PEASE PORRIDGE

cording chain (Chesky Records, New Yo rk City), microphone to A-D con­ WHAT'S ALL THIS verter (justto the input, even we don't like a room full of tubes), to 800-watt triode OTL (output transformerless) tube amplifiers.We enjoy a very good SPEAKER CABLE STUFF, reputation for our ability to get the most fr om the equipment. I must also add that I am a recording engineer and a professional musician with 25 ANYHOW? years of performance experience. It's easy to poke fun at the high end ack in July of 1995, Ken that you scoff at have tried exactly as Kenneth Lundgren does in the July Lundgren of Bloomingdale, that, and have failed. 24, 1995 issue of ED, because what is B Ill. had some comments We are retailers and hi-fi is by na­ discernible only by ear may not be about some manufacturers' ture, a fashion and fad business. If it measurable with knobs and dials. The claims for their speaker cables.1 wasn't so, you'd still see amplifiers art of audio reproduction can only, in "There is a special digital cable which with wooden cases. But as in any the final analysis, be evaluated by ear imparts an 'analog-like warmth' to business, some things withstand the (which like a musician's ear or a tech­ digital signals." He made several test of time, like tuxedos, and some nician's eye, takes time and effort to other skeptical and disparaging com­ things don't, like leisure suits. So train). I assert that neither you nor ments about cables that cost hundreds time will tell. But, I tell you there is your skeptical following have actually and thousands of dollars-and I one hell of a wire business out there. listened, which is what true high­ agreed with him that most of these We can't all be deaf! end audio is all about. If you were to claims are absurd. On the other hand, I have heard do so, in time as your ear learns the A few weeks later, I got a letter many types of speaker cables and subtle differences in things such as from a person who appeared to be a they all sound different. I swear! soundstaging, instrument place­ strong supporter of high-end audio, Now you have to consider my back­ ment and timbre, and bass extension George Kaye. Af­ ground of having the trained ear of and speed, you too would be scratch­ ter I wrote him, I a working New Yo rk jazz bassist, ing your head and saying, "Who'd received the fol­ and I've been building amplifiers have thunk it!" By the way, there are lowing cover letter since I was 10 years old. I've made a many well-trained audio engineers and a revised ver­ lot of mistakes and have learned who do things "by the book," and sion of his original from them. So I have 30 years of ear fortunately they provide us with letter. and amplifier training. Also, some many types of equipment that can be people just hear it naturally without improved with modifications. "B ob: much ear training. Usually your debunking of a ques­ In response to I'll tell you I may be fo oled one day tionable theory is excellent, as you one of your letters and be absolutely right on another. demonstrated with the Taguchi BOB PEASE about speaker ca­ The cognitive ear skills are not always Method, and is done with a step-by­ OBTAINED A bles, I want you to at the fo re when it comes to listening, step methodical approach. However, I BSEE FROM MIT know that it makes so I make any critical judgments over am astounded that you attack some­ IN 1961 AND IS absolutely no dif­ time, days or weeks. thing outside your field of expertise STAFF ference what cable I just sent you a rewrite of my first on such an emotional level, and aban­ SCIENTIST AT I use. It's what letter. I hope you don't take offense don the scientific method. I have no NATIONAL gives you the most at my claiming that you are not an problem if you can mount an argu­ SEMICONDUCT­ pleasure that is im­ expert in this fi eld, but hell, it makes ment that trashes something in audio, OR CORP., portant, and I for good reading. but it has to stand up to the finaltest, SANTA CLARA, want people to be GEORGE KAYE" which is the listener's acceptance of CALIF. happy with their the experience. hi-fi. So if you can "D ear Bob: Anyone can get music out of a hand­ doctor up $10 worth of lamp cord with I enjoy reading your column, but I ful of components. The music fr om an a little C and L so that it sounds indis­ must stand up for the audio high-end early Vi ctrola was touted as realistic, tinguishable from a run of any high­ which has come under scathing attack undistinguishable fr om the original. end wire, let me know and believe me, by you and several of your readers. Our experience has proven this claim we'll make much $$, and I'll buy you We design, modify, and manufacture has been exaggerated over time. your next VW clutch. Unfortunately, vacuum tube audio gear. We've built Quality is measured in the domain of many of the people who sell the cable everything fromthe complete tube re- taste. The listening experience is rela-

ELECTRONIC DESI GN/MAY 13, 1996 I PEASE PORRIDGE

tive in terms of quality, and what is is a transducer. The type of capacitor, what harm does it do? Why does it pleasurable or realistic to one, may the tightness of winding, the dielectric make dial twisters mad, anyway? sound bad or false to another. You may and physical deformation properties The truly ineffective fads (yes, find a particular brand of scotch the of the dielectric, resonance, and damp­ there are some charlatans who sell * finest for years and then, as someone ing, all make the difference between a floobydust) die quickly, but others did for me, you are introduced to some violin sounding like a violin, or a fu zz­ open questions that spawn new areas 18-year-old single malt, and suddenly tone electric guitar-whether it is of development, such as digital jitter your allegiance is torn, or maybe you played ppp or fff. Run a square wave reduction (yeah, you can hear that, can't tell the difference. from an amp into any cap of about l /-lF too), power supply improvements I can't go into all the technical and it will sing!2 Polarize a cap with, (WaIt Jung, where are you?), etc. I things, that at first glance, may seem say, 50 V, through a 1-MQ resistor and recommend reading an issue or two of insignificantbut are really important, hook it up to the input of an amplifier Audio Amateur, Glass Audio, Stereo­ so here are a few items that we in or scope. Tap the cap with a pencil. phile, etc. audio ponder and actually research: Thump-thump! The whole art of audio is about ex­ What is the mechanism at the mo­ There are two camps of audio de­ panding the choice and ability to intro­ lecular level that governs how wire vices-tubes and transistors. While duce (or not) anything that wasn't conducts low-level signals, and is it solid-state equipment has made big there in the performance. How close dynamic? What really goes on in 02- strides in the last decade, I have yet are we toward that objective? I'd say free copper wire vs. regular wire to hear what is considered the best we're still a long ways from making with higher oxidation? How pure a solid-state amp outperform many of you believe that there really is a drum conductor is copper? Silver has dif­ the less-than-best tube amplifiers. set in your living room. Think about fe rent electrical properties. Audio Why? One of my theories is that tubes for a moment: 5 drums, 10 vibrating hookups with silver sound different. act as little echo plates adding a rever­ heads, and 4 cymbals, each aimed in a Is different better? Sometimes yes, berant warmth to the music and the different direction, perceived through sometimes no, but different is al­ more tubes the better. 2 loudspeakers. The world of audio for ways different. There are lots of other theories, pleasure (as opposed to function, like Did ' you ever start your car with such as even/odd harmonic genera­ your phone or PA system) is an ART jumper cables only to see them stiffen tion, of an FM-type distortion in a sili­ that balances what you see on your up under the magnetic fields of the con device-ask RF expert Dick Se­ test bench, with what you hear. Any­ current flow? Some of the energy in­ querra about this one. Sufficeit to say one who has been down that path will tended fro the speaker goes into mov­ this debate goes on. The question is, attest to that. ing the two speaker wires instead of "Which sounds better?" Only you can It is understandable that those the speaker, or is lost in the wire's make the call for you. who live and die by measurements dielectric absorption. Current peaks Does an amplifier with 0.0003% are skeptical because they've never of 50+ amps are not uncommon in low­ THD sound better than an amplifier played in this area, and the reported impedance 1-2Q speakers. The dc re­ with 0.25% THD? Is high negative audible results will challenge com­ sistance of speaker wire audibly alters feedback better? Or none at all? Or monly-held engineering notions. the ability of the amplifierto damp the somewhere in between? Don't get me wrong, most of them motion of the speaker cone. The skin Then there is the "perfect" digital are explainable (eventually). So you effect, starting to take effect around format that was rammed down the can pass judgment, but it will be 10 kHz will change the timbre of the public's throat by the record industry valid for you only after you've hon­ treble. Different wires with different that saw huge profits in re-releasing estly put your UNUSED test instru­ resistance, capacitance, dielectric, software with minimal production ment, your EAR, into service. And, and microphonic characteristics will costs. How can you definethe shape of by the way, people do ABX compari­ sound different. a waveform at say 10 kHz with 4 son tests, however, it's more compli­ Another place to look is in the lowly points? The ear hears these subtle cated than you think because switch­ RCA connector, and the often-over­ phase distortions, my friend, and says, ing will degrade the audio signal, looked solder joint. Oxidation and cold "This ain't quite right!" oxidized contacts and all that, old solder joints cause diode-like junc­ But what are we talking about any­ chap. Enjoy your VW. tions with threshold voltages. Low­ way? Are we sending men to the GEORGE KAYE level information such as the rever­ moon? Saving lives? No. Just enter­ George Kaye Audio Labs. beration in a hall may live below the tainment, both for the listeners and Ne w Yo rk, N. Y." threshold of such a connection and the designers. The emotions stir upon * Note, "Floobydust" just means mis­ may get squashed or greatly attenu­ listening to a work of art. Yo ur hearing cellaneous, not weird junk. I don't ated. You end up turning the music up is connected to your emotions and know anyone that sells floobydust.­ louder because your ear is searching your physical being. ENJOY­ RAP for information and auditory clues MENT! !! There is a tinker's pleasure that are there in life. in a fieldor hobby that doesn't require Well, as soon as I read all ofthis, my Every capacitor is microphonic. It a PhD to get decent results. Besides, "emotions began' to stir," and it was

