Cirrus Logic AN Integrating CobraNet into Audio Products - Manual (Page 13)

EMI P RELIMINARY
11
CS18101 (CM-2) AppNote1 - rev 1.1 Jan, 2004
w w w . c i r r u s . c o m
EMI
Like thermal issues, the theory of EMI is well known but the practice is something else
entirely. There are so many factors that can effect EMI that dealing with EMI can be akin
to black magic. In fact, the title of a good book on EMI reflects that (and we highly
recommend that book):
High Speed Digital Design: A Handbook of Black Magic
Howard W. Johnson & Martin Graham
Prentice-Hall, Inc.
ISBN: 0-13-395724-1
We won't attempt to explain how to solve EMI issues, since someone already wrote that
book. But we will go over some things specific to the CobraNet Modules that may help
with EMI.
Follow the section “Maintaining Signal Quality” above. Proper signal termination, signal
loading, de-coupling, etc. can all have a dramatic effect on EMI.
CobraNet Modules have zero-ohm resistors and/or caps that connect chassis ground,
digital ground, and various mounting holes. Depending on the situation, it might be
advantageous to remove or change these components.
The module faceplate (or the RJ-45 connectors, if you're not using the faceplate) must
make good electrical contact with the chassis. It's common for paint or other "debris" to
cause electrical isolation, increasing EMI. During testing, you can try to use copper tape
to make and seal this connection. A bad connection here can increase radiated emissions
by 6db or more!
If possible, don't use a ribbon cable to connect the module to the host system. A direct
connection will work much better, both from a signal-quality point of view as well as an
EMI point of view.
When doing EMI testing, try a wide range of Ethernet switches. Experience has shown
that different switches can have as much as a 12db difference in common mode
emissions. So try different switches until a "good one" is found.
Your choice of power supply can also effect EMI and common mode emissions. Choose a
good one!