Electromagnetic interference manifests in various forms such as electromagnetic radiation, electromagnetic induction, or electromagnetic coupling. It refers to the electromagnetic radiation or fields generated during device operation, ultimately leading to varying degrees of interference and damage to nearby devices, electronic equipment, or communication systems. It’s worth noting that the extent of electromagnetic interference is often influenced by several factors, including signal frequency, signal rise time, and high-frequency noise, making it more prominent in high-speed PCB designs. In this article, TechSparks delves into a theoretical analysis of electromagnetic interference in high-speed PCB to help readers better understand and address this critical issue.
High Speed PCB EMI Radiation Types
In high-speed PCB, EMI radiation originates from routing and I/O cables, with cables having the most noticeable impact, closely related to factors such as length, routing method, and cable shielding. During signal transmission, although some cables may have lower transmission frequencies, their strong radiation capability can ultimately couple signals onto the cable, causing EMI radiation.
EMI radiation is classified into common-mode radiation and differential-mode radiation based on the different ways in which currents are transmitted. Common-mode radiation occurs when interfering currents flow in the same direction in two conductors, while differential-mode radiation is the opposite. This distinction can be explained using equivalent electric dipole and magnetic dipole models. In practical circuits, when currents flow in the same direction, a stronger electric dipole is formed, making common-mode current-induced radiation more pronounced. Therefore, special attention is required during high-speed PCB design.
Differential Mode Radiation
When a device is in operation, currents in the circuit form a loop instead of flowing directly back to the power source. This means that a small electromagnetic loop antenna is established, and the radiation electric field intensity can be calculated using the following formula:
- E: Electric field intensity, V/m;
- f: Frequency of the differential-mode current loop, Hz;
- A: Area of the current loop, cm2;
- I: Differential-mode current in the current loop, A;
- r: Distance from the radiation source to the affected device, m;
- θ: Angle between the vector and the Z-axis.
We consider the worst-case scenario for testing, where the phase of the reflected wave is in-phase with the direct wave, and the emitted wavelength is twice the wavelength of the direct wave. Assuming sinθ=1, as in practical field tests, ground reflections are inevitable, so the actual calculated maximum value may increase by a factor of two, i.e.,
According to the electromagnetic performance standards of different countries, the range of values for r is typically between 1m and 30m. According to the formula, E is directly proportional to f, I, and A. Therefore, TechSparks recommends the following measures to mitigate EMI’s impact on high-speed PCB: reducing the area of the current loop, decreasing the magnitude of the differential-mode current, and lowering the frequency of the differential-mode current loop.
Common Mode Radiation
To analyze common-mode radiation, we can use the dipole antenna model, and the formula for calculating electric field intensity is as follows:
- E: Electric field intensity, V/m;
- f: Frequency of the common-mode current, Hz;
- I: Intensity of the common-mode current, A;
- r: Height of the I/O dipole antenna, m;
- r: Distance from the measuring antenna to the dipole antenna, m;
- θ: Angle between the vector and the Z-axis.
To suppress common-mode interference, we can employ the following methods: use differential circuits, lower the ground potential of the excitation antenna, add common-mode chokes, and shunt common-mode currents to ground.
Impact of Passive Component High-Frequency Characteristics
Modern electronic systems demand higher performance and sensitivity, necessitating faster signal transmission rates. As frequencies rise and clock signals exhibit steeper rising edges, there is a risk of electromagnetic interference, even if designers employ shorter traces. The hidden characteristics of passive components often underlie these issues.
In circuit design, improper parasitic elements, capacitance values, or routing configurations can turn a circuit into an electromagnetic radiation source, emitting electromagnetic waves outward. Typically, radiation antennas operate within specific frequency ranges, with wavelengths around 1/4 or 1/2. However, the operating frequencies of high-speed PCB is often much lower, around 1/20 of the wavelength of fixed-frequency waves. This frequency selection aims to minimize the possibility of the circuit acting as a potential electromagnetic emitter.