ELECTRONIC DESI GN/MAY 13, 1996 PEASE PORRIDGE

not art causing that. I do not think Mr. argue that if they have the same char­ Mr. Kaye can buy any cables he wants Kaye should call me "my friend" be­ acteristics, they are likely to sound to. He can spend his money anywhere cause I am NOT his friend. the same. And there are a LOT of he wants to. He can counsel his custom­ Takingthe last topic first,the relays cables that have about the same basic ers to buy any equipment he wants to. in ABX boxes do not have to cause characteristics-and they probably And if the customers decide to buy it, distortion because of "oxidized con­ sound alike-despite esoteric claims. it's none of my business. It is not my tacts." Silver contacts are used for But Mr. Kaye wants to argue that primary task to prevent every starry­ high-current paths because the wip­ " ... differentis always different." As he eyed yuppy fr om being separated from ing action of the relays breaks is a professional salesman and pro­ his $$$. through the oxide, every time you op­ moter of high-end audio, I am not sur­ But when Mr. Kaye argues that he erate the relay. For low-level signals, prised to hear him say that. But my can tell subtle differences in cables gold relay contacts are used specifi­ rebuttal is to respond: "Different is that are really trivially different, I cally because gold does NOT oxidize. different, ONLY IF the difference have to be skeptical. When Mr. Kaye So don't try to confuse me with bad makes a difference." You can say that says I should trust him-I won't. science, old chap. If you think you can you like superior speaker cables, but One other reader, who often col­ hear the "distortion" caused by a relay if you can't tell the difference-if you laborates on high-tech, high-end audio placed in series with a speaker wire, cannot hear any significant differ­ projects wrote in to observe, "Bob, I'm sure some people woulq be happy ence-then how can you argue? You you don't have any credibility on high­ to set up a comparison test. could say that silver wires are "differ­ fidelityaudio." (For those of you unfamiliar with ent" than the best copper wires. But, It is literally true that I am not a ABX testing, it's a switch box that lets if there is no measurable difference, person who says, "I can hear that you perform comparisons on two and no audible · difference-and the these well-burned-in interconnect ca­ items-amplifiers, speakers, speaker difference can be computed to cause bles, connected between my tuner and cables, etc. If you push button "A," much less than 0.1 dB of difference­ my amplifier, sound much better, and you are connected to, for example, maybe the difference is not really a have much better clarity than any old speaker cable A. Pushing button "B," difference ... cables." If I don't go around saying connects you to speaker cable B. And When Mr. Kaye argues, "But what preposterous, incredible things like pushing button "X," connects you to a harm does it do?"-that's where we that, then you might say I don't have random choice of either cable A or part company. I am prepared to con­ any "credibility." cable B. See "What's All This Hoax cede that some audio components may But if you acknowledge that I am Stuff, Anyhow?" in the April 4, 1994 indeed have audible advantages. So if capable of suggesting a reasonable ex­ issue for more details.) a salesman of high-end Audio Equip­ periment so that any person can fig­ ' I do not want to get into any argu­ ment sells to any customer, whatever ure out for himself if the interconnect ment about vacuum-tube amplifiers he wants to buy, that's one thing. But cables are capable of making any audi­ versus solid-state amplifiers. If I if a salesman counsels his customer ble difference-then maybe I do have can't hear the difference and Mr. that one set of speaker cables pro­ some "credibility. " Kaye can, that doesn't bother me at vides "better soundstage depth," and I mean, do you really think that in­ all. Hey, I have tried using my ears. another one has more "transparency," stalling three feet of super-fancy in­ They are adequate for ordinary lis­ well, that's his problem. He can sell terconnect cable between a tuner and tening. I can hear what I need to rich yuppies the Emperor's new un­ an amplifier can cause any significant, hear. But I cannot hear and appreci­ derwear, or the Emperor's new san­ audible improvement or difference? If ate every little nuance. Hey, I have dals, or the Emperor's new speaker one cable has 36 pF and another one tin ears. And I am neither proud nor cables. He can sell them anything he had 50 pF, and if the tuner's audio ashamed about that. It's just a fact. wants to. He can sell them silver output impedance was 600 ohms, that I probably ran too many lawn mow­ speaker cables and silver ac line cords would cause a change of frequency ers in the last 40 years. (yes, people do sell such stuff) that response as big as from 7.3 MHz vs. 5.3 If Mr. Kaye finds that an amplifier cost many thousands of dollars. He MHz for f3 dB. The difference in phase with 0.25% distortion sounds better has to live with his conscience-that's shift at 730 kHz might be 2.2 de­ than one with fancy 0.0003% specs, I his problem, not mine. grees. The phase difference at 73 have no doubt he may be right. I am I won't say that the regulators or kHz might be as big as 0.22 degrees. perfectly prepared to agree with Mr. amplifiersor converters I recommend The phase difference at 20 kHz Kaye that our new CD standards are to my customers do not have some might be as big as 0.06 degrees. imperfect. I am only going to argue subtle or hard-to-measure charac­ Do you think anybody can hear that with Mr. Kaye about wires and cables. teristics that are quite good, but difference on an audio channel? Do you I am prepared to agree that when mostly we sell our parts on their mer­ really think an interconnect cable be­ speaker cables have significantly dif­ its, in terms of tested and guaranteed tween your tuner and audio amplifier ferent characteristics such as L, or C, specs. We think our customers get can add significant smoothness or clar­ or series resistance, they may sound their money's worth. I can live with ity? Can you hear an improvement after different. However, I am prepared to my conscience. the cables have been ''burned in?" Can

