Having spent the first week of August at the IEEE International Symposium on EMC, the one concept that is still ringing in my ears is what is often called “ground” but is actually a power or signal return or reference plane. The reason is that what is commonly called “ground” and what is a power or signal return, or that of a reference plane, is not the same. There were several talks on this concept given by many people who are much brighter than I am and have proven to me they know what they are talking about. And yet, I recently read an article whose author seemed confused about what a ground was.
The concept of electrical “ground” has its basis in the “earth-return telegraph” and the first telephone connections. Signal lines were routed between two points, but the return path used was the ground, literally the earth. The earth is also used for electrical safety return paths, neutral reference, and split phase power reference. In the United States, the electrical power panel in the home or commercial locations (when correctly installed) have the neutral lines and safety grounds tied to a common bus bar, which should be routed to a ground rod near the power panel.
When a signal line is routed in a wire with an adjacent return, or when a power or signal trace is routed over a return plane, the adjacent wire and return plane should not be considered “ground,” and I recommend not addressing it as such. Far too often, I have worked with an engineer who will create a wonderful filter for the power line and have a number of capacitors from power to return without having any inductance or other filtering on that return line. This is typically because the designer has a concept that the return is “ground” and treats it as a hole to throw all types of electrical noise into. The results are usually not desirable as found when the EMC laboratory performs conducted emissions on the return lead.
Consider this: A system uses 28 VDC aircraft power. There will be 28 VDC on the pins of the input connector. What if the same system used +14 VDC on one pin and -14 VDC on the other pin? The voltage applied would still be 28 VDC between the power and return pins, but now, instead of considering the return a “ground,” the designer must consider it a -14 VDC return power. How would the design change? When I asked designers this, they said the -14 VDC line would need the same type of filtering used on the +14 VDC line—correct answer.
But this is the same for the 28 VDC return line. It should not be considered as a ground but as a power line with a voltage that happens to be near a zero-volt reference. The return line is a current caring conductor, having inductance in the lead and traces, capacitance with other metal in the area (wires, traces, components, chassis, and so forth), and can allow radio frequency energy to flow in and out of the equipment on that line. Unimpeded, the return becomes a source of emissions and a path for susceptibility to enter the equipment.
But what of voltage? An aircraft may become charged during flight, and without earth nearby when flying at 33,000 feet, the voltage of the whole structure may shift up or down significantly. At the input connector of the system, the power line may be at 10,000 volts while the return is at 9,972 volts; a difference of 28 VDC appears across the power pins, and the system operating the same as it did when referenced to a 0-volt earth ground. Thus, voltages are relative between two points and may be significantly different than remote locations. However, this is not a concern for us.
A difficulty can arise when a system power or signal return is referenced to the structure of the vehicle. Power and signal return may be the aircraft frame, tank, or vehicle chassis, but this is avoided in Naval ship power to avoid both corrosion and magnetization of ferrous materials in the hull. When a power or signal is routed in a wire to some remote connection location, and the return path is the vehicle structure, a large loop area is created between the current flowing in the two paths. This large loop can create radiated emissions problems with less than 10 µA of radio frequency current. Similarly, currents induced into those wires or into the structure from external sources can induce susceptibility in the system. These signals are considered common mode signals since they flow in or out of one or more wires from the equipment but return by a path that is not adjacent to these wires.
In my next article, I will discuss the concepts of common and differential modes in more detail, loop areas, why they are a concern, and other aspects of current flow.
