home | list info | list archive | date index | thread index

Re: [OCLUG-Tech] How clock (cycle) distribution works ??

William Case wrote:

> 
> The one thing that is missing is how the clock signal is distributed to
> the various flip-flops.  This is a process that must involve more than
> just an oscillator and a line to a register.  I.e it must have have it's
> own unique problems such as skew and jitters, interrelating various
> clocks, plus, plus.  There surely is some discussions somewhere of
> various design considerations.

Most of that design is part of the Vodoo of the chips themselves.  At
those sorts of frequencies signals start acting very strange, indictors
look like wire leads (and Visa Versa) the circut board looks like a
whole lot of capacitors, a 6 inch lenth of wire acts as a long delay
line and signals travel on the outer edge of the wires.

I suscpect to undersatnd how to set up the system one would need some
courses in RF engineering, Microwave design (a microwave oven runs at
2.5 GHz, so a 3.8 GHz computer is well up in that range.) as well as
semiconductor techniques.  The chips must use a lot of "funny stuff" to
get things to work.

Years back my Wife worked on a telephone transmison system, and she
ended up making 3 dimensional circuts with the lead bonds on the
transistors acting as a tuned inductor.  That was shortly before they
invented Fibre Optic Cables.  Modern Computers run at a higher frequency.

> I have searched online and found lots of very technical papers that
> assume the reader has a level of knowledge about the problems and
> proposed solutions of clocking, that I do not have but would like to
> acquire.  For example, there was one paper that discussed H trees that I
> kinda sorta understood, but would like to learn more.  On the other side
> of the available information conundrum, is just very simplified
> explanations that I came to understand a several weeks ago.
> 
> As usual, I want to get from one level of understanding to the next.
> 
> Any direction to an appropriate text or online site would be very much
> appreciated.

At the edges, i suspact that it all boils down to applying the math to
the real world.  the variables are such that there is probaly still some
cut-and-try involved.



-- 
Charles MacDonald               Stittsville Ontario
 cmacd [ at ] TelecomOttawa [ dot ] net        Just Beyond the Fringe
        http://www.TelecomOttawa.net/~cmacd/
   No Microsoft Products were used in sending this e-mail.