ELECTRONIC DESIGN/MAY 13, 1996 I PEASE PORRIDGE you hear the molecules in the copper, relay contacts of the ABX box did not sounds good is right, and nobody can after they have been realigned by cause any significant error, as they talk you out of it. Mr. Doug Self has burn-in and stress? Can you hear the were transparent compared to straight­ done a very fine job of defining the absence of corrosion in 22-k gold­ wire bypasses, and the box did not reason why scientific types are not plated connectors? cause any error. supposed to be able to criticize and Do you think that ANYBODY can These tests were written up by Mr. disagree with subj ective observa­ hear such a difference? If anybody N ousaine and published in Sound & tions. I will simply quote Mr. Self's says they can hear a difference, would Vi sion, a magazine published in Can­ best definition of the "Subjectivist ,, you like to be able to challenge them ada. You can buy this magazine by Manifesto :7_"Objective measure­ to an ABX test? sending a check for $2.95 (D.S.) to ments of an amplifier's performance Do you argue, "What harm does it Sound & Vision, 99 Atlantic Avenue, are unimportant compared with the do?" when some self-proclaimed "ex­ Suite 302� Toronto, Ontario M6K 3J8, subjective impression gained by lis­ pert" tries to sell some fancy intercon­ Canada. Postage to Canada is 46 cents tening tests. Where the two contra­ nect cables, when the only audible dif­ per 0.49 ounces in 1996. Ask for a copy dict, the objective results may be dis­ ference is-the salesman's fatuous of Volume 11, Number 3. They'll know missed .. .It therefore follows that deg­ words of praise? that there is a lot of interest in Mr. radation effects exist in amplifiers NOW, I have mentioned in my oid N ousaine's story. that are unknown to orthodox engi­ columns3 that "Audio Experts" claim If any of you are tired of being told neering science, and that are not re­ they can hear the difference between about. the superiority of expensive vealed by the usual objective good speaker cables and "better" speaker cables, you'll want to read tests ... Considerable latitude may ones. One guy claimed he can even this taste-test. Maybe the Emperor's therefore be employed in suggesting hear a difference if there are splices4 new sportcoat is transparent, after hypothetical mechanisms of audio im­ in the cables. Other guys say that the all! Note, Mr. Kaye thought I should pairment, such as mysterious capaci­ expensive cables provide better subscribe to Audio Amateur, or Glass tor shortcomings and subtle cable de­ "presence." Advertisements claim Audio, or Stereophile. I'd rather fects, without reference to the plausi­ that these expensive specialty cables spend my money on a subscription to bility of the defect, or the gathering of provide better "sound-staging." Bet­ Sound & Vision. objective evidence of any kind." ter "harmonic integrity." More pre­ Mr. N ousaine indicated that at the In other words, if a guy says he can cise "imaging." More fluid highs. beginning of the tests, some Audio HEAR the difference between two "Less skin effect." I shall not comment Experts tried to duck the question of, amplifiers or two cables, we are not further on these claims. would they like to be involved in the permitted to contradict him, nor to try But an engineer, Tom N ousaine in comparison testing. They tried to ex­ to measure any differences or simi­ Illinois, set up some comparisons be­ plain, "It sounds better with better larities. And if we want to put that tween some high-end speaker cables cable. It must make a difference be­ person to a test of proving that he can costing $418 and $900, versus some cause everybody says it does." After really hear a difference-we can't 16-gauge zip cord, at 18 cents per foot. publication of these tests, it may be a force him to do the test. And if the test He used an ABX box for ease in con­ little bit harder for these guys to keep results show that he cannot really ducting double-blind comparisons. on arguing this... hear any difference, we are not per­ The audiophiles were confident that Mr. N ousaine has since run many mitted to doubt his excuses. they could clearly hear the superiority other tests-dozens of ABX tests. Explanations such as, "these oxy­ between their uniquely excellent So far, no one has been able to hear gen-free wires have better distortion speaker cables, in their own home any significant differences between at the sub-molecular level, greatly su­ audio systems, versus the cheap wire. cheap wire and fancy speaker cables. perior to the noise and distortion in Mt er many careful tests-double­ In fa ct, Mr. N ousaine has laid out a ordinary wires which is manifested by blind tests-the audiophile experts challenge, with no takers yet: If you tiny micro-bursts of noise just below had to admit they couldn't hear any can hear the difference between the threshold of audibility or measur­ difference. "The differences become your fancy speaker cables and cheap ability" must be accepted, and cannot vanishingly smalL" "They sound iden­ wire, you win $1000. But if you can't be questioned. Ye ah, sure. ticaL" They were devastated. "Fol­ hear the difference, he gets to keep I still think that Fred Davis' article lowing the session, Mr. B. was the expensive speaker cables. Mr. on "Speaker Cables, Te sting for Audi­ ,,5 crushed, likening the experience to N ousaine has also run many tests bility is one of the best technical that of finding his wife in bed with that show similar results for "audio explanations. And I really like Fred's another man." After 10 sets of trials, interconnect" cables. No one can proposal, that the broad array of 32 there was no (statistically valid) in­ hear the difference between the best flat twisted pairs of cable, such as dication that the Golden Ears could and the cheapest ones, notwith­ Spectra-Strip 843-138-2601-064, with tell whether the cables they were standing all the fancy claims. all the wires appropriately "paral­ listening to were the $900 cables, or I do not want to get drawn into a leled" is one of the best speaker ca­ the $9 cables. long argument about "subjectivism," bles. Because when you have 32 ca­ All of the parties agreed that the which is the premise that whatever ble pairs, and each one has about 75 ELECTRONIC DESIGN/MAY 13, 1996 PEASE PORRID"GE

