[comp.sys.ibm.pc.hardware] Info req: Mathco

mea1@uafhcx.uucp (Mark _E_ Amos) (07/24/90)

In article <292200022@trsvax>, ken@trsvax.UUCP writes:
> 
> 
> I was reading the July 16, 1990 issue of PC Week and saw an ad that caught
> my eye.  Advanced Micro Devices is advertising an 80C287 math coprocessor
> for only $99.00.  I could not believe it.  Does anyone know anything about
> it (ie. speed) or even how reliable the company's products have been in the past.  The only thing that makes me skeptical is that the company adds, in very
> small print, that you are "limited to two lifetime replacements per
> person."  I know components do not fail very often but this does not
> exactly install confidence in them.
>

I tried to post this reply once but I don't think it took so here goes - hope
it's not a repeat...

I recently read an article in the EE Times about the AMD 80C287, and it went
into some detail about the fact that AMD has the genuine Intel microcode
through a contract they signed in 1976.  Intel is currently battling to prove
the 80287 code is not included, but in the meantime we have a $99 80C287
available that SHOULD be 100% compatible.  I have dealt with AMD in the past
and they seem very responsive to customers, so the purchase is not of shady
equipment from an obscure company, but a reputable chip maker.  Further, the
target market (says AMD) for this chip is the "rest of us" who cannot afford
Intel's scalping.  They intend to sell tons of these, so I would speculate that
much testing has been done to ensure it is correct.

I have had direct experience with AMD as an OEM, buying and using tons of
EPROMs and microcontrollers, so I can vouch for their validity and quality.
The only precaution I would take is buying from a reputable distributor, so as
return would be easier if there were something awry.  

Further, the speeds are supposed to range up to the fastest '286 AT bus...  I
have no definate on this so please ask the seller - or AMD.

BTW, this litigation between Intel and AMD would mean 386 microcode availablity
to AMD if they win... (486 is not included, tho).

mea1@uafhcx.uark.edu - University of Arkansas, Fayetteville
			Computer Science Engineering

			Disclaimer: What, me worry?

bartling@mozart.amd.com (Steve Bartling) (07/27/90)

In article <292200022@trsvax> ken@trsvax.UUCP writes:
>
>
>I was reading the July 16, 1990 issue of PC Week and saw an ad that caught
>my eye.  Advanced Micro Devices is advertising an 80C287 math coprocessor
>for only $99.00.  I could not believe it.  Does anyone know anything about
>it (ie. speed) or even how reliable the company's products have been in the past.  The only thing that makes me skeptical is that the company adds, in very
>small print, that you are "limited to two lifetime replacements per
>person."  I know components do not fail very often but this does not
>exactly install confidence in them.
>
>Any info would be greatly appreciated.
>
>
>KEN                          trsvax.UUCP!ken
>#include  <disclaimer.h>

The 80C287 is a 'real' 80287. It is completely hardware and software
compatible with the Intel 80287. As far as reliability goes, I can
answer that by example. Did you know that 60% of all 80286 
microprocessors that are sold were made by AMD ? You can rest assured
that it would be very unlikely that you will even need to replace
the chip even once. The "limited to two lifetime replacements" clause
is required by the legal folks. 

As far as operating speed is concerned, it is designed to run at
10MHz. Since the 286 based AT uses a divide by two clocking scheme
for the processor and a divide by three scheme for the coprocessor,
the 10MHz 80287 can be used in up to a 16MHz 80286 system. The
arithmetic works out as follows :

A 16MHz 80286 system requires a 32MHz crystal.

32MHz / 2 = 80286 processor clock = 16MHz

32MHz / 3 = 80287 CO-processor clock = 10.66 MHz 

Thus the 80C287 can be used in up to a 16MHz AT.

As far as price goes, a 10MHz 80287 currently is available in
single quantities by mail order for about $280. The $99 dollar
deal is designed to buy market share ( IMHO ). AMD probably
will not over this good a deal for very long. I would jump at
the offer soon.

If anyone has any further questions, please feel free to email
me at bartling@mozart.amd.com.

-Steve Bartling
bartling@mozart.amd.com

P.S. For those of you who don't know much about AMD, they are
     the 12th largest manufacturer of semiconductors in the
     world. And they DON'T make DRAMS ( DRAMS skew the statistics
     since the volumes are very high ). You don't get this large
     without doing something right.

Disclaimer : I hate myself, my wife hates me, my dogs hate me,
	     and AMD hates me. None of the aforementioned 
	     ever believe a word I say, so why should you ?
	     In other words, I don't speak for AMD and they
	     don't speak for me. I just happen to work at AMD.