come on, but no POST no anything. Of course the push button still
works to power down. However, if I use the power switch on the power
supply power reboots work fine.

What is having the problem here? The CPU, or the Motherboard?

I had an almost identical problem with an Intel PentiumD820 on a new
aBit IP35Pro. The problem was the motherboard. I RMAed it and the
replacement worked perfectly.

If you power down the system using the power switch on the PSU, ie you
TOTALLY power down the system, I'll bet that it reboots fine.

I suspect that part of the mobo isn't being reset when with the soft
powerdown. that is why the hard powerdown works.

When
I purchased the system, the CPU has a huge glop of thermal paste
between the heat sink and CPU. It was about 2 mm thick. After fixing
that, I have this problem. Could there be thermal damage to the CPU?
Maybe even the Mobo?

It is almost impossible to thermally damage modern processors. They wil
throttle down then shut down before any damage is done.