From owner-freebsd-hackers  Tue Oct  3 03:30:06 1995
Return-Path: owner-hackers
Received: (from root@localhost)
          by freefall.freebsd.org (8.6.12/8.6.6) id DAA11005
          for hackers-outgoing; Tue, 3 Oct 1995 03:30:06 -0700
Received: from nanolon.gun.de (nanolon.gun.de [192.109.159.5])
          by freefall.freebsd.org (8.6.12/8.6.6) with ESMTP id DAA10988
          for <hackers@FreeBSD.ORG>; Tue, 3 Oct 1995 03:30:00 -0700
Received: (from uucp@localhost) by nanolon.gun.de (8.6.8.1/8.6.6) with UUCP id LAA20269; Tue, 3 Oct 1995 11:19:33 +0100
Received: (from andreas@localhost) by knobel.gun.de (8.6.12/8.6.12) id JAA19413; Tue, 3 Oct 1995 09:33:44 +0100
From: Andreas Klemm <andreas@knobel.gun.de>
Message-Id: <199510030833.JAA19413@knobel.gun.de>
Subject: Re: make world on FreeBSD-stable impossible. cc1: ... signal 11
To: joerg_wunsch@uriah.heep.sax.de
Date: Tue, 3 Oct 1995 09:33:44 +0100 (MET)
Cc: andreas@knobel.gun.de, terry@lambert.org, davidg@root.com, mark@grondar.za,
        hackers@FreeBSD.ORG
In-Reply-To: <199509291919.UAA04815@uriah.heep.sax.de> from "J Wunsch" at Sep 29, 95 08:19:08 pm
X-Mailer: ELM [version 2.4 PL24 ME7]
Content-Type: text
Content-Length: 1903      
Sender: owner-hackers@FreeBSD.ORG
Precedence: bulk

> As Andreas Klemm wrote:
> > 
> > Well not only a fan. A fan with a heat sensor. So I thought I'm
> > relatively save, not overheating the P90. And I think this is the only
> > thing that might happen, when overclocking a chip.
> 
> You're overclocking the chip?  Well, it's not only the temperature of
> the case that might cause you harm.  The more subtile problems do
> arise out of local overheating on the chip (remember that the thermal
> flow inside silicon is rather limited), which can finally cause a
> total damage of the chip due to shortcuts or interruptions in some
> conductor layer.  The most suspectible layers are the top-level two Al
> layers with the intervening insulator layer.  The Al expands about one
> order of magnitude more than silicon, causing "hillocks" which can
> finally break through the insulator layer.

Uh, thanks for the detailed description, what might happen.
Fortunately it turns out, that my P90 has survived the 120 MHz 
burnin.

I changed main memory and burst cache memory Sunday evenig and 
hadn't any problems since then. I made 4 make world's and am running
a real stable -stable now (I hope with fingers crossed :-).

> The second reason for heat-caused failures is called "electro
> migration".  Basically, hot spots of Al conductor pads start to get
> thinner, and consequently even hotter.  This is much comparable to the
> burndown of the tungsten wire in a lightbulb.  It's generally a long-
> term failure, but caused by overheating.
> 
> IMHO, overclocking the chip is not worth it.

Ok, ok !! I'm convinced now ;-)))

So I think I can say a very big

	"Thank you"

to all of you, who helped me in the last days !

-- 
$$ apsfilter - magic print filter 4lpd      @home : andreas@knobel.gun.de
$$ ftp://sunsite.unc.edu                    @work : andreas@sunny.wup.de
$$ /pub/Linux/system/Printing/aps-491.tgz   knobel: >>> powered by FreeBSD <<<