From: jiml@garfunkel.FtCollinsCO.NCR.COM (Jim L)
Subject: Re: SIMM Speed
Distribution: world
Organization: NCR Microelectronics Products Division (an AT&T Company)
Lines: 40

In article <1993Apr6.150808.27533@news.unomaha.edu>, hkok@cse (Kok Hon Yin) writes:
|> Robert Desonia (robert.desonia@hal9k.ann-arbor.mi.us) wrote:
|> : B
|> : BK>Is it possible to plug in 70ns or 60ns SIMMs into a motherboard saying
|> : BK>wants 80ns simms? 
|> 
|> : You shouldn't have troubles.  I have heard of machines having problems 
|> : with slower than recommended memory speeds, but never faster.  
|> 
|> --
|> It should run without any trouble of course but why do you want to buy some
|> 60ns and mixed them with 80ns?  60ns is more expensive than 80ns and
|> furthermore your machine will run the slowest SIMMs clock speed eventhough
|> you have 60ns.  Just my 0.02cents thought....
|> 


Your machine will run at whatever the bus is jumpered to/CMOS is set to
(usually wait states) regardless of what speed RAM is installed.  No
motherboard can sense the speed of the RAM installed, unless you call
failing as a sort of auto-sense.  This is how you can sometimes use
"slower" RAM in a machine.  You either set the number of wait states to
accomodate the slow RAM (in which case, all memory will run at that
slower rate) or you reduce the wait states and take the chance that the
slower RAM will act like faster RAM and you won't crash.

Putting faster RAM in won't speed things up unless you tell the machine
it has faster RAM.  

Mixing fast and slow RAM will not help you if you have to keep the bus 
slowed down to accomodate slow RAM.

JimL
--------------------------------------------------------------------

-- 
Mailer address is buggy!  Reply to: jiml@strauss.FtCollinsCO.NCR.com

James Lewczyk                                   1-303-223-5100 x9267
NCR-MPD Fort Collins, CO             jim.lewczyk@FtCollinsCO.NCR.COM