ohms, the impedance of all 32 together much of the error is due to cable Santa Clara, CA 95052-8090 is down near 2 or 3 ohms. inductance, and how much to Skin This does give very low phase shift Effect. In the real world, it may be P .S. I recall that a few years ago the from the amplifier to the speaker. It hard to tell. Bay Area Skeptics (Bay Area Skep­ gives good damping. It also gives low I wonder if Mr. Kaye can actually tics, 177723 Buti Park Court, Castro skin effect. And it is inexpensive. And show us some characteristics of "oxi­ Valley, CA 94546), along with an it is flat enough to place under rugs, dation with diode-like threshold volt­ Audio Engineering Society, ran a which you cannot do with big, fat, ex­ ages" on old RCA connectors, as he lecture and ABX comparison in pensive cables. mentioned in his letter. Does that Berkeley. In that case, a couple of Some people might say, "But such mean that some small signals just do guys who claimed to have good ears a cable must have a lot of capaci­ not get through? I'd love to play were able to hear some consistent tance." Well, IF you go around leav­ around with one of those. But I cannot difference between some 22-gauge ing your speaker cable disconnected say that cold-soldered joints do not lamp cord and some good cables. from any speaker, yes, it might act ever exist. However, most engineers admit that as if a big capacitance-perhaps as I sent George Kaye a copy of Mr. cables that have about the same R, big as 0.016 �F for 20 feet. BUT if N ousaine's article. His comments L, and C tend to sound alike. Cables you connect it to a 4-ohm speaker, it were quite limited. But he did admit that have markedly different R, L, just looks like a few ohms. There that while he can hear the difference or C can have significantlydiff erent ain't no pF-just as when you termi­ between different speaker cables on phase shifts, and might sound differ­ nate a long 50-ohm cable to its char­ some days, he may not be able to hear ent. So maybe we should propose an acteristic impedance, it looks resis­ the differences on any particular day. ABX test of a capacitive speaker ca­ tive. Conversely, such a cable has a In other words, differences that he ble, not versus a plain piece of lamp lot less inductance. In general, it has enthusiastically considers clearly sig­ cord, but versus a piece of lamp cord lower impedance, and is better nificant on one day, he may not be able with 0.01 �F on its output. Or com­ suited to the task of coupling from an to distinguish on another day. So pare an inductive speaker cable set amplifier to a low-impedance load. there's no point in trying to give Mr. to a piece of lamp cord with a few Ordinary pairs of speaker wires are Kaye an ABX test. (So if you are going hundred microhenries in series with typically closer to 70 or 90 ohms, in to challenge a friend to an ABX test, its output. ..Hey, let's have some their characteristic impedance. don't show him the Nousaine article fun !!-RAP Other people argue that the induc­ first, because he may try to weasel his tance of a speaker can cause a lot of way out.) References: phase shift. Even when a speaker ca­ I asked George Kaye to tell me ble's impedance is low and ideal, it can what kind of speaker cables he liked. 1. "Bob's Mailbox," Electronic Design, cause phase errors when driving a He demurred, saying that he liked July 24, 1995 p.105. load-a real loudspeaker-if that and used many different kinds of 2. I have not recently put together a 1 looks like L in series with R, such as speaker cables. In other words, he �F capacitor made of NPO ceramic or 100 �H in series with 4 ohms. This refused to be trapped into saying silver mica, but I bet if you drive a would appear to be a problem. there was one kind of cable he pre­ 1-kHz square wave into these capaci­ But, if you put a series R-C net­ ferred. If there were such a cable, I tors, they will not "sing." It is not true work parallel to the R-L load, such would surely try to emulate its char­ that "all capacitors are microphonic." as 4 ohms in series with a 6 �F My­ acteristics, using perhaps some I t is just that most capacitors that Mr. lar-that can cause the overall im­ Spectra-Strip cable with some Kaye is familiar with, are micro­ pedance to be nearly flat, and can cut added R and L or C. But if he won't phonic. down on the phase shift caused by name a type, I can't do that. Maybe 3. "What's All This Hoax Stuff, Any­

speaker impedance. This works best it's just as well that Mr. Kaye does how? ," Robert Pease, Electronic De­ when L/R = R x C. This is called a not live just down the street from sign, April 4, 1994. Boucherot cell. me, so I can't challenge him to an 4. "What's All This Splicing Stuff, 6 One article argued that skin effect audio test. I can't literally challenge Anyhow? ," Robert Pease, Electronic can cause a nasty little tail of error if you readers to an audio test. But I Design, Feb 1991. you apply through the speaker cable can challenge you to think. Mr. Kaye, 5. "Speaker Cables-Testing for to the speaker a brief transient-such I'm not so sure about. Audibility," Fred E. Davis, Audio, as a tone burst of a few cycles of a sine July 1993, p. 34-43. wave. These transient errors may be All for now. I Comments invited! 6. "The Essex Echo," Malcolm Omar more audible than a mere phase error RAP I Robert A. Pease I Engineer Hawkesford, Stereophile, October on a continuous tone sine wave. And 1995. p. 53-64. they are fairly easy to measure with a" Address: 7. "Science vs. SUbjectivism in Audio 'scope. However, nobody said these Mail Stop D2597A Engineering," Douglas Self, Elec­ differences, are SIGNIFICANT. And National Semiconductor tronics & Wireless World, July 1988, the analysis never indicated how P.O. Box 58090 p. 692-696.

ELECTRONIC DESIGN/MAY 13, 1996 ♦ Verso Filler Page ♦ ENGINEERING REPORTS

Effects of Cable, Loudspeaker, and Amplifier Interactions*

FRED E. DAVIS

Engineering Consultant, Hamden, CT 065 J 7, USA

Loudspeaker cables are among the least understood yet mandatory components of an audio system. How cables work and interact with loudspeaker and amplifier is often based more on presumption and speculation than on fact. The literature on loudspeaker cable behavior and effects is minimal. Measurements were made with 12 cables covering a variety of geometries, gauges, and types. The measured data indicate distinct differences among the cables as frequency-dependent impedance, subtle response variations with loudspeakers, and reactance interactions between amplifier, cable, and loudspeaker. In some cases the effects of the amplifier overwhelm the cable's effects. Mathematical models that provide insight into the interaction mechanisms were constructed and compared to the measured data.

are much too long compared to the length of the cables); o INTRODUCTION phase shift and dispersion effects are too small to be A variety of specialty loudspeaker cables can be audible (typically less than 0.3 deg/m at 20 kHz, and found advertised in almost any audio magazine from the differences of less than 60 ns/m for most cables between last 10 years. All promise the same result-better 100 Hz and 10 kHz); and the skin effect has only a sound-yet they span the gamut of electrical charac­ small effect on heavy conductors (skin depth in copper teristics, geometries, and materials. How loudspeaker at 20 kHz is 0.5 mm). cables work is often based more on presumption and It is no secret that loudspeakers offer a complex load speculation than on fact. Few articles are published to the amplifier [12]- [13]. While an isolated loud­ exploring the behavior of these mandatory components speaker is predominantly inductive, the complex in journals [1] and popular magazines [2]- [6]. Debates impedance of most loudspeaker systems with multiple continue on computer network news groups on audio drivers and passive crossover networks exhibits both [7]. "White papers" available from manufacturers (but negative and positive phase angles at given frequencies, otherwise unpublished) are frequently more marketing indicating capacitive reactance as well as inductive l than science [8]- [11]. reactance. Otala and Huttunen [13] show that given Using a simplistic view of how loudspeakers and complex waveforms, commercial loudspeakers require cables work, conventional wisdom would suggest that up to 6.6 times more current than an 8-0 resistor for since loudspeakers exhibit a low impedance (nominally the same waveform, suggesting a dynamic impedance 4-80), then the cable should have even lower resist­ as low as 1.2 O. ance. As a result, "monster" cables were introduced. The ideal loudspeaker cable should transfer all audio Then a more complex view of cables emerged, sug­ frequencies into any loudspeaker load with fiat voltage gesting that loudspeaker cables would perform better response. Real cables will always show some loss due with less capacitance or more inductance, or the skin to resistance, but better cables will both minimize this effect, phase shift, and dispersion were veiling high loss and still transfer all frequencies unscathed. The frequencies, or they behaved like transmission lines. acoustical result will depend on many factors, but the These factors are the essence of 'high-end' cables. electrical interaction of loudspeaker, amplifier, and Greiner addressed some of these issues in his papers cable forms an essential foundation. This engineering [1]- [3]. In short, he proposes that loudspeaker cables are not transmission lines (audio frequency wavelengths 1 Some issues of Audio that illustrate Nyquist plots of loudspeakers' complex impedances are vo!. 74, p. lOO (\ 990

* Manuscript received 1990 August 21; revised 1990 De­ Nov. ); vo!. 74, p. 94 (\990 Aug. ); vo!. 73, p. III (\989 cember 24. June); vo!. 73, pp. 88, \08 (\989 Sept. ).

