[Comp.Sci.Dept, Utrecht] Note from archiver<at>cs.uu.nl: This page is part of a big collection of Usenet postings, archived here for your convenience. For matters concerning the content of this page, please contact its author(s); use the source, if all else fails. For matters concerning the archive as a whole, please refer to the archive description or contact the archiver.

Subject: FAQ: rec.audio.* Wire 7/07 (part 8 of 13)

This article was archived around: 15 Jan 2009 06:01:56 GMT

All FAQs in Directory: AudioFAQ
All FAQs posted in: rec.audio.tech, rec.audio.opinion, rec.audio.misc, rec.audio.marketplace
Source: Usenet Version

Archive-name: AudioFAQ/part8 Last-modified: 2007/07/12 Version: 2.17
15.0 Wire More than any other topic, speaker cables and equipment interconnects seem to use up rec.audio.* bandwidth echoing the same theoretical arguments, testimonials, and opinions. Controversy can be stimulating, educational, and also amusing. Please try to keep postings aimed at one of those three goals, and avoid the insults and emotion. Also, try to avoid echoing a common position or principle, as described below. 15.1 Do speaker cables matter? To avoid confusion and repetition, here is some terminology, thanks to Steve Lampen of Belden Wire & Cable Co. A wire is a single conductor made up of one or more conducting elements, but all configured (as in a stranded design) to act as a single conductor. Mostly, this is coated or covered by plastic, rubber, enamel or similar insulators. Groups of wires are called cables. So zip cord is a cable, because it contains more than one insulated conducting element. Coaxial cable is also cable. Cables can introduce noise into the signal, act as a filter (and thus change the frequency response of the system), attenuate the signal (change the amplitude), and provide nonlinearities from oxidized or otherwise poor connections. Nonlinearities can distort the signal which add harmonics. Nonlinearities can also rectify or demodulate higher frequency signals into audible signals. It is quite scientifically conceivable that some cables do cause a difference in sound, because of the differences in DC resistance, interconductor capacitance, and connector attachment alone. The effects of exotic conductor weaving and materials are not so well established. In general, these effects (once we eliminate DC resistance), seem to be small. However, if your system is at least fairly good, then some folks have observed (although not in an objective, double-blind experiment) significant differences in system performance with different cables. The effects are said to be quite system specific; the only real guideline is to try them and see which ones seem to sound better to you, in your system. Roughly speaking, the price ranges for speaker cables is low (under $1/ft), medium (under $6-8/ft), and high (up to $100/ft and more). Experiments can be expensive, so don't commit your money unless the seller has a practical money-back guarantee. In any system or experiment, it is essential that the differences between cables be separated from the differences between connectors. You should have an EXTREMELY solid connection between cable and speaker. Speakers carry high peak currents, so bad connections will create significant artifacts or signal losses at any power level. For example, if the connection has a linear resistance of just 1 ohm, the speaker damping factor may be changed, and the bass may suffer. If the connection contains imperfect metal oxides, then a slightly rectifying junction will block the signal, producing compression, distortion, and other non-linear effects. There are many magazine articles on cables with various perspectives which are worth reading, including: "Speaker Cables: Testing for Audibility" Fred E. Davis Audio, July 1993, pgs. 34-43 15.2 What speaker cables are available and how good are they? There is a wide range of speaker wire available, ranging from 30ga zip cord (~$.10/ft) to exotic wires costing over $300/ft. The material used ranges from copper to oxygen-free copper (OFC) to silver. (There are a bunch of others as well.) Oxygen-free copper is probably NOT any different from common copper in sound. If you hear a difference between two cables, it is not a difference between oxygen-free and common copper. Resistance may be significant for speaker cables. The higher the resistance, the more the cable will affect the sound, all else equal. The resistance characteristic of metals is called resistivity. The resistivity of copper is 1.7 microohm-cm. Silver is very slightly lower, 1.6. Gold is a bit higher, 2.4. Silver and gold are different from copper in other ways than resistivity. Gold does not oxidize in normal environments, so gold contacts will not need periodic cleaning and will not create rectifying junctions. Silver will oxidize, but the oxide of silver is conductive, so oxidized silver will still make good contact. Copper oxide is a bad conductor. Oxidized copper contacts may insulate, may conduct, or may rectify. Copper is a bad material for cable terminals, but this may or may not mean anything for the conductor itself. 15.3 What can I use for budget speaker cables? First, a few words on terminology. Wire is sized by AWG or BS gauge number. Larger numbers represent smaller wire. AWG 40 (also called 40 gauge) is as fine as human hair. AWG 12 is 2mm or .081" diameter. If you reduce the AWG number by 3 (such as from 29 to 26) then the wire cross-sectional area increases by a factor of 2 and the series resistance drops by a factor of 2. Some wire is classified as solid, because it contains one strand per conductor. Other wire is called stranded, because it consists of many strands per conductor. Stranded wire is far more flexible than solid wire. Most wire is made from drawn copper. Some wire is sold that is claimed to be made with a process that produces oxygen-free copper. Oxygen-free copper has a different metallurgical structure than common copper and may or may not conduct current better. Some critical listeners have reported excellent sound from large diameter solid copper wire, such as home wiring "Romex 12-2". At least one expert has said that common 18-gauge solid copper hook-up wire sold by Radio Shack also works very well. Also recommended on a budget is Sound King wire, a 12 gauge oxygen free copper stranded cable. This is available from MCM Electronics for $.39/ft. Scientifically, thinner wire has more resistance than fatter wire, so fatter wire will have less resistance-related effects. Resistance effects can be eliminated by using at least 12 gauge wire, particularly for long runs. Of course, shorter runs are always preferred, because they come much closer to the ideal zero-length wire, with no resistance, no capacitance, no inductance, and no change in signal. 