Dave,
sorry for the late respons about the split ground planes (or slots).
It is fascinating to see that fast-edge digital signals can be transmitted
over slots without much reflections and distortions. I have done experiments
on digital signals with a 750ps rise-time (20-80%) that cross a 16cm slot
without any problem. Also frequency-domain measurements indicate that the
transmission is good, except for a few resonance frequencies were there is some
attenuation.
The explanation for the fact that slots (which are relative small in length)
can be crossed by both wideband digital and analogue signals is that the return
current is provided by a slot transmission line, which is formed by the
slot. In general the slot is part of a moat in the plane and the slot line is
terminated by a short circuit (that is, the slot is short circuited by the
plane). However, some slots are open on one-side and the general picture is
that the return path of the current of a signal crossing a slot is carried
by a slot line which is terminated at both sides in at least one short circuit.
RF engineers use slots and microstrips to design filters at very high
frequencies (over 2GHz) or as antennas. The design and synthesis is very
complicated and special "numerical" tools are required to calculate the
coulping of the signal line (microstrip) with the slot line. The basic
problem is the slot line which supports a non-TEM mode (consider it a wave
guide). This non-TEM gives rise to a frequency-depent line impedance and a
propagation velocity. The slot line impedance is very small for low frequencies
and it rises to several tens of ohms for frequencies over several hundreds
of MHZ (or even more).
Note also that different signals crossing the same slot become coupled by
the slot, that is, there is crosstalk from one signal to the other signal
because of the common slot.
Slots are used as antennas and can be very effective, from the EMC viewpoint
a signal crossing a moat in the groundplane is equivalent to a slot antenna
that is excited by the crossing signal. Almost always there is no functional
problem, that is, the digital signal does its job. However the EMC performance
is degraded. I once worked on a "small" OEM scanner contract of 10 million
dollar, the scanner had EMC problems which were related to the CCD circuit. We
solved the problem by providing the signals crossing slots with a ferrite bead.
I can cite several other problems that were solved in this way.
I have been interested in these phenomena for some time and I have followed
the literature very well. It is very difficult to find simple explanations
for these slot lines, however I can list several very technical articles
(e.g. IEEE trans on Microwave Tech. and Tech.) that discuss these matters in
detail. The mentioned books in one of the responses ("Printed Circuit Board
Design Techniques for EMC Compliance", "High Speed Digital Design A Handbook
of Black Magic") do discuss slots and the concept of the return path, however
I would caution you about some of their conclusions, they are not always
backed by serious experiments or analysis.
My conclusions is that it is unlikely to have functional problems with
slots (however, much depends on the slot dimesions and length). However,
from an EMC viewpoint signals crossing a slot should be avoided, because
slots can be efficient radiators (depends on slot length).
I know of one very good article on the signal-slot coupling. It appeared
in the IEEE Circuits & Devices magazine. It is clear and simple, the
mathematics are limited. I will try to get the exact reference.
Regards,
Jan Vercammen
Agfa NV
Mechelsesteenweg 430
B2650 Edegem
Belgium
[log in to unmask]
################################################################
DesignerCouncil E-Mail Forum provided as a free service by IPC using LISTSERV 1.8c
################################################################
To subscribe/unsubscribe, send a message to [log in to unmask] with following text in the body:
To subscribe: SUBSCRIBE DesignerCouncil <your full name>
To unsubscribe: SIGNOFF DesignerCouncil
################################################################
Please visit IPC's web site (http://www.ipc.org) "On-Line Services" section for additional information.
For technical support contact Hugo Scaramuzza at [log in to unmask] or 847-509-9700 ext.312
################################################################
|
