P4C800-E Northbridge - Too Hot to Touch (Random Reboots)

G

Greg Wilder

I have a homebuilt P4C800-E PC, with a 3.2GHZ P4. Not doing any
overclocking.

Worked fine for about 4-5 months, then started getting random reboots
and BSOD. Have unistalled, upgraded, new ram, etc.

My CPU temp is running around 49c. However, my Northbridge heatsink is
too hot to touch. Do you think I have a problem with the MB, or should
I look at an active cooler for the Northbridge.

THanks.
 
P

Paul

Greg Wilder said:
I have a homebuilt P4C800-E PC, with a 3.2GHZ P4. Not doing any
overclocking.

Worked fine for about 4-5 months, then started getting random reboots
and BSOD. Have unistalled, upgraded, new ram, etc.

My CPU temp is running around 49c. However, my Northbridge heatsink is
too hot to touch. Do you think I have a problem with the MB, or should
I look at an active cooler for the Northbridge.

THanks.

If you look up the power dissipation specs for the Northbridge, the
family runs in the 12W range. Temperature can depend on how aggressive
the memory components are, that you are using (i.e. DDR500 will make
it warmer). When you bump up the Vdimm for your memory, that also
bumps the 2.5-2.6V that feeds the Northbridge, which makes it hotter.

I suspect, however, that some supply tied to that chip is way out of
wack. It may not be a voltage which registers in the hardware monitor
(a shortcoming I don't appreciate - all supplies should show up in the
monitor, as all it would take is an analog mux to add them). Take a
look in the hardware monitor anyway, to see if anything there looks
out of line.

I would RMA the board, before it destroys some of your other hardware,
like your DDR memory for example.

My rule of "thumb" is, if you cannot hold a finger on a heatsink
for more than 2 seconds, it is running at 50-55C. By your description,
yours is a lot hotter than that.

Here are the absolute max voltage values for the 875P Northbridge:
http://developer.intel.com/design/chipsets/datashts/25252501.pdf

Absolute Maximum Ratings
Symbol Parameter Min Max Unit
VCC 1.5 V Core Supply -0.3 to 1.75 V
VCC_AGP 1.5 V AGP Supply -0.3 to 1.75 V
VCCA_AGP 1.5 V Analog AGP Supply -0.3 to 1.75 V
VCC_HI 1.5 V HI/CSA Supply -0.3 to 1.75 V
VTT VTT Supply -0.3 to 1.75 V
VCC_DDR 2.6 V DDR System Memory Interface Supply -0.5 to 3 V
VCCA_DDR 1.5 V Analog Supply for System Memory PLLs -0.3 to 1.75 V
VCC_33 3.3 V Supply -0.3 to 3.6 V
VCCA_FSB 1.5 V Host PLL Analog Supply -0.3 to 1.75 V

Based on that list, I would expect to see a 1.5V supply perhaps tied
to the same supply as the AGP I/O, a 2.6V Vdimm supply shared by
Northbridge and DIMMs, and a 3.3V supply (could come straight from the
PSU). VTT is also a separate supply. So, if any of those is
radically wrong, or if the clock frequency being fed to the
Northbridge is wrong, that will make the chip hot.

It doesn't take a lot of air movement through the Northbridge
heatsink, to cool it down. I worked on a 865G board last year
that had a heatsink running at over 75C. That is because the
heatsink got zero air flow. (It was in a weird low profile
enclosure.) Adding a 40mm fan to the top of the heatsink
brought it down to 37C, as measured by a portable
digital thermometer. Again, if you find that blowing a little
air over it, is not bringing it into the "touchable"
temperature zone, there must be something seriously wrong.
Amazingly, there wasn't a hint of instability on that board,
even though the silicon die was probably running at 100C.
That means your Northbridge is probably hotter than that.

To measure any of the voltages above, you would probably have
to flip the board up on edge, and access connections on the
back of the board. This is not too practical for hobbyist
purposes - you would need a clamp or something, to hold it
vertical while you work on it.

Another test you might want to try, is feel all the MOSFETs.
These are the devices with the three leads coming out of them,
of which there are a large number on the board. I expect you
are going to find some of those are hot as well. Knowing
which one or one(s) are hot, can help narrow down which supply
voltage has gone nuts. (I have a P4C800-E, and no, I haven't
figured out what all the MOSFETs do...) A poster yesterday,
reported hot MOSFETs on another member of that family of boards,
so you might see something similar. MOSFETs soldered to the board,
are only good for one or two watts of power dissipation, before
they get good and hot. That is one reason I'm not too happy to
see Asus using them as part of linear regulators - they are fine
until something on the board starts drawing a bit too much current.

HTH,
Paul
 
G

Greg Wilder

Yes, it is too hot to touch via my finger for more than a couple of
seconds. Took the case off to see if that helps. Also, used a can of
compressed air upside down to cool it temporarily. Has been up 24hrs
now, but sometimes it will go 2-3 days, then crash 3-4 times in one
day.

I did have the RAM voltage up to about 2.65 - 2.75 to see if that was
the instability issue, so that may have caused a temperature problem.

I will probably swap out the MB just to make sure.

By the way, this is the error I am getting. Windows online support
indicates a driver issue. However, I have updated everything, and
removed the latest SW I installed right before the problem started.

Error code 1000008e, parameter1 c0000005, parameter2 bf851b3c,
parameter3 b7b289e0, parameter4 00000000.


I am using Mushkin memory, so I may swap that too.
Thanks to all
Greg
 

Ask a Question

Want to reply to this thread or ask your own question?

You'll need to choose a username for the site, which only take a couple of moments. After that, you can post your question and our members will help you out.

Ask a Question

Top