P4C800 died !

  • Thread starter Thread starter marecki
  • Start date Start date
M

marecki

Hi!

I got P4C800-Deluxe with P4 3.2E , 1 GB of RAM.
Everything worked excellent.
One day I plugged in USB Mustek 1200CS scanner to front USB port (working
all the time with other stuff)
Then everything got frozen so I shut PC off and tried to turn on again. Now
the mainboard is dead , went RMA.
POST reporter was saying:
"System failed , CPU test"

Rest of hardware is OK , also scanner .

What and why it happened ?

marecki
 
Hi!

I got P4C800-Deluxe with P4 3.2E , 1 GB of RAM.
Everything worked excellent.
One day I plugged in USB Mustek 1200CS scanner to front USB port (working
all the time with other stuff)
Then everything got frozen so I shut PC off and tried to turn on again. Now
the mainboard is dead , went RMA.
POST reporter was saying:
"System failed , CPU test"

Rest of hardware is OK , also scanner .

What and why it happened ?

marecki
Click to expand...

I think I know what happened. I bet your Southbridge chip
got very hot. This is another Southbridge failure, via
USB. USB seems to be a common root cause for this failure.

The mechanism is currently unknown, but it could be
some static electricity entered the USB port when you
were plugging in your Mustek scanner. Or, it could be
a phenomenon called "latchup", where a PNPN phantom
junction forms when too much current is jammed into
a signal lead. Once the PNPN junction starts conducting
current, the only way to make it stop, is power off the PC
via the switch on the back (or pull the plug). Even pressing
reset cannot stop it.

http://groups.google.ca/[email protected]

Return the board under warranty and get another one.

I cannot guarantee that this will not happen again. If you
are nervous about this, buy a USB2 PCI card and connect
all your USB devices to it. Don't use the motherboard USB
ports, if you think this will happen again. The odds should
be very low for something like this to happen, but without
knowing the exact failure mechanism, it is hard to know
for sure.

Also, have a look at the connector on the end of the USB
cable you were using, to see if any pins/wires are bent
or shorted. If the +5V signal gets connected to a D+
or D- signal, that might be enough to cause destruction
via latchup.

This article provides a partial explanation about the latchup
effect. My problem is, I cannot see any good mechanism for
the failure, unless the +5V signal on the USB port is
involved somehow. And a paper I just glanced at, mentions
that static discharge can indeed cause latchup as well.

http://www.analog.com/library/analogDialogue/archives/35-05/latchup/

Since other chipsets don't do this, I would have to suggest
that this is an Intel problem.

Paul
 
I think I know what happened. I bet your Southbridge chip
got very hot. This is another Southbridge failure, via
USB. USB seems to be a common root cause for this failure.

The mechanism is currently unknown, but it could be
some static electricity entered the USB port when you
were plugging in your Mustek scanner. Or, it could be
a phenomenon called "latchup", where a PNPN phantom
junction forms when too much current is jammed into
a signal lead.
Click to expand...

Thank You Paul , I wish everyone could answer like this ...

Hmm so maybe putting the heatsink to southbridge can decrease opportunity of
this failure ?

marecki
 
Thank You Paul , I wish everyone could answer like this ...

Hmm so maybe putting the heatsink to southbridge can decrease opportunity of
this failure ?

marecki
Click to expand...

Once latchup starts, no heatsink on earth can stop it.
Only if you managed to switch off the computer in
around 1 second or so, could you save the Southbridge.

Latchup can also occur at the gate level on a chip.
I have had a chip in the lab, that would simply cease to
function (no outputs would wiggle). If I turned off the
power, and turned the power back on, the chip would still
refuse to work. When I left the chip with the power off
overnight, and turned on the circuit the next morning, the
chip was working 100%. This phenomenon is called "micro-latchup",
and doesn't cause severe heating like has happened to your
Southbridge.

When people first reported this problem, I couldn't believe
it. That is because modern chips usually have a pretty good
rating against this happening. It is a surprise for an
Intel chip to be doing this. I don't know if any motherboard
manufacturer will admit publically to a problem like this or
not.

Paul
 
Paul said:
Once latchup starts, no heatsink on earth can stop it.
Only if you managed to switch off the computer in
around 1 second or so, could you save the Southbridge.

Latchup can also occur at the gate level on a chip.
I have had a chip in the lab, that would simply cease to
function (no outputs would wiggle). If I turned off the
power, and turned the power back on, the chip would still
refuse to work. When I left the chip with the power off
overnight, and turned on the circuit the next morning, the
chip was working 100%. This phenomenon is called "micro-latchup",
and doesn't cause severe heating like has happened to your
Southbridge.

When people first reported this problem, I couldn't believe
it. That is because modern chips usually have a pretty good
rating against this happening. It is a surprise for an
Intel chip to be doing this. I don't know if any motherboard
manufacturer will admit publically to a problem like this or
not.

Paul
Click to expand...

Gigabyte has an entry in their FAQs for motherboards regarding an ICH4/ICH5
Southbridge vulnerability when using front USB ports.

Titled:

Q. The ICH4/5 SouthBridge chipset on my Intel® motherboard appears damaged
or "burnt". How did this happen?

http://www.gigabyte.com.tw/Motherboard/Support/FAQ/FAQ_456.htm

Cheers,

John S.
 
Back
Top