Re: [SI-LIST] : Other materials for GHz application

Paul Franzon (paulf@eos.ncsu.edu)
Tue, 2 Dec 1997 09:45:11 -0500

Cheah:

There are no hard and fast rules for clock rate vs. board technology.
Here are some things to think about, however:

o To operate even point to point connections at this rate will require
careful modeling and analysis, including skin effect and especially
process variations. Everything in a board varies +/- 30%. You will
need a carefully designed termination scheme to suppress reflection
noise. The materials used in a PCB can support very wide frequency
ranges but designing low-latency low-noise nets gets tricky and
time-consuming.

o Clock speed by itself is not a driver to MCM over PCB. Common drivers
include size, inter-chip latency, noise control, routability,
and chip I/O count.
However, I will say this. Designing MCM wiring to meet a 200 ps delay
slack spec (to pick a number) is a *lot* easier than designing the same
PCB wiring.

Good luck

Paul

On Dec 2, 14:55, csoolan@dso.org.sg wrote:
> Subject: [SI-LIST] : Other materials for GHz application
> Hello,
>
> Does anyone have information on the following:
>
> 1. Why and What is the limitation of using FR4 up to 500-700 MHz?
>
> 2. Beyond this clock freq, what should be recommended for the PCB and
why?
>
> 3. Under what circumstances, we should migrate MCM and why?
>
> Thanks in advance.
>
> Best Regards.
> ******************************************************************************
> Cheah Soo Lan
> DSO NATIONAL LAB.
> 20 SCIENCE PARK DRIVE
> SINGAPORE 118 230
>
> EMAIL: csoolan@dso.org.sg
>
>-- End of excerpt from csoolan@dso.org.sg

-- 
-----
Paul Franzon, Associate Professor, North Carolina State University
Rm. 443, Engineering Graduate Research Center, Centennial Campus
smail: ECE, Box 7914, NCSU, Raleigh NC 27695-7911
Fedex: Rm. 416, EGRC, 1010 Main Campus Dve, NCSU, Raleigh NC 27606
919 515 7351, fax. 919 515 2285, www.ece.ncsu.edu/erl/faculty/paulf.html