J. Audio Eng. Soc., Vol. 39, No. 6, 1991 June 461 DAVIS ENGINEERING REPORTS report examines the mechanisms for this interaction exactly the same rate of speed, " hence the name constant and shows how it can affect the response of the system. velocity transmission (CVT). Two groups of three Vari­ Lay bundles form the two main conductors, with a 1 SAMPLE OF CABLES TESTED coaxial cable connected in an unknown fashion (due to potting compound) inside a proprietary coupler at The sample of cables gathered for this test represents the amplifierend. At $419 per meter, the most expensive a variety of commonly and uncommonly available wire. cable tested. Equivalent to 12 AWG. Most of the samples were 3.1 m in length. Some are 9) Kimber 8LPC. Very similar to sample 5. except very expensive (over $419 per meter), others cheap eight independent wires, woven in a flat cable, Teflon ($1.91 per meter), and some are not loudspeaker cables insulation. Each individual wire is equivalent to 19 at all. This is not an exhaustive examination of every AWG, and is composed of seven strands of variable loudspeaker cable available. The following is a brief gauge from 31 AWG to 24 AWG. Equivalent to 10 description of each type with sample numbers as they AWG. appear in Figs. 1- 3. They are presented in vrder of 10) Kimber 4PR. An unusual cable made from eight ascending resistance per meter. When known, the or­ independent wires of 23 AWG (7*31) braided together, ganization of the strands is shown in parentheses as PVC insulation. Equivalent to 14 AWG. (quantity*gauge). Unspecified gauges were estimated 11) Spectra-Strip 843-191-2811-036 Ribbon Cable. from conductor diameter and resistance. Abbreviated 191-036. Made of 36 wires of 28 AWG 1) Levinson H F 1OC. Many very small copper strands (7*36), arranged in a flat ribbon; intended for digital in two parallel conductors (each about 6.4 mm in di­ interconnections. The least expensive cable tested at ameter) spaced about 12.7 mm apart (between centers $1.91 per meter. Equivalent to about 15 AWG. of the conductors). Approximately 3 AWG. Extremely 12) Belden 19123. 18 AWG (41*34) "zip " (lamp) flexible for such a heavy conductor. cord. Brown PVC insulation, parallel construction. 2) Auto Jumper Cables. Literally from the garage. Two thick parallel (9. 5-mm diameter) conductors of 2 ELECTRICAL PARAMETERS OF CABLE approximately 7 A WG (19*20). SAMPLES 3) Krell "The Path." Independent wires of about 15.9-mm diameter, each of complex layer construction. The standard electrical parameters of the cables were The conductor is 4.8 mm in diameter, the remainder measured with an ESI model 252 impedance meter and ...... - ....1� '""...-1 1 "'rh ., ... 1+ .,...... ,h ,.. .. , ...... � ..... "C� 1 is insulation. It has several groups of tightly twisted iJUIIIlQJI'-,vU...... L"-' I Ill....- lll�...... IL,"\UIL;., ..... O-lv ;''HI\''/VV11...... III .I .0'''1...... 1- very thin enameled wires wound in helices around 3. It is resistance, capacitance, and inductance that heavier enameled wires. (This construction is similar will decide the performance of the cable, since exotic to Music Interface Technologies' "Vari-Lay" and materials and layer geometries can only affect these Monster Cable "Time Correct"). All conductors are fundamental characteristics. soldered together at each end with heavy, crimped ter­ Cable resistance in milliohms per meter is shown in minations. Approximately 5 AWG. They are labeled Fig. 1 (remember that this includes the resistance of "transconductant speaker cable." both conductors). Resistance is not a major factor in 4) AudioQuest Green "Litz." Six conductors (ap­ cables of reasonable length. Based on resistance alone, proximately 10 AWG) of many small enameled copper it would require about 23.4 m of 18 AWG cable to wires, lightly twisted over a stranded plastic core, al­ show -1 dBV drop with an 8-D load. 12 AWG seems together about 12.7 mm in diameter. Equivalent to about more than adequate even for demanding systems, high 6 AWG. 5) Kimber 16LPC. These are 16 independent wires, Resistance (milliohms per meter) woven together in a flat cable, Teflon insulation. Each 50 �------�------�----�------� individual wire is equivalent to 19 AWG, and is com­ e posed of sev n strands of variable gauge from 31 AWG 40 to 24 AWG. Equivalent to 7 AWG. 6) Spectra-Strip 843-138-2601-064 Ribbon Cable. 30 Abbreviated 138-064. Made of 32 twisted pairs of 26 AWG wire (7*34), arranged in a flat ribbon. Intended for high-speed differential digital data transmission. 20 Equivalent to about 8 AWG. 7) Belden 9718. Belden's 12 AWG (65*30) loud­ 10 speaker wire with clear PVC insulation and parallel construction, like "zip" cord (sample 12). 345 78 10 11 12 8) Music Interface Technologies' CVT. A large 18- Sample Number mm diameter cable using MIT's Vari-Lay construction (multiple conductors of different gauge and length). Fig. I. Cable resistance in milliohms per meter. I-HFIOC; 2-jumper; 3-Krell; 4-Litz; 5-16LPC; 6-138-064; The manufacturer claims this will permit "all frequen­ 7-9718; 8-CVT; 9-8LPC; 10-4PR; 11-191-036; cies to travel through a given length of MIT cable at 12-19123.

462 J. Audio Eng. Soc., Vol. 39, No. 6, 1991 June ENGINEERING REPORTS CABLE, LOUDSPEAKER, AND AMPLIFIER INTERACTIONS power levels, and reasonable lengths. The maximum with frequency which permits flatter voltage response. current for 12 AWG wire with PVC insulation in an A current of 1 A at a given frequency will cause a ambient temperature of 30°C, allowing for a 50°C tem­ voltage difference equivalent to the magnitude of the perature rise, is 36 A. This seems fine for audio ap­ cable's impedance in ohms at that frequency. For this plications, since 36 A into 8 0 is greater than 7 kW test, a resistive load of 1.00 (with approximately 0.06- rms (1.8 kW rms into 2 0). j.LH inductance) was driven at a current of 1.0 A rms Fig. 2 shows the cable capacitance. As expected, at 12 frequencies between 30 Hz and 20 kHz. All mea­ flatcables show the highest capacitance (samples 6 and surements were made with a Fluke 8050A digital volt­ 11), multiconductor cables less (samples 4, 5, 8, 9, meter and waveforms monitored on a Tektronix 2215 and 10), and two-conductor cables the least (samples oscilloscope. The amount of current was determined 1, 2, 3, 7, and 12). by driving the amplifier until the voltage across the Fig. 3 shows the cable inductance. Cables with only load was 1.000 V rms at each frequency and for every two separated conductors show the highest inductance, cable, thus removing frequency response variations from while most multiwire cables show the lowest inductance. signal source, attenuator, and amplifier. The voltage An exception is Music Interface Technologies' CVT difference from the output of the amplifier to the load due to its construction. was then measured and recorded, and the impedance calculated. The results of these measurements are shown in Figs. 3 CABLE IMPEDANCE VERSUS FREQUENCY 4 and 5 as cable impedance versus frequency, where The impedance of a cable across the audio spectrum the value of impedance reflects the contribution of both shows the influence of reactive and skin effects. Better conductors. Cables with the most constant impedance cables will have a low impedance that remains constant were the flat cables with higher capacitance (Fig. 4, 138-064; Fig. 5, 191-036). Other multiconductor cables such as Kimber 16LPC and AudioQuest Green Litz Capacitance (nanafarads per meter) -----�------'-----'------2.5 � ----, (Fig. 4, 16LPC and Litz) and the lighter gauge Kimber 8LPC and 4PR (Fig. 5, 8LPC and 4PR) display a small impedance rise. Of the two conductor cables tested,

