The Bridge - Issue 2, 2022 - 27

Mitigating Self-generated EMI for Wireless Devices
Feature
Fig. 3. A circuit board with Wi-Fi antennas (black), a processor (orange),
and memory (red) [6]
Based on the understanding from (1), two
approaches are possible to mitigate self-generated
EMI from the system design perspective. First, the
magnitude of the transfer function can be decreased
by moving the digital ICs and interconnects to
locations on the board that are less susceptible to
EMI. Placing shielding/absorbing materials in the
coupling path would reduce the transfer function,
but the same goal can be achieved by shifting the
location of the noise source. Second, the radiation/
susceptibility interaction can be decreased by rotating
the digital ICs and interconnects. Due to the inner
product relationship, the coupling can be suppressed
through orthogonality between two vectors (the
noise vector and antenna susceptibility vector). Both
solutions are discussed in greater detail below.
In the example shown in Fig. 3, the noise source type
is My (a magnetic dipole pointing in the y-direction)
[4]; therefore, the transfer function associated with
Hy (the magnetic field in the y-direction) must be
determined. The measured |Hy| with respect to the
embedded Wi-Fi antenna's excitation is shown in Fig.
4 [4]. The near-field scanner shown in Fig. 1 was
used for the measurement. From (1), we know that
weaker coupling occurs in the locations where the Hy
fields are weaker in the field map. Thus, moving the
noise source to the smallest Hy location candecrease
self-generated EMI. However, the minimum transfer
function location is not always possible for the actual
layout. Because the processor and memory along
with its interconnects must be moved as a group, the
space might not be sufficient to place all components
in practical designs.
A method that can work better in practical designs is
rotating the digital ICs and interconnects. Because the
entire circuit is rotated, substantial redesigning of the
layout is unnecessary. To explain this concept, linearly
polarized antennas are illustrated in Fig. 5. When
two antennas communicate, the antennas must be
similarly polarized to ensure optimal performance.
Antennas operating with orthogonal polarization will
not perform well, owing to substantial polarization
losses. Orthogonality therefore allows an antenna
with a given polarization to avoid interference created
by energy from an antenna with an orthogonal
polarization. In an ideal case, two orthogonally
polarized antennas have infinite isolation (i.e., zero
interference). Isolation using orthogonality is a wellknown
concept in the antenna community, but the
case of self-generated EMI differs from antenna
applications because 1) coupling occurs at the nearfield
region of an antenna, and 2) the noise source is
not a well-defined antenna. Using the dipole moment
representation of the digital noise radiation sources
and the EM framework developed in [5] makes it
clear and straightforward to exploit the orthogonality
in self-generated EMI problems.
Fig. 4. Measured |Hy| field map of the circuit shown in Fig. 1.
Fig. 5. Orthogonality to avoid interference between two linearly
polarized antennas.
HKN.ORG
27
https://hkn.ieee.org/
The Bridge - Issue 2, 2022

Table of Contents for the Digital Edition of The Bridge - Issue 2, 2022
