NetNews Usenet Archive 1992 #16/NN_1992_16.iso/spool/comp/sys/intel/1333

Text File | 1992-07-22 | 1.9 KB | 41 lines

Newsgroups: comp.sys.intel
Path: sparky!uunet!gatech!europa.asd.contel.com!darwin.sura.net!mips!mips!munnari.oz.au!labtam!labtam!scott
From: [e-mail] (Scott Colwell)
Subject: Re: 386/486 clock (1x or 2x)?
Organization: Labtam Australia Pty. Ltd., Melbourne, Australia
Date: Thu, 23 Jul 1992 00:24:50 GMT
Message-ID: <scott.711851090@labtam>
References: <[e-mail]> <[e-mail]> <[e-mail]>
Lines: 30

[e-mail] (Thomas Hoberg) writes:
>In article <[e-mail]> [e-mail] (Paul De Bra) writes:
>   The 386 uses a 2x clock, whereas the 486 uses a 1x clock.
>   A 2x clock has the advantage of delivering a perfectly symmetrical
>   1x clock after halving. A 1x clock needs better clock circuitry to
>   make sure it is symmetrical.

>Hmm, I was always under the impression that even the 486 used a 2x
>clock up until the 486DX[2]50, which uses a 1x clock. Anyway I have a
>66MHz oscillator on my 486/33 motherboard.

The '486 and '386 both require a more symmetrical clock than can be
provided by standard crystal oscillators. (They come with 60%/40% duty
cycle normally.)  You can buy 55%/45% ones but these are more expensive,
come from fewer suppliers and have longer lead times.

Hence you will find that '386 machines have an oscillator with a
tighter duty cycle spec at twice the cpu frequency(*) and '486 machines
will often have an ordinary oscillator at twice the cpu frequency which
is divided by two to give the correct duty cycle.  However, with a
'486DX-50, this becomes very difficult.  Particularly if you need to
meet FCC and CISPR RFI standards.

*note: TTL crystal oscillators top out at ~80MHz which is twice the top
frequency that you can buy '386s in.  Funny about that :-)

--
Scott Colwell                Labtam Australia Pty. Ltd.
net:    [e-mail]    41 Malcolm Rd., Braeside
phone:    [phone]            Melbourne, Australia  3195