The Bridge - Issue 2, 2022 - 25

Mitigating Self-generated EMI for Wireless Devices
Feature
Modern technology has enabled extreme levels
of amplification to be easily achieved within a
receiver; thus, receiver amplification is not a factor
limiting receiver sensitivity. Instead, the limiting
factor in modern receiving antennas or receiver
systems is noise: a weak signal is not limited by
the actual strength of the signal but rather by the
noise that masks it out. This noise can come from
a variety of sources. At frequencies above 30 MHz,
the noise generated from neighboring electronic
devices becomes far more important than external
environmental noise. The main type of noise limiting
receiver sensitivity is internally generated by digital
integrated circuits (ICs), called self-generated EMI.
For example, RF modules in smartphones, which are
composed of an RF antenna and its receiver, can
detect signals as weak as -120 dBm in a 200 kHz
bandwidth if the module is not disturbed by nearby
electronics. However, the clock frequencies of a
smartphone can reach the GSM 880-1800 MHz
band, as well as Bluetooth and Wi-Fi bands [2], thus
limiting the receiver's ability to detect low-level signals
and therefore decreasing the overall range of
the receiver.
Because of the complexity of real products, there
are several approaches, including trial and error
approaches, that are available for troubleshooting
EMI/EMC issues including for self-generated EMI.
Typical solutions involve adding shielding or absorbing
components (usually later in the development cycle),
which not only can alter RF antenna performance
(detuning its resonant frequency) when placed in
proximity but also add extra cost to the product. An
alternative mitigation solution is damping the signal
edges (intentionally degrading signal quality) to
decrease the energy of the noise radiation source;
however, this approach comes at the cost of the data
rate and thus is not desirable in modern high-speed
digital systems. Departing from the conventional
mitigation approaches, we introduce a new paradigm
of self-generated EMI mitigation without absorbing/
shielding components. Our approach also does not
compromise signal integrity. The concept is based on
and derived from a solid EM framework to model the
self-interference EMI problems, and is demonstrated
with real smart speaker designs.
II. Self-generated EMI
A three-step radiation source-coupling path-victim
antenna process has been widely used to model
EMI coupling. However, self-generated EMI requires
a different approach. In wireless devices, the RF
antennas are realized on the printed circuit board
(PCB), called on-board antennas, by using the
entire PCB as a part of the radiating structure. The
digital ICs are also populated on the PCB, which is
a part of the RF antenna. Thus, we use a two-step
radiation source-transfer function process to model
self-generated EMI. The transfer function represents
how much noise couples to the RF antenna given
the noise location and type, that is, the RF antennas'
susceptibility depending on the location/type of
noise source.
First, we model noise radiation from digital ICs
and their interconnects. In the circuit domain,
any circuit's sources can be represented by using
either an equivalent voltage or current source,
with corresponding source impedance, according
toThevenin's and Norton's theorems. Modeling of
EM radiation is quite different from circuit domain
modeling and requires modeling of the EM field
radiating out in three-dimensional space. The most
common method uses Huygens' equivalence
principle: an imaginary closed surface with impressed
electric (Js
) and magnetic current (Ms
) sources
Fig. 1. A 2D orthogonal near-field scanner from API [3]
HKN.ORG
25
https://hkn.ieee.org/
The Bridge - Issue 2, 2022

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