Well I'm going to have to fess up to being an idiot on this one.  After
several of your suggestions (and my own low confidence in the determination
of bad RAM as that has never happened to be before) I did some more
testing.  Turns out the culprit was overheating in the system.

When I was doing my swapping of RAM between banks, it turns out I had left
the side panel off the case.  Which means that it was running much cooler
than it was with both sticks of RAM in.  Putting both sticks back in and
still leaving the side panel off confirms that it wasn't the RAM causing the
crashes but overheating.

Here are my temps at idle according to Speed Fan:

Mobo1 - 102F
Mobo 2 - 194F
Mobo 3 - 106F
HD0 -  86F (as reported by the RAID controller)
Core - 48F

At those temps the machine is rock solid stable.  But that 194F temp scares
me.  I'm trying to figure out what it is and all I can think is that it's
for the integrated video chip.  The integrated video is covered by a gold
heatsink and is very, very hot to the touch but didn't come with a fan.

I did some googling and it seems that I have rev 1.0 of the mobo and in 1.1
they added a better heatsink.  I followed the lead of a few other rev 1.0
owners and pulled off the stock heatsink, scraped off the crappy thermal
glue stuff and applied some good themal paste.  I then put the heatsink back
on.  Didn't seem to make much of a difference.  I think the biggest problem
is that there are only two small plastic clips holding it on and it's fairly
loose.

Thoughts?  I've got two 120mm Scythe fans in the case, one intake and one
exhaust.  I have another 80mm fan blowing across the HDD cage with the 3
drives for the RAID.  I'm using one of the monster Scythe CPU heatsinks that
supposedly doesn't need a fan but I do have another 120mm fan I can use on
that as well.

------
Brian Weeden
Technical Consultant
Secure World Foundation

Reply via email to