Keeping an audio system (or any electronic system) free of noise and hum requires at least these three important and interrelated factors: a good grounding system, well-designed equipment, and proper interconnections.
Grounding is whole topic unto itself, but one design goal of a good technical ground system should be the minimization of shield currents carried on interconnect cables.
Well-designed audio equipment should not have any pin-1 problems. Any cable shield current should not be allowed to enter the equipment chassis and find its way onto the circuit board's ground.
Good interconnections are equally important, as is the choice of cable.
Even if the pin-1 problem is solved, there can still be gremlins lurking about. Under the wrong circumstances, cable shield currents can induce noise into the signal wires.
A common type of extensively used audio cable is composed of a twisted pair of signal wires surrounded by an aluminum foil shield plus a copper drain wire. This type of cable is fairly easy to install, but it happens to be the worst for shield current induced noise (SCIN).
The drain wire is the culprit. Any shield current present (up to a frequency limit of around 5 MHz, according to research done by Jim Brown), will flow mainly in the drain wire due to its lower resistance compared to the foil shield.
However, the drain wire is not wrapped uniformly around the signal conductors, and that's the problem. The drain wire is wrapped at the same rate as the signal pair and is closer to one of the wires. That means more noise will be induced in the signal wire closest to the drain wire.
In other words, the shield current causes a differential voltage to be produced on the signal pair, and that voltage noise is superimposed on the signal itself.
The type of noise that finds its way to cable shields can originate from any number of sources, such as motors, dimmers and RF interference. SCIN tends to increase with frequency in a fairly linear fashion.
Are there any audio cables that are less susceptible to SCIN? Yes, cables with uniform shields.
Brown evaluated and compared the SCIN performance of a variety of audio cables, including foil/drain, foil/braid, braid and braid/drain shields at frequencies up to 30 MHz. He found that, in general, "the braid shields were superior below about 15 MHz, the foil shields were superior above 30 MHz, and a foil/braid shield was effective throughout the RF spectrum."
Foil/drain shielded cables performed the worst.
These results point out that braid and foil shields have different characteristics when it comes to shielding. While both are effective against electrical interference, braided shields are not as effective as solid conductors for magnetic shielding, especially at higher frequencies when the braid holes become large compared to the wavelength.
But at lower frequencies of magnetic interference, both types of shields are ineffective because of the relative thinness of the shields compared to wavelengths at these frequencies.
What really prevents magnetic interference is the twist in the cable. The more twists per inch (or other unit of length) and the more uniform the twists, the better. This is important to remember as cables are stripped and prepped. The twist should remain up to and as close to the point of connection as possible.
I often see installations where the cable is untwisted back several inches before the connection point. That is not good installation practice, however easier it may be to dress the cables. These are the sections of the cable that are most vulnerable to magnetically induced hum. And here is also the place where the cable may be nearest to a power supply and power transformer--something to watch out for.
Twisted cable is also effective for RF rejection, another point in its favor--but back to shields A twisted-pair cable with a combined foil-and-braid shield would be ideal, but not easy to find.
Braided shielded cable has fallen out of favor for installers because it generally takes more work to wire. And some have been reluctant to use it because of its magnetic shielding properties at higher frequencies. However, its superior performance in reducing the more dominant SCIN could easily outweigh these factors when foil/drain wire-shielded cables just don't work.
Manufacturers should also be aware of the role SCIN plays in systems and provide RF filtering at audio inputs to eliminate a good portion of this noise. Despite marketing claims (probably moreso on the consumer side), bandwidths from DC to light are not needed in audio gear. A roll-off of around 100 kHz or so should be sufficient for good phase response in most equipment.
Systems designers need to consider SCIN and the environment in which the cabling will be installed. Some questions to ask: Is the system located in a place with high RF energy (like near a transmitter or in a major metropolitan area), or near a subway substation? What kind of staging and lighting equipment will be used? What about motors used in the building--HVAC, elevators, etc.? All of these and more could be sources of interference. What kind of ground system is present or being planned? To just blindly wire with the usual foil/drain wire cable without understanding the environment ultimately could cost more in the end in troubleshooting time and possible rewiring.
Some consultants refuse to spec foil/drain wire shielded cable anymore because of the SCIN problem. Some integrators may balk at the use of braided shielded cable because of the potential increased labor costs. Systems design is a balancing act. Designers need to be aware of the potential problems and the likelihood of occurrence, and weigh such factors as the cost of cable, wiring and installation against the cost of commissioning, debugging, and maintenance.
For systems already installed that have annoying noises and hums, start by identifying and correcting all pin-1 problems. Then if noise problems persist, look next for SCIN. Note what type of cable is used, and try a braided-shielded cable to see if the problems go away. If magnetic interference is suspected, check how the cable is prepped at the connection end. Take a systematic approach, correcting one layer of problems at a time.
References for this article include papers written by Jim Brown and Neil Muncy (AES preprints and journals available at www.aes.org ) and the book "Noise Reduction Techniques in Electronic Systems," by Henry W. Ott.
