Path: utzoo!utgpu!news-server.csri.toronto.edu!mailrus!uunet!fernwood!portal!cup.portal.com!ISW
From: ISW@cup.portal.com (Isaac S Wingfield)
Newsgroups: sci.electronics
Subject: Re: A 74F74?
Message-ID: <28826@cup.portal.com>
Date: 12 Apr 90 06:50:17 GMT
References: <9561@sdcc6.ucsd.edu> <37746@mips.mips.COM>
  <1990Apr8.000826.16803@utzoo.uucp> <1990Apr8.061809.6501@ns.network.com>
  <29798@amdcad.AMD.COM> <1990Apr10.225733.27152@ns.network.com>
Distribution: na
Organization: The Portal System (TM)
Lines: 24



John Logajan writes:

>ACT parts have TTL compatible input circuits, whereas AC parts need CMOS
>levels.  Both ACT and AC outputs swing the full CMOS range, which certainly
>would seem to imply a greater need for termination in high speed use.

I believe that edge rate (transition time) is a more significant indicator of
the need for termination (and also how much noise radiation is likely) than
voltage swing per se. As edge rate goes up, more and more energy shows up at
higher frequencies; at higher frequencies, shorter and shorter traces begin to
exhibit the transmission line effects that make termination necessary.

For equal propagation delays (and therefore equal system speeds), different
logic families have different edge rates, and so some will require
terminations, and similarly will be much noisier, while others will not.
Sorry, but I've been out of hard-nosed design for a while, and can't comment
on which of the current families are which...


Isaac
isw@cup.portal.com