X60 overheating with 7.1

[Hiroharu Tamaru]

Hi,

At Fri, 16 Jan 2009 16:05:40 -0500,George Neville-Neil wrote:
> On Jan 16, 2009, at 15:28 , Josh Paetzel wrote:
> >
> > A few things I've never posted before to the mailing lists.
> >
> > 1)  My heat problems were somewhat intermittant.  Enough so that it
> > would fool me in to thinking software was fixing things.  I'd try some
> > change, and it would seem to work, or not, but every once in a while I
> > was able to run a mke -j4 buildworld without it hitting 100C and
> > shutting down.
> 
> This sounds quite familiar.
> 
> > 2) IBM has admitted they had assembly problems.  If you browse around
> > you'll find some people who think their T60/X60/R60's are loud or have
> > unusably high heat, and some people who think the are quiet and
> > perfectly fine.  When I took mine apart the thermal compound was very
> > excessively applied (think 2 year old applying jelly to a sandwich)  
> > and
> > after I cleaned it up and reassembled it I thought I had damaged it
> > somehow as the cpu fan wasn't running at all.  It turns out that  
> > with a
> > proper amount of thermal paste it can handle low cpu load with only
> > passive cooling.  My CPU temps went from 65-70C at idle with the fan
> > running full out to 42-45C at idle with the fan either off or barely
> > running.  At full load it went from unusable to 65C with the fan
> > eventually ramping up to full speed after an hour or so of intensive  
> > use.
> >
> > You may be hitting a software issue, but my issue was caused by  
> > hardware
> > and very intermittant, which lead me to believe it was software  
> > related
> > for months on end.
> >
> 
> I started down the path of opening the box but it's pretty complex and  
> even with the manual
> getting to the CPU seemed fraught with peril.  If there were a site  
> like ifixit
> that had a good set of "to get to the CPU" I might attempt it again.
> 
> There are some factors that lead me away from a hardware problem:
> 
> 1) This machine was previously owned by my other half, who ran XP and  
> Vista on it
> with no problems.
> 
> 2) It really does seem related to 7.1 RELEASE vs. 7.1 STABLE and  
> CURRENT.  I have re-installed
> RELEASE twice and built -j 4 on it with no problems.
> 
> That being said, I may still look at the CPU paste.

Yes, I advise you so too.

My colleague and I both have X60, purchased at around the
same time but from totally different suppliers.  Mine was
fine, but for my colleague's X60, it started to shut itself
down very frequently after upgradeing it to 7.x.  We also
suspected some software problem, and we actually couldn't
prove it otherwise, but re-greasing the CPU paste solved our
issue as well, and we were done with it.
IIRC, some one claimed that the current ULE scheduler makes
a better use of the CPU than previous ones, not to mention
windows. Well, probably yes, but probably not...;-)
And yes, even after this surgery, mine run a little cooler
than his, with a very similar load, watching them side by
side.

What was a little different from most reports (in English)
was that in our case, the paste on the heatsink didn't look
like it reached the CPU (or was it the north bridge..?), so
we rather added some silicone paste.  There are some of
these reported in Japanese if you google around, FWIW.

This may help you to get to the CPU:
http://www-307.ibm.com/pc/support/site.wss/MIGR-62866.html
which links to:
ftp://ftp.software.ibm.com/pc/pccbbs/mobiles_pdf/42x3550_04.pdf
See Page 92 (pdf page 102/242)

I do remember that there were some subtle places (though we
worked without any documents), and I do remember that there
were some screws that had more than one thing to hold, on
front and backside of the main board. So look carefully
before touching each screw!  YMMV, of course.

Hiroharu Tamaru