1.5

0.5

0.4

0.3

0.2

2 3 4 5 67 8 10 11 12 0.1 Sample Number o � 10 100 1000 10000 Fig. 2. Cable capacitance in nanofarads per meter. I-HFJOC; Frequency (hertz) 2-jumper; 3-Krell; 4-Litz; 5-16LPC; 6-138-064;

7-9718; 8-CVT; 9-8LPC; 10-4PR; 11-191-036; 10 HF10C -+- 2. Jumper ---.- 3. Krell

12-19123. --e- 4. Litz --*- 5. 16LPC -+- 6. 138-064

Fig. 4. Cable impedance versus frequency for cable samples

Inductance (microhenries per meter) 1-6. 1.6,-----'------'------'------,

0.5

0.4

0.3

0.2

0.1

O r-�-��Tn�-r_�rrrITn--+_1_TTTTnr-� 10 100 1000 10000 2 4 5 67 8 10 11 12 Frequency (hertz) Sample Number - - 7.9718 -+- 8. r::vr -- 9. 8LPC Fig. 3. Cable inductance in microhenrys per meter. 1- --e- 10. 4PR -x- 11. 191 -036 � 12.19123 HF10C; 2-jumper; 3-Krell; 4-Litz; 5-16LPC; 6- 138-064; 7-9718; 8-CVT; 9-8LPC; JO-4PR; 11- Fig. 5. Cable impedance versus frequency for cable samples 191-036; 12-19123. 7-12.

J. Audio Eng. Soc., Vol. 39, No. 6, 1991 June 463 DAVIS ENGINEERING REPORTS

12 AWG wires (Fig. 5,9718 and CVT) performed the factored out,showing only the cable and loudspeaker best, since both heavier and lighter gauges showed interactions. In Sec. 7, amplifier effects are included greater high-frequency impedance. The complex layer with loudspeaker and cable effects for a total system construction of the CVT cable has duplicated almost response. exactly the impedance characteristics of the 12 AWG Belden 9718 (Fig. 5, CVT and 9718; coefficient of 5 CABLE RESPONSE WITH LOUDSPEAKER correlation = 0.997). LOADS The effect of inductive reactance in this sample of cables is far more significant than the skin effect. For Obviously, a loudspeaker can only perform to the example,3.1 m of the largest diameter cable sampled, quality of the electrical input to its terminals, so the Levinson HFI0C (sample 1), will show a 3.42 times best cable will show the flattest frequency response increase in resistance at 20 kHz due to the skin effect, but the inductive reactance will be 9.8 times greater than resistance at that frequency. When driving 8 n at 20 kHz through 3.1 m, the skin effect alone would 40 produce a drop of -0.044 dBV relative to 20 Hz,while the combined reactance and skin effects would produce 20 a drop of -0.43 dBV. 0 Higher cable capacitance will tend to reduce the -20 combined reactive component of the cable, thus low­ -40 ering cable impedance at high frequencies and im­ proving the high-frequency response. This effect is -60 10 100 1000 10000 contrary to the popular belief that high frequencies Frequency (hertz) will be attenuated more with higher cable capacitance -B [5],[8]. Such conclusions are drawn from a cable model .. * .. Impedance Phase consisting of series resistance and shunt capacitance; Fig. 6. Impedance and phase response of loudspeaker A. but no series inductance. Spectra-Strip 138-064 (sample Note that frequencies are sampled and lines connecting data 6) showed the highest capacitance (6.847 nF for 3.1 points do not reflect valid data. m),lowest inductance, and flattest cable impedance. \Vcll designed amplifiersare not affected by this amount of capacitance, but some amplifiers may become un­ stable. 25 40

20 20 4 TEST LOUDSPEAKER AND AMPLIFIER CHARACTERISTICS

10 The impedance and phase characteristics of loud­ 5 speakers A and B used in these tests are shown in Figs. OL-__l�-LLL� __ ��-LLUJV ___ L��LU�U 6 and 7,measured at the same frequencies used in the 10 100 1000 10000 cable impedance test. Please note that the lines con­ Frequency (hertz) necting the data points in these graphs are intended to .. * .. Impedance ---B Phase simplify reading the plot and do not reflect valid data between the sampled frequencies. Loudspeaker A is a Fig. 7. Impedance and phase response of loudspeaker B. three-way design with an acoustic suspension woofer, Note that frequencies are sampled and lines connecting data points do not reflect valid data. three dome midrange drivers, and three dome tweeter drivers. It exhibits mostly capacitive reactance (negative phase angle) at the frequencies sampled between 127 Damping Fador Response (dBV) 300 ...... 0.4 Hz and 12 kHz, with its lowest impedance of 4.8 n above 8 kHz. Loudspeaker B is a two-way system with 250 0.2 a bass reflex enclosure and dome tweeter. It shows much more inductive reactance (positive phase angle) 150 -0.2

than loudspeaker A around I kHz, and a capacitive 100 -0.4

reactance peak around 8 kHz. Its lowest impedance is 50 -0.6 5.8 n around 500 Hz. Amplifier frequency response and damping factor 10 100 1000 10000 Frequency (hertz) are shown in Fig. 8. Amplifier A exhibits more sig­

nificant frequency response variationsand a large drop - Amplifier A Freq. -+- Amplifier 8 Freq. in damping factor above 1 kHz. Amplifier B has a flat ---a- Amplifier A Damping ...... - Amplifier B Damping frequency response and a high,almost linear damping Fig. 8. Frequency response and damping factor for amplifiers factor. In Secs. 5 and 6, the effect of the amplifier is A and B.

464 J. Audio Eng. Soc., Vol. 39, No. 6, 1991 June ENGINEERING REPORTS CABLE, LOUDSPEAKER, AND AMPLIFIER INTERACTIONS despite loudspeaker impedance or phase angle. The loudspeaker and cable resistive and reactive components cable electrical response was measured using two com­ together. The cable is mode led at each frequency as a mercial loudspeakers as a load. resistance in series with an inductive reactance using A constant amplifier output of 1 V (0.00 dBV) was the measured values of resistance and inductance. The used at each frequency to remove any variations due skin effect was calculated and applied to the resistance to amplifier or signal source. The amplitude of the where appropriate. The capacitive component of the voltage at the loudspeaker terminals was measured in cable is too small to have much influence at audible dBV and recorded. frequencies, and is thus omitted from the model. The The low-inductance multiconductor cables show the loudspeaker is modeled at each frequency as a resistance most linear response (Fig. 9, Litz, 16LPC, and 138- in series with a reactance that can be either inductive 064; Fig. 10,8LPC, 4PR, and 191-036). Also note the or capacitive. The expression for the cable response at relatively flat response of the 12 AWG cable with both the loudspeaker terminals for a given frequency is loudspeakers (Figs. 10 and 11,9718) when compared to other two-wire cables (Figs. 9 and 11, HFlOC and VR� + Ps VaC! Krell). Another common effect is the high-frequency ) (R + ( Y w Rs)2 + Xw ± Xs)2 loss with the higher inductance two-conductor cables. Fig. 9 also shows the interaction of a cable's inductive where reactance with loudspeaker A's capacitive reactance where the level rises above 0 dB V in the 1-kHz to 10- voltage at loudspeaker terminals at fre­ kHz region. At this point the loudspeaker terminal quency f voltage has exceeded the amplifier'soutput. The cause VaC!) voltage at amplifier output at frequency f of this will become apparent with the loudspeaker cable cable resistance, including skin effect,at model introduced in Sec. 6. frequency f Four cables representing a variety of types were tested cable inducti ve reactance at frequency f with loudspeaker B (Fig. 11). Loudspeaker B shows loudspeaker resistance inductive reactance and low impedance between 300 loudspeaker reactance at frequency f, in­ Hz and 3 kHz and the response dips. When the reactance ductive (+) or capacitive (-). of loudspeaker B becomes capacitive around 8 kHz, it The response in dBV was found by taking the log- shows the same rise with the more inductive cables (HF lOC and Krell). Response (dBV per 3.1 meters) Ql ...... 6 LOUDSPEAKER CABLE MODEL