15.4 What can I use for budget speaker connectors? The worst connectors are push-down, or spring terminals. Screw terminals with solid copper wire are much better. Gold-plated binding posts and gold spade lugs are inexpensive by audiophile standards and are extremely stable. Binding posts with spade lugs can be tightened to get a very good mechanical joint, and may offer the lowest electrical resistance of any connector. Gold plated banana plugs and jacks can be very good speaker terminals, depending on construction. Some banana plugs have weak springs and fall out rather than staying in place. Good ones are more expensive than gold spade lugs, however, they also provide a bigger area of contact, and are more convenient when you must frequently reconfigure the system. Banana plugs should be periodically monitored for corrosion and loss of spring tension. Monster offers a banana-plug connector with an expanding center pin that forms an even better connection than common gold banana plugs. At approximately $25 per pair, the Monster banana plugs aren't a budget connector. All else equal, connectors with gold surfaces are better than connectors with any other surface. This is for two reasons. First, gold is extremely inert, meaning that unless gold is exposed to very harsh chemicals or harsh vapors, it will not corrode or oxidize. It will remain a pure, low-resistance conductor. Second, gold is quite soft, so that if a gold-plated connector is squeezed between two metal surfaces, it will deform slightly to fill scratches and voids, giving a very broad, low-resistance contact area. Corrosion of connectors is often a problem. Gold-plated terminals and connectors somewhat avoid this problem; problems with other connectors can be mitigated by unplugging and replugging the connector on a regular basis, cleaning the contact areas with a pencil eraser, or by using a contact enhancer such as Cramolin or Tweek. When you use a contact enhancer, be very sure to follow the directions, and avoid spreading enhancer about your equipment. 15.5 What about interconnects, such as the cable between tuner and amp? Line-level interconnects conduct smaller signals than speaker cables; the typical signal ranges from -2V to +2V (the CD output standard) with currents in the microamps (the corresponding values for speaker cables attached to a largish power amp might be -70V to +70V and currents of many amps). Line-level interconnects can be divided into single-ended (or unbalanced), and balanced interconnects. Home audio is almost always single-ended interconnects. Single-ended interconnects almost always use a form of the RCA connector (or phono plug). RCA plugs form fair to poor connections that degrade with time as corrosion works into the metal-metal contact and as the spring tension of the connectors relax. Gold-plating reduces the effect of corrosion and locking RCA connectors solve most of the mechanical problems. However, these premium phono connectors are rare and expensive. For example, a gold-plated Vampire locking RCA plug costs approximately $23/pair. If RCA connectors weren't a de facto standard, we'd recommend against them. Unbalanced interconnect wires vary in geometry, material and price. Cheaper wires have a single conductor (normally stranded) and a shield and cost $.20-$2/ft. Medium (complexity and price) wires have two conductors (often arranged as a twisted pair) surrounded by a shield and cost from about $3-$20/ft. Exotic wires have all sorts of geometries and materials (such as stranded silver conductors, or ribbon cable braided around a core, or in one extreme case, a tube filled with mercury!). Prices may be as high as $200-$300/ft. Balanced interconnects have three conductors: two for the signal one for ground, and additionally a shield. The standard connector for balanced cable is the ITT/Cannon XLR connector, which is quite good mechanically (they lock). Equivalent connectors are also available from Switchcraft, Neutrik, and other vendors. If you have to run cables longer than 12 feet or 4 meters, the greater noise immunity of balanced interconnects is often a good idea. For this reason, balanced connectors are standard equipment in professional installations such as studios, public address systems, and broadcast stations. There is not much variation in balanced cables. The three brands mentioned above are known to be rugged, high quality and moderately priced. Slightly weaker imported connectors are available, but they aren't dramatically cheaper. Better connectors are also easier to assemble and have a more durable cord strain-relief. For most systems, the most important aspect of a cable are the mechanical reliability of the connectors; in particular, the joint between connector and wire, and the joint between connector and socket. Typically, interconnect cables are short. It is worth getting just the right length; cables often come in .5 meter increments. With quite good systems, some people observe differences in sound between various interconnects. This is quite system-specific and the same advice as given above applies: try several brands. Most good dealers will loan interconnects for home evaluation. In cables where the shield does not carry the signal or ground, the shield is normally only connected to ground at one end. In systems where there are significant differences between ground levels on various components, it may make a difference which way such cables are connected. Typically, the end where the shield is grounded should be at the source of the signal. Often, such cable has arrows on it pointing in the direction of the signal flow. In any case, try both orientations. There are many objective reasons why cables might cause differences in sound by interacting with the drivers in the signal sources as well as by providing non-linear effects in the RCA connector. Most of these effects are again related to interconductor capacitance and resistance, and the quality of the shielding provided by the "shield" conductor. In balanced cables the quality of the "twisted pair" inside the shield is also important. One might note that a shield protects from only capacitively coupled interference, and not from any magnetic field interference. The twisted pair in a balanced line provide some magnetic rejection, as does steel conduit. However, steel conduit has other characteristics which make it undesirable for audio in general. 