Path: utzoo!attcan!utgpu!jarvis.csri.toronto.edu!mailrus!ncar!ico!vail!rcd
From: rcd@ico.isc.com (Dick Dunn)
Newsgroups: comp.sys.ibm.pc
Subject: Re: BEWARE OF SAI SYSTEMS IN SHELTON, CT
Summary: an analogy for running out-of-spec
Message-ID: <1989Nov16.002229.5457@ico.isc.com>
Date: 16 Nov 89 00:22:29 GMT
References: <6268@tank.uchicago.edu>
Distribution: usa
Organization: Interactive Systems Corporation
Lines: 43

phd_jacquier@gsbacd.uchicago.edu writes:
...a discussion with a vendor about using a 10-MHz chip in a 12-MHz
system:
> I called and they stand by it even though they don't tell you anything at 
> purchase time. Here are some excerpts of the phone call:
> 
> Q: " It is not recommended to push CPU's beyond their guaranteed speed.
>      You should give me a 80286-12 if i paid for a 12Mhz system"
> A: " Have you had problems? You have a one year guarantee anyway"
> Q: " What if it dies after one year because it's been used at too fast 
>      a speed?"
> A: " If there is a problem because of the speed, it's going to happen right
>      away, not after a year."

Let's try a contrived example:

Q:  "The tires on my car are not rated for highway speeds of 65."
A:  "Have you had any flats?  The tires are warranted anyway."  (I told you
	it would be a contrived example.:-)
Q:  "What if the tires give out because they're being used above their
     rated speed?"
A:  "If there is a problem because of the speed, it's going to happen right
     away."

Of course, running a PC with chips outside spec is not (likely) a physically
dangerous thing to do...but it certainly can affect the reliability of your
system.  Moreover, if there is a problem because of the speed, it will show
up in one of two ways:
	- The system will fail right out of the box.
	- It will work for a fair while (perhaps a year?) but eventually
	  fail as components age.  There is a popular misconception that
	  "no moving parts" ==> "won't wear out"
There really IS something to worry about with running CPUs outside spec.
Also, as I've commented before, if the board designers don't even stay
within spec on the CPU--which is stamped with the right speed so any fool
can see it--what are the chances they've cheated in other parts of the
design?  How much of your data do you want to bet that they got it right?

The hope was expressed that this is not a common practice; I think that
pushing CPU chips IS unfortunately common.
-- 
Dick Dunn     rcd@ico.isc.com    uucp: {ncar,nbires}!ico!rcd     (303)449-2870
   ...Keep your day job 'til your night job pays.