Meeting Review, January 20, 2004
Brown described several
simple principles that can be used to prevent radio frequency entry into
microphones and microphone preamplifiers.
He showed the results of research that he conducted over the last few
years to identify causes of the problem and the effectiveness of the
solutions. Mr. Brown studied the
problem first in controlled laboratory conditions, and then verified his
observations in field trials in the
Broadcast radio signals
and cellular telephones are the most common sources of interference, but
fixed and hand-held two-way radio transmitters are also potential
sources. In tests conducted in the
Key elements of Brown's research are the extension of the work of Neil Muncy, published in the June 1995 JAES, that identified two mechanisms for the coupling of audio frequency interference into audio systems. They were 1) the mis-termination of shields within equipment (the "pin 1 problem"), and 2) non-uniform magnetic coupling of shield current in balanced audio cables to the two signal conductors, a mechanism he called shield-current-induced noise (SCIN). Brown's research, published during 2003 in a series of four AES papers, shows that these mechanisms are also principal causes of RF interference to audio systems to at least 1 GHz.
Brown explained that RF interference enters equipment due to five principal design defects, ranked in importance from most to least. 1) the pin 1 problem; 2) Insufficient differential-mode filtering of input and output wiring; 3) Insufficient common-mode filtering of input and output wiring; 4) Insufficient shielding of the equipment itself, typically an unshielded enclosure; 5) Insufficient filtering of power and control wiring. Brown's work has so far focused on the first two of these mechanisms.
Muncy's "pin 1 problem" is simply another name for common impedance coupling. Two steps are needed to avoid it. First, within equipment, each cable shield (pin 1 of an XLR connector) must have a very short, low impedance path to the shielding enclosure of the audio equipment. While not a component of the pin 1 problem, a connection to the outside of the enclosure is best, both because skin effect keeps RF on outer surface of the enclosure, and also because the connection, however short, cannot act as an antenna inside the enclosure.
A short connection alone is not a solution - if the shield is also connected to circuit common, the voltage drop produced by RF shield currents will be added to the signal, detected (rectified) by signal circuitry, and heard as interference. Thus the second component of the solution is that circuit common must be connected to the shielding enclosure, not to the shield. Brown showed photographs of the connector construction of many popular and expensive microphones where this principle was either neglected or poorly implemented.
The construction of the audio cable can have a large impact on the RF signal that the equipment sees. Brown often “accidentally” referred to audio cables as antennas, reminding attendees of their dual nature. Using an RF generator to impose an RF current through the shield, Brown measured shield-current-induced noise (SCIN) in more than 20 cables up to 4 MHz, and some to 300 MHz. His data showed that RF interference coupled by shield-current-induced noise (SCIN) in foil/drain shielded cables was typically 20-30 dB greater than with braid-shielded cables below about 4 MHz, but that foil shields are superior above about 30 MHz. His research suggests that a foil/braid shield comparable to that commonly used for VHF/UHF antenna distribution can provide the best performance at all frequencies.
Brown also showed examples of new XLR-style cable connectors that are being developed by two companies to control RF interference. The designs are based on work by Brown and members of the AES Standards Committee. The cable-mounted connectors have a concentric capacitive termination of the shield to the connector shell, a dc connection of the shield to pin 1, and a ferrite bead around the pin 1 connection.
Connecting the cable shield directly to both pin 1 and the shell would help equipment with a pin 1 problem, but would add noise if the connector were used with a grounded wiring panel in a building. The concentric capacitive connector avoids this dilemma. Brown showed this connector greatly reduces interference in all of the products tested, provided good contact is made to the shell. The results were the same both in the lab and in the field.
Another prototype connector is intended for use with in microphones as the output connector. It connects pin 1 to the microphone case via a very short strap and places ferrite around all three pins. No test data was available for the latter connector.
Brown played a recording
made on a DAT recorder that had very poor VHF immunity, made in a downtown
The second essential step in controlling RF interference is to block the radio frequency signals carried on the signal pair from entering the equipment. Audio equipment is often designed with excessive bandwidth, allowing AM broadcast signals coupled onto the signal pair (by SCIN) to be amplified and detected. Brown urges limiting bandwidths to the minimum required for good phase response, and certainly no more than 100 kHz. Filtering with simple bypass capacitors is sufficient for interference below 10 MHz, but additional small capacitors and series chokes are needed to extend immunity to VHF and UHF. Brown noted that carefully selected ferrite beads could provide the inductance.
Brown showed the results of RF susceptibility measurements of over 50 microphones, microphone preamplifiers, and mixers. By using appropriately designed RF networks, he was able to selectively inject current into the shield path or into the audio path. His measurements confirmed that microphones and input equipment built with poor grounding techniques were very effective as radio receivers. The worst microphones had 80 dB more audio interference than the best. The price and prestige of the microphone often had no relation to its RF immunity. Several microphones that had poor immunity were modified by Brown to create excellent immunity by applying the principles described above.
Finally, Mr. Brown showed how hand-held VHF and UHF transmitters and cell phones can be used for a quick check of susceptibility to interference. He noted that the legal use of such transmitters is limited, and must conform to FCC rules. The localized nature of the transmitted radio field can help the operator locate the point where RF enters into an audio system. Brown explained that audio cable is very lossy at VHF and UHF, so the locally generated RF signal is greatly attenuated after traveling a wavelength or two through audio cables.
Many of those present at this meeting work in the broadcast industry or for audio equipment manufacturers. They found Mr. Brown’s informative talk particularly relevant to their work.