Edge Radiation from High Speed PCB Power/Ground Planes
When the harmonic components of the via switch current match the resonant frequency of the resonant cavity formed by the power/ground planes on a high-speed PCB, the PCB planes resonate, leading to the emission of a significant amount of electromagnetic waves outward near the resonant frequency. This type of radiation propagates through the PCB edges and is known as edge radiation.
In the ideal case of a perfect electrical conductor boundary (edge short-circuit), the electric field is zero, meaning that the electric field is effectively shielded within the edge and does not radiate from the plane’s edge. Conversely, in the ideal case of a perfect magnetic conductor boundary (edge open-circuit), the electric field’s intensity is maximum near the edge, potentially causing significant radiation.
When high-speed digital circuits operate, the gate circuits inside integrated circuits are responsible for processing digital signals, switching them from 0 (low level) to 1 (high level), or vice versa. This switching process involves the conduction and cutoff of transistors, which control the flow of current. When transistors in the gate circuit turn on, current flows from the power supply circuit into the gate circuit, and when transistors turn off, current returns from the gate circuit to the ground. Due to minor mismatches or asymmetries in practical circuits, the returning current may not be perfectly balanced, resulting in transient current variations. When this current variation passes through inductive components in the circuit, it induces an alternating voltage drop across the inductance, leading to noise. This includes two scenarios:
- When multiple components are simultaneously in a switching state, the current demand changes rapidly. Since this current change is high-frequency, the voltage variation induced on the inductance is also high-frequency, i.e., alternating voltage, which is a source of noise.
- When a signal passes from one PCB layer through vias to another layer, it must traverse the power/ground planes between layers. This process leads to changes in the reference plane. These changes in the plane cause the return current path for the signal to be discontinuous, resulting in ground bounce noise on the PCB.
Microstrip Antenna EMI Mechanism
The principle behind microstrip antennas causing EMI issues in high-speed PCB is fundamentally similar to the radiation principles of power/ground planes. Microstrip antennas are a type of small, lightweight, flexible, and easily integrated antenna. They consist of a conductive material (usually copper) mounted on an insulating substrate, with the copper serving as the antenna radiator.
A microstrip patch antenna can be thought of as a section of microstrip transmission line, with one end being an open circuit with no electrical connection, and the other end forming the antenna’s edge. When a microstrip patch antenna operates, electrical signals propagate along the microstrip line with a length represented by “L.” As the microstrip patch antenna’s endpoint is an open circuit, the electrical signal reflects at this point, creating a voltage wavefront. This voltage wavefront can be used to emit electromagnetic waves, thus enabling the antenna’s transmitting function. Typically, the length “L” is chosen to be approximately L ≈ λm/2, where λm represents the wavelength on the microstrip line.
Consequently, at the other end of the antenna, there is also a voltage wavefront. At this point, the electric field can be approximately expressed as:
The radiation of the antenna is formed around the microstrip patch, mainly in the narrow gap between the patch and the ground plane. According to the equivalence principle, the radiation of the electric field on the narrow gap can be equivalently described by surface magnetic current radiation. The equivalent surface magnetic current density is:
On both wide edges of the microstrip patch antenna, the flow direction of the magnetic field is the same. Such magnetic currents reinforce each other in the normal direction (usually represented by the z-axis), producing the maximum radiation field. As the angle deviates from this normal direction, the radiation field’s intensity gradually decreases, forming a radiation pattern. Along each of the “L” edges, the magnetic current is composed of two anti-symmetric parts, canceling each other out at various locations on the H-plane (yz-plane). However, the magnetic currents along the two “L” edges are anti-symmetrically distributed with respect to each other. Thus, on the E-plane (xz-plane) at various locations, their fields cancel each other out. On other planes, the radiation from these magnetic currents does not completely cancel out but is considerably weaker compared to the radiation along the two “W” edges.