Radio Frequency Interference Sources Today, the electronic system has a clock frequency of several hundred megahertz. The leading and trailing edges of the pulses used are in the sub-nanosecond range. The network interface transmits data rates of 100 Mbit/s and 155 and 622 Mbit/s (ATM-Asynchronous Transfer Mode). High quality video circuits are also used for subnanosecond pixel rates. These higher processing speeds represent constant challenges in engineering.

One of these challenges is radio frequency (RF) interference due to rapid changes in electromagnetic energy. The rate of oscillation on the circuit becomes faster (rise/fall times), the voltage/current amplitude becomes larger, and the problem becomes more. Therefore, solving electromagnetic compatibility (EMC) is even more difficult today than before.

Before the two nodes of the circuit, the rapidly changing pulse current represents the so-called differential mode noise source. The electromagnetic field around the circuit can be coupled to other components and intruded into the connection. Inductive or capacitively coupled noise is common-mode interference. The RFI currents are the same as each other and the system can be modeled as: consisting of the noise source, the "victim circuit" or "receiver" and the loop (usually the backplane). Several factors describe the size of the interference:

● The intensity of the noise source ● The size of the interference current in the surrounding area ● The rate of change Therefore, although there is a high probability of unwanted interference in the circuit, the noise is almost always modeled. Once the cable is connected between the input/output (I/O) connector and the chassis or ground plane, there are certain RF voltages that can cause a few milliamps of RF current to be sufficient to exceed the allowable emission level.

Noise coupling and propagation common mode noise is due to an unreasonable design. Some typical reasons are different lengths of individual wires in different wire pairs, or different distances to the power plane or chassis. Another reason is component defects such as magnetic induction coils and transformers, capacitors and active devices (such as application-specific integrated circuits (ASICs)).

Magnetic components, especially so-called "core choke" type energy storage inductors, are used in power converters and always produce electromagnetic fields. The air gap in the magnetic circuit is equivalent to a large resistor in the series circuit, which consumes more power. Thus, the core choke, wound on a ferrite rod, produces a strong electromagnetic field around the rod and has the strongest field strength in the vicinity of the electrode. In switching power supplies that use a retrace structure, there must be a gap in the transformer with a strong magnetic field in between. The most suitable element for keeping the magnetic field in place is a coil, which distributes the electromagnetic field along the length of the die. This is one of the reasons why the magnetic element preferably operates in high frequency at a spiral structure.

Inappropriate decoupling circuits often also become sources of interference. If the circuit requires large pulse currents, and local capacitors do not guarantee a small capacitance or a very high internal resistance, the voltage generated by the power supply circuit will drop. This is equivalent to ripple, or equivalent to a rapid change in the voltage across the terminals. Due to the stray capacitance of the package, interference can couple into other circuits, causing common-mode problems.

When common-mode current contaminates the I/O interface circuit, this problem must be solved before passing through the connector. Different applications propose different methods to solve this problem. In video circuits, where the I/O signals are single ended and share the same common loop, to solve it, noise is filtered out with a small LC filter. In low-frequency serial interface networks, some stray capacitance is enough to shunt noise onto the board. Differentially driven interfaces, such as Ethernet, are usually coupled via transformers to the I/O area and are coupled at the center taps on one or both sides of the transformer. These center taps are connected to the backplane via a high voltage capacitor, shunting common mode noise onto the bottom plane so that the signal does not distort.

Common mode noise in the I/O area does not have a general approach to solve the problem of all types of I/O interfaces. The main goal of designers is to design the circuit, and often overlook some of the details that are considered simple. Some basic rules can minimize noise before it reaches the connector:

1) Set the decoupling capacitor immediately before the load.

2) Rapidly changing pulse currents on the leading and trailing edges, and the loop size should be minimized.

3) Keep high-current devices (ie, drivers and ASICs) away from the I/O ports.

4) Determine the integrity of the signal to ensure minimal overshoot and undershoot, especially for high-current critical signals (eg, clock, bus).

5) Use local filtering, such as RF ferrite, to absorb RF interference.

6) Provide a low impedance bridge to the backplane or in the I/O area on the backplane.

RF Noise and Connectors Even if engineers take many of the above listed precautions to reduce RF noise in the I/O region, there is no guarantee that these precautions will succeed enough to meet the emission requirements. Some of the noise is conducted interference, that is, the common mode current flows on the internal circuit board. This source of interference is between the backplane and the circuit. Thus, this RF current must flow through the path of the lowest impedance (between the base plate and the carrier signal line). If the connector does not exhibit a sufficiently low impedance (at the junction with the backplane), this RF current flows through the stray capacitance. When this RF current flows through the cable, emission is inevitably generated (FIG. 1A).

Another mechanism for injecting common-mode current into the I/O region is the coupling of strong interference sources nearby. Even some "shielded" connectors are useless because the source of the interference is near the connector, such as a PC environment. If there is a gap between the connector and the backplane, the RF voltage induced here can degrade EMC performance (Figure 1B).

Shielded connector methods include finger reeds or pads. The lap of the connector is to fill the empty space between the connector and the chassis. This method requires a pad (Figure 2A). The metal liner is preferably as long as the treatment is appropriate, that is, as long as the surface is not contaminated, as long as the hand does not touch or damage the liner and as long as there is sufficient pressure to maintain a good, low-impedance contact.

Another method is to attach the connector to the connector or install the connector on the case. At this time, the maximum contact surface is slightly smaller, and the size and elasticity of the tab should be strictly controlled. When the shielded connector is installed, it should be opened on the casing, and the side of the opening should be free from oil stains (Figure 2B). Care must be taken if the tolerances are not appropriate, causing the connector to sink too deeply in the casing and interrupting the connection. Every EMC engineer knows that in an "excellent" system, this problem must meet the launch requirements and be checked on the production line. Unsecured or curved gaskets, which are installed on the oily area of ​​critical areas (such as the opening where the connector is installed), will fail.

The EMI connector was chosen for the following reasons;

1) Conductive foamed plastic is extremely soft and can be placed around the connector. This eliminates the problems associated with another housing and gasket.

2) The mechanical engineer can install the connector within the acceptable tolerance of the system casing.

3) The connector and chassis are connected with low impedance to ensure good contact. The liner on the inside of the housing wall can be made of a softer material when it is painted to provide shielding.

4) For designs requiring forced cooling, the liner should preferably have another feature: the seam between the connector and the housing wall should be sealed to reduce air leakage. In a dusty environment, the gasket must be kept clean inside the system.

Conclusion There are a variety of connectors on the market today that enable designers to interface with a particular design for optimal design.