Feb 1, 2003 12:00 PM,
A predictable amount of constant, random noise is inherent in all electronic devices, such as background hiss in a sound system or the grainy movement in a video image. But other types of noise are generally much more noticeable and irritating — hum, buzz, clicks, or pops in audio; slow-moving bars or specks in video; or unexplained misbehavior in data systems. The connection of system components to utility AC power unavoidably creates small voltage differences between any two ground points in the system.
GROUND VOLTAGE DIFFERENCES
The voltage differences are generated in two main ways. First is magnetic inductive coupling in the wiring of the branch circuit. Because the same current flows to and from an outlet load in the black and white wires, the magnetic fields created around them are of equal but opposite polarity. The magnetic fields completely cancel at points precisely equidistant from the two wires. However, if a third conductor is closer to one of them than the other, the cancellation is only partial, and the resulting magnetic field induces a voltage in the third. This is the essential principle of transformer operation. Because the spacing between the wires inside the conduit is not exact, significant voltage can be induced in the safety ground wire (or the conduit itself) when load current flows in the circuit.
The second way that voltage differences are generated is as a result of AC power leakage current flow in ground and signal wiring. In all real equipment, capacitances (shown as CL and CN in Fig. 1) exist between each of the power line connections and the equipment chassis. These are the unavoidable parasitic interwinding capacitances of the equipment’s power transformer. If the equipment contains internal power line RFI/EMI filters, the capacitances are larger. Because current must flow to change the voltage across any capacitor, harmlessly small but significant AC current flows from power line to chassis inside each piece of equipment. This current is usually called leakage current.
Because ground conductors have resistance, these currents can create small but significant voltage differences between grounds at different locations in the system. Likewise, cumulative leakage currents from all equipment with 3-prong AC plugs will flow in the building safety ground wiring, causing ground voltage differences between outlets. This interoutlet ground noise generally increases with physical distance between outlets and commonly reaches 1V or more. Generally, voltage is higher between two branch circuits and higher yet if any equipment is also connected to an “alien” ground, such as a CATV feed.
Any wire, including a signal cable, that connects the chassis of two devices thus forms a ground loop. Loop current can reach 100 mA or more. For equipment with a 2-prong AC plug, leakage current can’t flow to ground through its power cord. Instead, it will seek a path to ground through the signal cables. In UL-approved equipment, leakage current is limited to 0.75 mA, which causes an unpleasant but harmless tingle if it flows through your body.
Tiny leakage currents and ground voltage differences are a fact of life. Noise problems occur only when they are allowed to couple into the signal path. Some technicians have a strong urge to short out these voltages with massive copper conductors, but the results are most often disappointing. Others hold a fanciful notion that noise can be eliminated by skillfully directing it to a “quiet” earth ground, where it will disappear forever. Actually, because the earth has considerable resistance, earth grounds are not at 0V, with respect to each other or any other mystical or absolute reference point.
Unbalanced interfaces are susceptible to coupling because leakage current flows in the grounded signal conductor (generally the shield), which causes a voltage drop that adds noise directly to the signal. Less than 0.001V (1 mV) causes serious noise. This common-impedance coupling can become severe between two pieces of safety-grounded equipment. The ground noise in the building safety-ground wiring is effectively forced across the ends of the unbalanced cable shield, making noise larger than the reference-level signal in many situations.
Balanced interfaces, though theoretically immune, can fall victim to noise coupling, too. Equipment design shortcomings, such as the insidious “pin 1 problem” or balanced inputs with poor common-mode rejection, are still common. In addition, some shielded twisted-pair signal cables exhibit a form of magnetic noise coupling called shield-current induced noise.
The most effective approach to solving system noise problems is to locate and safely eliminate the real problem: coupling. When systems are free of noise-coupling problems, special power treatment is rarely necessary. If a system is disconnected from AC power, hum and noise disappear. One could easily jump to the conclusion that all noises are brought in with the power and that the utility company or the building wiring is to blame. Because this idea has overwhelming intuitive appeal, devices to cleanse and purify AC power (which will be covered in a future column) and new wiring schemes are often installed without hesitation. However, they don’t always help.
With conventional wiring, AC outlets on any branch circuit are wired daisy-chain style to the breaker panel as shown in the diagram, sharing a single safety ground. Isolated, or technical, grounding separately routes all three wires from each outlet back to a subpanel or the main breaker panel. Safety grounding is provided only through a separate insulated green wire that makes no connection to the J-box or conduit using a special — usually orange in color — insulated ground or IG outlet.
Especially in large commercial buildings, conduit may contact (at unknown locations and often intermittently) water pipes, gas pipes, air ducts, or structural steel that carry noisy currents. This scheme stops those currents from entering the safety ground circuit. It also prevents the ground noise caused by leakage current and inductive coupling from other devices on an ordinary daisy-chained branch circuit, because each outlet essentially becomes a separate branch circuit. Often simply adding a new branch circuit used exclusively for an audio/video system can be as effective yet far less expensive than implementing an isolated ground system.
Isolated ground must be connected to neutral to serve its safety purpose. Code requires that neutral and safety ground be bonded only at “the power service disconnecting means,” which usually is the main breaker panel. Serious system noise problems can occur when another neutral-to-ground connection exists elsewhere in the building wiring or an outlet is miswired with safety ground and neutral swapped. Special tests must be used to find such problems. Technicians who believe that earth grounds somehow absorb noise have an urge to add more ground rods or even use them as quieter substitutes for existing safety ground connections, which is desperate and dangerous thinking. Beefing up a conventional ground system with massive copper conductors is costly, especially if installation labor and downtime are considered. Generally, the improvements pale in comparison with the results of separate branch circuits or isolated grounding.
Never defeat the purpose of a safety-ground connection by using a 3-to-2-prong plug adapter to solve a noise problem. Signal interconnection cables will also carry lethal voltages throughout a system if that ground lifted device ever fails.
Bill Whitlock is president of Jensen Transformers. He has designed audio and video circuits systems for 30 years. He can be contacted by e-mail at email@example.com.
Ralph Morrison and Warren Lewis, Grounding and Shielding in Facilities, 1st ed., John Wiley and Sons, 1990.
Neil Muncy, “Noise Susceptibility in Analog and Digital Signal Processing Systems,” Journal of the Audio Engineering Society, June 1995, p. 435-453.
Bill Whitlock, A New Balanced Input Circuit for Maximum Common-Mode Rejection in Real-World Environments, Audio Engineering Society 101st Convention, 1996, preprint No. 4372.
Bill Whitlock, “Ground Rods: Magical Noise Sinks?” S&VC, November 2000.
Bill Whitlock, Handbook for Sound Engineers, 3rd ed., Focal Press, 2001, Glen Ballou, ed., chapter 32, “Grounding.”
Bill Whitlock, Hum and Buzz in Unbalanced Interconnect Systems, Application Note AN004, Jensen Transformers, 1996.
Bill Whitlock, “Neutral Ground,” S&VC, June 2001.
Bill Whitlock, “Never Assume Anything,” S&VC, August 2000.