“Follow the currents.” That statement was made by Dr. Bruce Archambeault, who says that current flow is the most important concept of EMC. I have to agree with him because if we know how currents are being generated and how they move through our circuits and chassis, we can understand our sources of emissions and coupling mechanisms of susceptibility in greater ways.
In EMC engineering, we try to classify currents into two categories: Common Mode (CM) currents and Differential Mode (DM) currents. Differential mode currents are easier to understand. A unit demands power from a power line, e.g., a 28 VDC power bus. The current then flows back on the return line. This is the differential mode current.
However, in the operation of the equipment, some current may be inductively or capacitively coupled to the chassis, to other circuits, or used by the system and routed to other lines. This can result in an imbalance in the power line currents. In this case, 1.001 amps may flow in the power line, but only 1.000 amps flow in the return line. The result is that we have 1.000 amps of differential mode current and 0.001 amp of common mode current, which has a remote return path.
Note these currents are due to inductive or capacitive coupling; thus, I am assuming they are high frequency currents since DC is coupled neither inductively nor capacitively. To be high frequency, some function must occur, such as being chopped by a switching power supply or used in digital circuits. Considering power supplies, FETs and transformers used in switch mode power supplies create fields by their operation. A FET turning on and off will have high frequency transients from the sudden starting and stopping of the current flow.
Looking at this FET as the noise source – the voltage will spike when the FET transitions from a conducting to a non-conducting mode. This FET voltage spike will be with respect to the reference plane or the chassis. High frequency, short duration voltages between two conductors will induce a current between them. However, once generated, currents must find a path back to the source (currents flow in loops). The trouble comes when that return current path is either unknown or uncontrolled. Our last Military and Aerospace EMC article stated that radiated emissions can be over limit with as little as 10 µA of uncontrolled common mode current.
Common mode inductors, ferrites, and the like are often employed to control the flow of these currents out of the equipment. However, the addition of impedance in series is most successful when there is a local return path for the currents to flow. This means the use of capacitors from the reference plane or chassis back to the line connected to the FET in this case. In doing so, a closed loop for the current is provided, and it will be a preferred path when that common mode inductance is placed on the outboard side of the capacitance.
Realize that loop areas produced by these current paths are also receiver antennas as much as they are transmitting antennas. They can receive and then inject interference-like energy into the circuit, which may disrupt the operation of sensors, digital circuits, and the like, which may be connected to the same current path. Thus, the control of current loops has two beneficial effects: the reduction of emissions and greater immunity to outside signals.
When differential noise is an issue, the solution will be to use capacitance from line to line, and linear inductance, preferably in both the power lead and the return. When a switching circuit demands current, providing a local source for this current in the form of capacitance can reduce or soften the demand from the power line. Series inductance should not be common mode inductance but linear inductors designed to handle peak current demand without saturation. Using ferrites as the core material for linear inductors can be problematic. When used on an AC line and run to saturation, ferrites have been found to create more noise than when they are not in circuit. Proceed with caution when using inductors.
Finally, single-ended signals are not truly single-ended. Current will not flow from here to there and will not have a return path. So, returns tend to be other than adjacent lines. If this is a pure DC line, there is no issue. The problem is that radio frequency energy is easily coupled and can appear on these lines. With a remote return, the loop sizes increase, and so do emission and susceptibility problems. Thus, single-ended lines are not a cure for EMI problems, and they can have their own set of issues.
Ultimately, if we follow the currents and know how they are generated and where they travel, we can understand a great deal more about how to return the currents locally and avoid them from leaving the equipment in the first place.
