I believe contributors to this thread have been citing the following paper
(unnamed until now):

"Article for Printed Circuit Design,  By Lee W. Ritchey,  3Com Corporation
'Differential Signaling Doesn't Require Differential Impedance
Or,  How to Design a Differential Signaling Circuit' "

I feel compelled to report that even though the author presents some
noteworthy viewpoints about the practical aspects of differential logic
transmission circuits, some of his assertions are in error, therefore his
ideas must be applied with caution.

In the fifth paragraph, the author presents the premise,
".................[differential signaling provides] immunity from common
mode noise coupled into the pair by outside noise sources.
...........because any magnetic noise field couples equally into both wires
resulting in a common mode signal."

This statement is fundamentally WRONG. The pair in question forms a loop
that is in effect a single turn pickup coil with the drivers serving as a
virtual center tap. An emf is generated at the single turn coil terminals by
varying magnetic flux fields cut by the plane of the coil. This is a
differential voltage, NOT a common mode voltage. The greater the coil area,
the more flux lines pass through the plane of the coil, consequently, the
larger the emf. The shape of the loop doesn't matter; only the area. Any
standard physics text or electromagnetics reference will verify this.

The following sentence goes on to say, "[This] is of [no] real benefit when
the wires are routed over the planes of a PCB."  As a consequence of the
initial wrong premise, this statement is also wrong.

First of all, magnetic fields penetrate copper planes, with the degree of
penetration dependent on the frequency. Therefore, trace loops adjacent to
the planes will pick up these fields, the degree also dependent on the
frequency. Finally, a trace loop voltage (emf) is a function of the area of
that loop. Since a differential pair is effectively a loop, minimizing the
pair separation minimizes the area, hence the voltage pickup as well.

Clearly, in a magnetic noise environment, the differential pair spacing
cannot be ignored. Once again, be careful.

Fred A Rupinski

----- Original Message -----
From: "Brad Velander" <[EMAIL PROTECTED]>
To: "'Protel EDA Forum'" <[EMAIL PROTECTED]>
Sent: Monday, September 16, 2002 8:47 PM
Subject: Re: [PEDA] Autorouter and Differential Pairs.

> I knew the article and had read it thoroughly in the past several times. >
...................................he has done so much trying to
> quantize or just plain dispel some of the old wives tales that plague this
> industry.

* * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
* To post a message: mailto:[EMAIL PROTECTED]
* To leave this list visit:
* http://www.techservinc.com/protelusers/leave.html
* Contact the list manager:
* Forum Guidelines Rules:
* http://www.techservinc.com/protelusers/forumrules.html
* Browse or Search previous postings:
* http://www.mail-archive.com/proteledaforum@techservinc.com
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * *

Reply via email to