Path: utzoo!utgpu!news-server.csri.toronto.edu!bonnie.concordia.ca!thunder.mcrcim.mcgill.edu!snorkelwacker.mit.edu!apple!usc!samsung!caen!uflorida!reef.cis.ufl.edu!jdb
From: jdb@reef.cis.ufl.edu (Brian K. W. Hook)
Newsgroups: comp.sys.ibm.pc.hardware
Subject: Re: Buying a 386: 33 MHz problems, C&T chipset problems?
Message-ID: <26368@uflorida.cis.ufl.EDU>
Date: 17 Jan 91 23:12:42 GMT
References: <1991Jan17.081458.3191@cs.ucla.edu>
Sender: news@uflorida.cis.ufl.EDU
Organization: UF CIS Dept.
Lines: 57

In article <1991Jan17.081458.3191@cs.ucla.edu> wales@valeria.cs.ucla.edu (Rich Wales) writes:
>I'm seriously considering buying a 386 motherboard in order to upgrade
>my 8-MHz 286 system at home.
>
>One dealer gave me the following recommendations, which I'd appreciate
>any feedback on:
>
>(1) He advised me not to get a 33-MHz 386, but to get a 25-MHz instead.
>    He claims there are engineering problems with motherboards running
>    at such a high clock speed as 33 MHz.

That's funny.  I haven't a single problem.  Maybe those problems explain
why 33mhz machines are considered the ultimate in price and performance!
I really doubt that "engineering" problems play a big factor in 386
design at high speeds.  Now I do know that 286s at high speed will be
real flaky.
>
>(2) He advised me to avoid motherboards using the Chips & Technologies
>    chip set, as these supposedly have problems at high speeds (even
>    25 MHz).  He claims he's had personal experience with flaky 386'es
>    with C&T chips.

I find that HIGHLY difficult to believe since the C&T is the industry 
standard chipset for AT compatibles.  I mean, it's only competition is
VLSI or INTEL or OPTI, neither of which are really posing a big
problem.  Calling the C&T chipset flaky is akin to calling the Western
Digital ST506 a flaky interface or Intel 80386 a flaky chip.  Their
industry standard for a reason, folks.

>(3) He recommended that I should look for a 386 motherboard with support
>    for a 287 (that's right, 287) coprocessor -- since 287's are =much=
>    cheaper than 387's and can still give an acceptable increase in math
>    speeds for all but the heaviest number-crunching applications.

True, a 386 witha 287 will be faster than one without a 287, but that is
kind of useless, don't you think?  You pay for speed, and then you 
cripple yourself with a slow coprocessor?  I don't know, kinda like
buying a Ferrari and putting on a gas cut off valve at 55MPH.
>
>Again, I am not making these claims myself; I am passing them on from a
>dealer -- albeit one who runs his own company (i.e., not a salesdroid),
>seems to know what he's doing, and with whom I've done business before
>(I bought a RAM card from him for my laser printer last year).
>
>I'd welcome any comments.
>

I'd like to hazard a guess that this guy you have been talking to has
for sale:

	A 386 motherboard that has a slot for an 80287 and that it does
	not use a C&T chipset.  Oh, I bet it isn't a 33mhz, is it?

	:)

Brian