Expressions for transmission lines (such as charac­ teristic impedance, impedance matching, reflections) do not fit audio applications, since the cable lengths involved are minute fractions of the shortest audio wavelength (about 16 km at 20 kHz in copper). This is discussed thoroughly in Greiner [1]- [3]. 10 100 1000 10000 Frequency (hertz) Therefore, cable and loudspeaker should be treated - as lumped-circuit elements. The cable response model -+- 7. 9718 - 8. cvr -- 9. BlPC in this engineering report is simple and is based on the 10. 4PR --""*- 11. 191-036 � 12. 19123 ratio of the vector sum of the loudspeaker's resistive Fig. 10. Measured cable response with loudspeaker A for and reactive components to the vector sum of both cable samples 7-12.

Response (dBV per 3.1 meters) Response (dBV per 3.1 meters) 0.1 ::: : : : :: : Ql ......

-0.1

-0.2

-0.3

-0.4

-0.5 L----'---''--LJ--LJJJ1--_L-.L....I.-Ll--L.J...Ll...-_-'---'---'LLLLl.il.----l 10 100 1000 10000 Frequency (hertz) 10 100 1000 10000 Frequency (hertz) --t- 1. HF10C -- 2. Jumper ----*- J. Krell

-4- 6. 138-064 4. Utz ----Jo(--- 5. 16LPC -I- 1. HF10C - 3. Krell -- 6. 138-064 -El-- 7. 9718

Fig. 9. Measured cable response with loudspeaker A for cable Fig. 11. Measured cable response with loudspeaker B for samples 1-6. cable samples I, 3, 6, and 7.

J. Audio Eng. Soc., Vol. 39, No. 6, 1991 June 465 DAVIS ENGINEERING REPORTS arithm of the ratio of the response at a test frequency phase shifts due to capacitive loads from causing in­ and the l-kHz response, stability in the feedback loop. Both amplifiers A and B include this network. Obviously, this inductance is Vs(f) in series with the cable inductance,and in some cases 20 10 g Vs(l kHz) can exceed the cable inductance. The damping factor of an amplifier can also shape Three different styles of cables are modeled and the frequency response. The damping factor (and the compared to measured values in Fig. 12. The model output impedance of the amplifier)is controlled by the gives a very good approximation to the measured re­ frequency-dependent loop gain of the amplifier,the de­ sponses (coefficient of correlation = 0.999,0.948, and gree of negative feedback,the impedance of the output 0.997 for HF1OC, 16LPC, and 19123, respectively). devices, and any other components in series between The results are for the full 3 .I-m length of the cable the amplifieroutput and the output terminals. The am­ since they are not directly scalable to other lengths. plifier output voltage will be lower where the damping The rise above 0 dBV in the measured responses factor is lower or where the load impedance is lower. occurs when the combined magnitude of the impedance An amplifier with low damping factor is less able to of loudspeaker and cable (as seen by the amplifier) is control back EMF and reactive effects of the loud­ lower than the loudspeaker's impedance alone. This speaker. results when the reactance of the loudspeaker is ca­ The responses of all cables were tested with the same pacitive and subtracts from the cable's inductive re­ loudspeaker, but using two different amplifiers. Figs. actance. The result is a lower total reactive component, 13 and 14 present the responses of loudspeaker A and which reduces the magnitude of the impedance seen amplifier A,while Figs. 15 and 16 present the responses by the amplifier. Since the amplifier output is held at of loudspeaker A with amplifier B. These graphs il­ a constant voltage for the cable impedance test, the lustrate the combined responses of loudspeaker,cable, current through the loop is higher than the loudspeaker's and amplifier. Immediately obvious is that the response impedanc� alone would require. This higher current of amplifier A overwhelms the individual cable effects results in a voltage across the loudspeaker terminals (Figs. 13 and 14). The damping factor for amplifier A that is higher than the amplifieroutput. Low-inductance and the impedance of loudspeaker A both drop in the cables will provide a more ideal response since cables same frequency range, which exacerbates their inter- whose inductive reactance is much less than the loud­ speaker's capacitive reactance wiii reduce this "hump" effect and present little more than the loudspeaker's Response (dBV per 3.1 meters) complex impedance to the amplifier as a load. When 0,1 " ...." . the effective impedance of cable and loudspeaker is lower,it should not prove difficultfor a well-designed amplifier because the effect is small with short cables (approximately 0.6% for the worst case in these tests, sample 2, auto jumper cable). The lowest impedance seen by the amplifier and the greatest rise in loudspeaker voltage as a result of this effect occur at resonance, 10 100 1000 10000

= Frequency (hertz) when Xcable - Xspeaker' The impedance will then be limited by the resistive components of both cable and --+- 1. HFl DC mode! ----- 5. 16LPC model ------12. 19123 model loudspeaker. For example,loudspeaker A would require 1. HF1 QC meos. � 5. 16LPC meos. � 12. 19123 meas. just over 12.4 m of Be1 den 9718 cable to provide enough Fig. 12, Modeled and measured response with loudspeaker inductance to achieve resonance at 10 kHz,where the re­ A for cable samples I, 5, and 12. sistance seen by the amplifier would be about 4.84 O.

AMPUFIER EFFECTS Response (dBV per 3,1 meters) 7 0 �=:=±2��#F��� Now that the relationship between loudspeaker and -0,25 cable is better understood, the effects of the amplifier -0.5 will be considered. As seen with the cable model,added -0,75 inductance will cause frequency response deviations -1 due to interactions with the loudspeaker's reactive -1.25 components. Therefore it would be desirable to min­ -1.5 L-�����L-�__ ��iDL-�L-Li���� imize reactive effects from the amplifier as well. Most 10 100 1000 10000 amplifiers include added inductance (typically 0.5-10 Frequency (hertz)

J.LH) paralleled with a resistance (typically 2. 7-27 0) --+- , . HFl QC -- 2. Jumper -¥- 3, Krel1 between the output of the amplifier (generally from the --a---. 4, Litz -)(- 5. 16LPC --+- 6. 138-064

point that negative feedback is taken) and the amplifier's Fig. 13, Complete system response for amplifier A with output terminals. This inductance is added to isolate loudspeaker A, cable samples 1-6.

