Date: Thu Apr 12 2001 - 19:18:09 PDT
Using a demonstration board I designed for a high-speed design seminar for
Hewlett-Packard, I tested multiple 50-Ohm configurations (microstrip, guarded
microstrip, stripline, and guarded stripline) for losses, radiation off the
board, crosstalk, and signal fidelity. The FR4 board demonstrated losses of
~3 dB at 1.25 GHz and 6 to 8 dB at 2.5 GHz for just a 9-inch length. The
average attenuation of 0.78 dB/inch for 20 inches gives 15.6 dB attenuation
(for 2.5 GHz). The rolloff (combined losses caused by both dielectric and
skin effect) descended at ~3 dB/octave above 1.5 GHz. Therefore, I would
predict ~6 dB attenuation for the fundamental, ~18 dB attenuation for the
third harmonic (3.75 GHz) and ~21 dB attenuation for the fifth (6.25 GHz) for
20-inch traces in FR4. (Please don't treat these numbers as highly accurate,
as I am drawing on memory for expediency.)
Taking the magnitude of the fundamental as unity, the initial 3rd and 5th
harmonics would be 1/3 and 1/5 of the fundamental amplitude to start with.
Were the attenuation to be equal over frequency, the waveshape would be
retained, only reduced in amplitude. However, the greater attenuation of the
higher harmonics ultimately leads to a rough facsimile of only the
fundamental arriving at the termination/destination. I have beautiful
examples of this situation using Vitesse's VSC880 and VSC870 parts that are
used in a Nokia system I am working on.
Note that, since the amplitude of the harmonics RELATIVE TO THE FUNDAMENTAL
are reduced, and even though the amplitude of the fundamental is reduced, the
peak of the fundamental wave is not canceled/reduced by the higher harmonics.
For the same reason, the leading and trailing edges are not as sharp as when
the wave began its journey. This condition yields a peak-to-peak amplitude of
the arriving signal that is close to the original transmitter peak-to-peak
signal, but it looks more like a sine wave.
Michael L. Conn
*** Serving Your Needs with Technical Excellence ***
**** To unsubscribe from si-list or si-list-digest: send e-mail to
firstname.lastname@example.org. In the BODY of message put: UNSUBSCRIBE
si-list or UNSUBSCRIBE si-list-digest, for more help, put HELP.
si-list archives are accessible at http://www.qsl.net/wb6tpu
This archive was generated by hypermail 2b29 : Thu Jun 21 2001 - 10:11:34 PDT