Crossing Ground Plane Breaks - Part 2
Tracing Current Paths
Figure 1. Test Board with Square Loop Positioned for Measurement
Address: P. O. Box 1457, Los Gatos,
CA 95031
TEL:
800-323-3956/408-356-4186
FAX:
408-358-3799
Mobile: 408-858-4528
URL:
www.dsmith.org
Email: doug@dsmith.org
Figure 1. Test Board with Square Loop Positioned for Measurement
Abstract: EMC and signal integrity engineers know that a signal crossing over
a break in a ground plane often causes reflections as well as immunity and
EMC problems. In
this Technical Tidbit, a method is shown to determine where signal currents
flow when a signal path crosses a break in a ground plane. Scope waveforms
are included to show typical results. The results clearly show the loop
(antenna) that is formed by the signal and its return path in the ground
plane.
Discussion: Figure 1 above shows a test board with two paths, both
run about 12 cm from a BNC connector to a 47 Ohm load over a ground plane.
One path stays over the solid ground plane while the other path crosses a
5 cm cut in the ground plane. This board simulates a 4 layer board with cuts
in both the power and ground planes. The path crossing the break in the ground
plane will be of primary interest for this Technical Tidbit article. The
short wire loops soldered to the ground plane on the left and right sides
are for measuring ground plane voltage in another experiment and are not
used for this article.
Figure 3 shows the loop output at position A. The loop output is the
derivative of the signal current (M di/dt), but it can be used to sense where
the current is flowing and its direction. The signal source is a square wave
from the 5-50 MHz oscillator
described on this site. The positive and negative going peaks correspond
to the edges of the square wave. The edges of the square wave are a little
"softer" or rounder than the oscillator would normally have because of the
heavy load presented by the 47 Ohm resistor. The scope is triggered directly
from the oscillator so the relative directions of the edges of the waveform
can be compared.
When the loop is moved to position B with the same orientation, the plot
in Figure 4 results. Notice that it is inverted from Figure 3. This
means that the current is flowing in the opposite direction around the end
of the ground break. The amplitude is a little smaller in Figure 4 likely
because the current is not parallel to the loop for its full length (it bends
around the end of the break) and the lower inductance of the ground plane
compared to the signal wire.
| 10 mV/div |
|
||
| 10 ns/div |
|||
| Figure 3. Loop Voltage (Position A) |
|||
| 10 mV/Div |
|
||
| 10 ns/div |
|||
| Figure 4. Loop Output Voltage (Position B) |
|||
Figures 5 and 6 show the loop output at positions C and D respectively
with the loop held in the same orientation and just slid from C to D. Note
again the reversed current directions indicating the current is flowing down
one side of the break and up the other. If a smaller loop (for better resolution)
was scanned over the board, one could trace out the complete signal path
from the source to the load and back to the source again.
| 10 mV/Div |
|
||
| 10 ns/div |
|||
| Figure 5. Loop Output Voltage (Position C) |
|||
| 10 mV/Div |
|
||
| 10 ns/div |
|||
| Figure 6. Loop Output Voltage (Position D) |
|||
One can see from the data above that the signal return current flowing
in the ground plane is diverted to the end of the break and thus forms a
substantial loop area with the signal path for the signal current. This large
loop has many implications for system operation including being more susceptible to external EMI (electromagnetic interference).
For related information pertaining to magnetic loops on this website see:
The waveforms in this article were taken with an Agilent
Infinium 54845a oscilloscope.