466 J. Audio Eng. Soc" Vol. 39, No, 6, 1991 June ENGINEERING REPORTS CABLE, LOUDSPEAKER, AND AMPLIFIER INTERACTIONS action. The response with amplifier B (Figs. 15 and flattest response will occur by keeping the reactallce 16) closely resembles the response of the cable and of the amplifier and cable as low as possible. loudspeaker alone (Figs. 9 and 10). The high damping factor of amplifierB maintains better control of reactive 8 CONCLUSIONS effects with the more inductive cables, producing a flatter response (Fig. 15). If loudspeakers were only simple resistance, then The effect of the amplifier can be added to the cable large, low-resistance cables would not be a bad idea. response model by including the additional resistance However, loudspeaker systems exhibit a frequency­ and reactance of the amplifier's output: dependent complex impedance that can interact with the reactive components of amplifier and cable. The Vs(f) best response was obtained with low-inductance cables and an amplifier with low-inductance output and a high, frequency-independent damping factor. These tests have shown that the best way to achieve adequately low resistance and inductance in a cable is where Va(f)' = amplifier voltage at frequency f. by using many independently insulated wires per con­ ductor rather than one large wire. Efforts to reduce the Fig. 17 illustrates the results of this model, using skin effect (such as Litz construction) will help, but amplifier B' s voltage response with loudspeaker A's due more to the reduction of inductance than the re­ impedance and phase (converted to dBV relative to the duction of the skin effect. Inductive reactance is more 1-kHz response as before). The model fits well with significant in large cables than the skin effect. If an the measured data (coefficient of correlation = 1.000, amplifier does not disagree, larger capacitance in a 0.997, and 0.999 for HFI0C, 16LPC, and 19123, re­ cable is not significant since this component is com­ spectively). Because the model is very simple and am­ paratively small and reduces amplifier and cable in­ plifier dynamic responses are more complex, it does ductive reactance effects. not fitas closely with all amplifiers,especially the ones The best performance was measured with the multi­ that have a more complex output reactance (which may conductor cables Spectra-Strip 138-064, Kimber include capacitive effects). The model infers that the 16LPC,and AudioQuest Litz. Smaller multiconductor

Response (dBV per 3.1 meters) Response (dBV per 3.1 meters) ° f===S::::��ffiF� -0.25

-0.5

-0.75

-1

-1.25

- 1 5 . L----'.---'.-L-LLLLLL----'.--'...... L..LLJ..LLL- ---'.---1...... L..L..L.Ll.l.l----l

10 100 1000 10000 10 100 1000 10000 Frequency (hertz) Frequency (hertz)

-+- 7. 9718 - 8. cvr """""*'- 9. 8LPC --f- 7. 9718 - 8. cvr --- 9. 8LPC

---G---- 10. 4PR --¥- 11. 191-036 --+--- 12. 19123 --a-- 10. 4PR -)(-- 11. 191-036 --+- 12. 19123

Fig. 14. Complete system response for amplifier A with Fig. 16. Complete system response for amplifier B with loudspeaker A, cable samples 7 -12. loudspeaker A, cable samples 7-12.

Response (dBV per 3.1 meters) Response (dBV per 3.1 meters)

-0.2: F=��:rmF:::��:;::mF�������

-0.5

-0.75

-1

-1.25

-1 .5

10 100 1000 10000 10 100 1000 10000 L----'.--'...... L..L..LJ..LLL---'--'...... L...L..L.l..liL----'.--..l...... L...L..L.l..liL---'Frequency (hertz) Frequency (hertz)

-+- 1. HF10C -- 2. Jumper ----¥- 3. Krell -+- 1. HF10C model -- 5. 16LPC model ----*- 12. 19123 model 6. 138-064 --e- 4. Litz � 1. Hr, DC meos. ---+i- 5. 16LPC meos. � 12. 19123 meos.

Fig. 15. Complete system response for amplifier B with Fig. 17. Model of complete system response for amplifier B loudspeaker A, cable samples 1-6. with loudspeaker A.

J. Audio Eng. Soc., Vol. 39, No. 6, 1991 June 467 DAVIS ENGINEERING REPORTS cables such as Kimber 8LPC,Kimber 4PR,and Spectra­ (1980 May). Strip 191-036 also performed well. [2] R. A. Greiner, "Cables and the Amp/Speaker Of the two-wire cables, 12 AWG provided the best Interface," Audio, vol. 73, pp. 46-53 (1989 Aug.). performance with reactive loads, while both smaller [3] R. A. Greiner, "Another Look at Speaker Ca­ and larger gauges (3-7 AWG and 18 AWG) showed bles," BAS Speaker, vol. 7, no. 3, pp. 1-4 appended greater high-frequency drop and interaction with ca­ (1978 Dec.); addenda, vol. 7, no. 6, pp. 6-7 (1979 pacitive reactance in a load. 12 AWG seems more than Mar.). adequate, even for demanding systems, high power [4] C. Ward,1. Thompson and M. Harling,"Speaker levels, and reasonable lengths. Cables Compared," BAS Speaker, vol. 8, no. 7, pp. The effects of 3.1-m cables are subtle,so many sit­ 25-29 (1980 Apr.). uations may not warrant the use of special cables. Low­ [5] R. Warren, "Getting Wired," Stereo Rev., vol. inductance cables will provide the best performance 55,pp. 75-79 (1990 June). when driving reactive loads,especially with amplifiers [6] D. Olsher,"Cable Bound," Stereophile, vol. 11, having low damping factor, and when flat response is pp. 107- 118 (1988 July). critical,when long cable lengths are required,or when [7] B. Jones, "Speaker Cable Electrical Tests," perfection is sought. Though not as linear as flat cables, ACSnet/UUCP: [email protected] 1990; a series 12 AWG wire works well and exceeds the high-fre­ of discussions and rebuttals can be found referencing quency performance of other two-conductor cables ([email protected]),newsgroup: rec.audio. tested. By the way,keep the auto jumper cables in the [8] D. Salz, "The White Paper on Audio Cables," garage! Straight Wire Inc.,Hollywood, FL (1988). [9] B. Brisson, "How Phase Shift in Audio Cables 9 ACKNOWLEDGMENT Influences Musical Waveforms," Musical Interface Technologies,Auburn, CA (undated). The author would like to thank Robert A. Pease for [10] "Cable Design, Theory Versus Empirical his invaluable suggestions on measurement techniques, Reality," AudioQuest,San Clemente,CA (1990). Dave Sales fVi his amplifier and exotic cables, and [11] "Sumiko Reports: OCOS-The Formula," Su­ Brian Converse,Dr. Robert Milstein,and Dianne Davis miko,Berkeley, CA (1989). for their support and suggestions. [12] J. H. Johnson, "Power Amplifiers and the Loudspeaker Load," Audio, vol. 61,pp. 32-40 (1977 A ...... \ 10 REFERENCES �Ub')' [13] M. Otala and P. Huttunen, "Peak Current Re­ [I] R. A. Greiner, "Amplifier- Loudspeaker Inter­ quirement of Commercial Loudspeaker Systems," 1. facing," 1. Audio Eng. Soc. , vol. 28, pp. 310-315 Audio Eng. Soc., vol. 35, pp. 455-462 (1987 June).

THE AUTHOR

Fred E. Davis is currently a consulting electronics controls, portable instrumentation and tools, computer engineer, and previously has been director of engi­ hardware design and software, and consumer elec­ neering for machine vision and optical data storage tronics. Besides a passion for music, his interests in­ companies. He has worked with biomedical electronics, clude photography, books (especially old and rare sci­ audio, recording studio and video electronics, industrial ence books), and film.

468 J. Audio Eng. Soc., Vol. 39, No. 6, 1991 June