15.6 What about Phono Interconnects: Phono interconnects are part of the link between a cartridge on a turntable and a preamp (or head amp or receiver). They are a special case of line-level interconnects because the signal is much lower, typically 1 to 50 millivolts. They are also intended to operate into a higher impedance, typically 47K ohms, and form part of the capacitive load for the cartridge. The low signal levels mean that the shielding of the cable, and the presence of a separate drain/shield are more important, as is a good ground. A separate solid ground should come along with the cable as a separate lead co-routed with the cable. In addition, the low signal levels make a good solid connection to and through the connectors MUCH more important, because of the greater sensitivity to low-level nonlinearities. Wire capacitance is often ignored in line-level interconnects; however, in a phono interconnect, it may constitute half of the total capacitive load of the cartridge. Obviously, then, two cables with significantly different capacitances should sound differently. In this sense, the "right" cable for one cartridge may be too low or high in capacitance for another cartridge. For low-impedance cartridges (most moving coil cartridges), the wire must have low resistance to prevent cartridge unloading and frequency-dependent signal loss. In addition, as the signal levels are quite low, shielding is important. Unfortunately, copper shields do not block stray magnetic fields, so in the case of phono cables, careful routing may be even more effective at reducing hum than special wire. 15.7 Is there really a difference in digital interconnects? There are now three kinds of digital interconnects that connect transports to D/A converters: coax, plastic fiber (Toslink) and glass fiber (AT&T ST). In theory, these should sound EXACTLY the same (bits are bits). However, this assumes good circuit design (in particular, the clock recovery circuits of the DAC, and careful consideration of electronic noise) which may be compromised because of cost considerations or ignorance. Note: different signaling schemes are used on plastic and glass fiber. In any case, some people claim to hear a difference; of those who do, most seem to prefer the glass fiber. However, the technology of fast digital data transmission in consumer electronics is evolving very quickly now. Any specific recommendation should be treated with suspicion until the industry matures. 15.8 Can I make very good interconnects myself? Yes. You will need to be the judge of whether or not they are as good as $100 interconnects, but it is easy to make interconnects that are better than the $2.00 set which comes with new equipment. There are two necessary ingredients: two-conductor shielded cable and RCA connectors. There is a lot of debate over what is the best cable, but in general, the lower the capacitance per foot, the better. Choice of insulation is harder. There may be an advantage to polypropylene or teflon over polyester or rubber, but even that is debatable. If you are buying wire from an electronics distributor, some have successfully used Belden 1192A microphone cable. It is rubber insulated, so very flexible. Another recommended cable is Belden 8451. This is a polypropylene cable with foil shield. Finally, consider Belden 89182. This is foamed teflon insulated, so very low capacitance, and foil shielded. If you plan to make a long cable, this low capacitance cable may be the best choice. There is also a variety of RCA connectors available. A good connector would be gold plated and machined to tight tolerances. A poorer connector will not fit as well, will make poorer contact as the connecting surface oxides, and will lose its springiness with use. When wiring the cable to the connector, use one wire for signal, (the tip of the RCA connector) and one wire for ground (the shell or outer conductor of the RCA connector). Some cables use a foil shield which is difficult to solder. These cables typically have a drain wire parallel to the foil which can be used for soldering. Others use a braided shield. Regardless of which type of wire you have, connect the shield or the shield drain wire to ground on only ONE SIDE. This will stop noise picked up by the shield from causing ground noise. It can be a touchy job soldering RCA connectors. Before you use your new cables, check with an ohmmeter or a continuity tester to make sure that you have not accidentally sorted the signal and ground leads together, either with a stray drop of solder or a loose wire strand. 15.9 Is there a standard for wiring balanced XLR-3 cables? Yes. Connect pin 1 to ground/green, pin 2 to white, and pin 3 to black. Herb Hamilton suggests that you remember "George Washington Bridge" and then use the first letter of each word (GWB) to help you remember Green=1, White=2, and Black=3. This same wiring convention works for balanced line level signals and balanced microphone cables. COPYRIGHT NOTICE The information contained here is collectively copyrighted by the authors. The right to reproduce this is hereby given, provided it is copied intact, with the text of sections 1 through 8, inclusive. However, the authors explicitly prohibit selling this document, any of its parts, or any document which contains parts of this document. -- Bob Neidorff; Texas Instruments | Internet: neidorff@ti.com 50 Phillippe Cote St. | Voice : (US) 603-222-8541 Manchester, NH 03101 USA Note: Texas Instruments has openings for Analog and Mixed Signal Design Engineers in Manchester, New Hampshire. If interested, please send resume in confidence to address above.