Muriel, you asked:
>>For example, in a simple circuit (a DC battery feeding a resistor), do I
>>have common-mode emissions? Are the common-mode emissions inherent from
>>any physical system? Can I model them in HF?
The DC situation you describe will not produce AC emissions, but is a good
model to start with in understanding what happens. Draw a square. Draw a
battery in the left side. Draw a resistor (RL) in the right side. This is
your internal load. Now place a resistor in both the top and bottom sides.
For now, let them both have the same value (Rs). This is the "stray"
impedance in the paths to and from RL. Place ground at the lower left hand
corner. Now draw horizontal lines from both ends of RL resistor to the right.
This represents your signal/ground cable going to a "remote" load, which we
can assume for now is high impedance, i.e. no significant current.
Notice that even in this simple DC situation, the signal/ground signals are
shifted by an amount equal to I x Rs, where "I" is the current around your
square circuit.
Now replace the battery with a very low frequency signal generator. Now the
signal/ground cable shift is varying at some frequency, and results in some
amount of common-mode noise being emitted by the cable acting as an antenna.
Suppose you add braided wire shield around your cable. If you ground the
shield to the "lower left" ground, you may reduce the emissions. However,
if you ground the shield at outgoing ground (below RL), it merely oscillates
with the signal/ground wires.
Now increase the frequency of the signal generator. At some point, we can no
longer represent the stray impedances as RS, but a very complex combination
of parasitic resistance, capacitance, and inductance. At extremely high
frequencies, the currents may not even flow in the intended conductors, but
through chassis and cover parts.
Now let's convert the single circuit model into a siple PCB model.
Draw a square. This is the typical logic PCB.
Ground symbol the lower left corner. This represents the incoming voltage/
ground connector from the power supply, and is the "best" ground in the
system.
Draw a line leaving the PCB at the upper right corner. This represents
a signal cable leaving the card and/or device, and usually includes an
internal ground signal. This cable acts as an antenna.
It is not uncommon for a PCB to have the incoming power at one corner,
and outbound cables at an opposite corner, hence this worst case model.
Now draw a voltage generator from the grounded corner to the "output"
corner. If the ground impedance across the PCB were perfect, i.e. zero
resistance at all frequencies, there would be no voltage driving the
cable/antenna, hence no common mode noise.
However, in the real world, the signal generator represents the time
varying shift in the output "ground" across the PCB due to the impedance
in ground across the PCB at the many frequencies involved.
The resulting AC voltages/frequencies driving the cable/antenna create
a common mode effect that will not usually be reduced by adding a shield,
if the shield is attached at the output ground of the card.
Two of the main tricks in reducing this common mode noise are to (1)
partition high powered/high frequncy components on the PCB from low level
logic, and (2) create as many ground paths as possible across the PCB.
The latter usually includes filling unused space with ground patterns,
stitching multi-level grounds wherever they cross on the PCB, etc. Since
it is difficult to predict where electrons will choose to go at MHz and
GHz frequencies, the idea is to provide as many paths as possible.
Hope this helps.....
George Alspaugh
Lexmark International Inc.
-------------------------------------------
This message is from the IEEE EMC Society Product Safety
Technical Committee emc-pstc discussion list.
To cancel your subscription, send mail to:
[email protected]
with the single line:
unsubscribe emc-pstc
For help, send mail to the list administrators:
Jim Bacher: [email protected]
Michael Garretson: [email protected]
For policy questions, send mail to:
Richard Nute: [email protected]
