kony wrote:


> Did I just not look long enough to find data on this board,
> or is it either a different MSI model, or an ECS KN1? A
> link to a good picture of it might help us assess the
> heatsink.

Sorry, it is indeed a ECS board, the video card is MSI.

>
> What are you using to determine this temperature reading?
> Is it possible the reading is incorrect? When you feel the
> heatsink, is it very hot? I mean hot enough you cannot
> leave your fingers on it for more than a second?

This computer is running Gentoo linux, using the lm_sensors to monitor the
fan speed and temperatures.

>
> Basically, if the heatsink is making good contact with the
> northbridge, and if it is 82C, it will feel very hot. If
> the heatsink is not making good contact with the
> northbridge, and that lack of good contact is the cause of
> it's temperature, it will not feel so hot.

I will do this, thanks.

>
> Take the heatsink off again and examine the contact area, to
> see if there is the clear impression of the northbridge on
> it as a gauge of whether it was making contact. Examine
> the mounting clips and how it is situated when you remount
> it. The point here is that the stock heatsink, mounted
> properly, with a working fan, is sufficient to keep it cool
> enough no matter what load the system is under unless there
> is a very high ambient (room) temperature, or the case has
> very poor intake/exhaust airflow.
>
> Key is determining which factor is the cause, IF the
> temperature reading is correct. You have not written of any
> system instability resulting so it further puts into
> question whether the temp reading is correct. Is this temp
> also within the same range when the system is sitting idle
> in the bios health/hardware-monitor page, (if that page
> does display northbridge temp)?

> What has changed since then? Have you done any work inside
> the system, or moved it and potentially the mounting studs
> popped out as a result, just prior to onset of the problem?
> Had the (summertime where you are?) ambient room temps
> increased significantly, had you installed new software? If
> not, is it possible there is a bios setting for northbridge
> voltage that was unintentionally increased?

I don't think the bios has been changed, but I will check to be sure.

> If none of these, I would wonder if there might be a
> virus/etc that is causing more of a system load than usual
> to account for an temp increase, BUT even at full load the
> virus, if present, would only be revealing the problem not
> the cause since any system should be able to run at full
> load indefinitely.

I am running Gentoo linux behind an Ipcop firewall, and I don't use p2p
software or other "risky" software. Probably not a virus. The system does
not exhibit any signs of being under a heavy, or even moderate load.


>
> The increase in temp from the northbridge alone is not
> enough to feel an increase in temp from the case exhaust.
> While it isn't exactly a cool running chipset, even it's
> peak power->heat generation is not much relative to the
> total system power consumption. There must be other
> components also generating more heat than usual in order for
> the difference to be enough to perceive. Check your
> (windows?) Task Manager for processor loading % and which
> process(es) are the cause. This, as mentioned above, should
> be rectified if it is not something you intended to have
> running, but ultimately no matter what is running it should
> not cause such a high temp increase IF the temp is
> accurately reported.

The next time it spikes, I will open top to see what is running.

> Is the rear case fan mounted on a grill made of small
> stamped out case holes or slots? A mostly obstructed fan
> will not move enough air for a typical system as you've
> begun to describe, especially if there is also inadequate
> front and/or side panel intake area.... and yet, as you
> wrote it seems to have worked ok for 3 years but is it
> possible this is the temp it had reached in past summers if
> the ambient temp had gone up?

The computer is dusted out on a regular basis, once or twice a year.
>

> The northbridge is not likely to be the first part to fail
> with the system's other components also running at the
> elevated temp they must be at for there to be a perceivable
> temp difference from case exhaust. It will definitely
> reduce system lifespan (if some other unrelated failure
> doesn't cause a problem first, we would have to have a
> crystal ball to predict this or how much shorter the
> lifespan will be from temperature, though some say a 10C
> rise in temp of certain components will cut their lifespan
> in half, but this is usually other components like
> capacitors and fans, while silicon will typically withstand
> 80C for longer than the viable life of the system unless the
> excessive heat is more than the PCB solder junction can
> withstand, but of course if a certain threshold is exceeded
> it will cause instability and potential data corruption.
>
> Since your case exhaust is feeling hotter, regardless of the
> cause it is an indication that the system at peak load is
> potentially not cooled well enough by the case cooling
> subsystem, so whether it is the final resolution or not, one
> thing that should be done is to improve the case cooling
> through whatever measures are needed. Since we don't' know
> what other components are in the system, what the ambient
> temp is, or the exact state of the airflow in an unknown
> system chassis, no specific recommendations can be made
> there. Generally the idea is that the rear case fan and
> rear PSU fan should both have no stamped-out metal covering
> them, at most a tubular (usually chromed-type) fan guard on
> them, and the front should have a fairly large intake area
> totaling at least as much as a 80mm fan cutout (which could
> be many small holes cumulatively, it need not be one giant
> 80mm hole) that is situated roughly such that the intake
> airflow goes past the hard drive to cool it as well.
The case is an Ultra, I don't remember the model. It would appear to have
adequate cooling, a large 80mm grill in front of the cpu and many small
holes along the lower left cover.
>
> I will assume the case doesn't have excessive dust buildup,
> though if you haven't taken the front bezel off the case
> ever, there could be a fair amount of dust trapped in there
> even if you've cleaned out the rest of the case. I wouldn't
> think it excessive enough to cause this by itself though in
> an effort to leave no stone unturned I mention it anyway.
>
> You might also try temporarily leaving the case side panel
> off and pointing a desk fan at the system interior. If
> doing so, note what difference in temp it makes. If that
> difference is more than about 5-10C, it is a sign you need
> to modify the case to improve airflow before putting further
> consideration into the resultant temp of the northbridge. If
> all else fails and you don't want to strip the whole system
> down to cut out metal (stripped down so the metal fragments
> don't end up loose inside, potentially shorting out
> something), then consider cutting a hole in the side panel,
> towards the lower front area roughly across from the hard
> drive bay to help, instead of rob airflow from that area,
> unless you have a very hot running video card too in which
> case it might benefit even more from the addition of a side
> panel fan blowing on it, or ideally even two fans blowing on
> each area.
>
> Since the system has been running hot, examine the
> motherboard, and video card if present, for failing
> capacitors. Generally they'll have a leaky residue or domed
> tops, sometimes the bottom plug is out instead. Visually
> they'll be obvious enough. If any are present, the life
> of the component is gone unless you want to bear the repair
> costs to fix, replace capacitors.
Thank you for the very detailed reply, you have given me several areas to
look into, which I am in the process of doing now.
Terry