a quick, cheap speaker pair for the basement · a quick, cheap speaker pair for the basement ... a...

AUGUST 2001US $7.00 Canada $10.00

IT JUST SOUNDS BETTER

A QUICK, CHEAP SPEAKER PAIR FOR THE BASEMENT

www.audioXpress.com

Continuing Audio Electronics, Glass Audio & Speaker Builder

EditorialOut, Out, Damned Dot

Nearly everyone in the North Americanelectronics establishment seems to mecontent to sit, seemingly happily andpermanently, between two stools. Thesituation concerns the unhappy choiceMichael Faraday made in choosing aquantity for that relationship betweentwo layers of metal separated by nearlyany dielectric that we refer to as capaci-tance. We refer to these units as micro-farads, which isn’t exactly the way it iswith capacitance. Faraday’s farad isvery, very large in relationship to thecomponents we normally use.

The microfarad is actually one mil-lionth of a farad: 10-6 farads. Notice theminus sign before the exponent, whichmeans that the number is smaller thanone. We represent the micro with the12th lowercase letter of the Greek alpha-bet, the µ, which is equivalent of our“m.” We do not use the “m” for microbecause it has already been taken bythe next less small mathematical quan-tity, the thousandth, or milli (10-3).Which is why when writing about mi-crofarads we do not designate a capaci-tor as being 40mF. Unless we mean a40,000µF device. We don’t generally dothat since we seem to have this passion-ate devotion to the little Greek µ and itsdecimal derivatives.

At least in North America we do. Wealso have a positive, abhorrent aver-sion for the perfectly respectable math-ematical representation for numberswhich are in the billionths, namely thenano. No, we would rather resort to thecreaky, circumambulating, ditsy busi-ness of slicing up micros with deci-mals. What sense decimals make whentalking about billionths escapes me. Itis either laziness, ignorance, the herdinstinct, or perverseness, take yourpick.

Any quantity of farad slices thatcome in billionths can be designated asnanofarads. These are units which are10-9. The whole extent of decimated

farads in this range, from 0.001µF to0.999µF, is more properly and easilydesignated as 1nF to 999nF. The mostoften used 0.1µF as a bypass for opamps is far more easily written 100nF—same number of characters, and with-out the Greek letter (Alt 0181, on yourcomputer keypad).

Fortunately North American elec-tronics practitioners have made somesmall progress in the last three or fourdecades or so since we have, at least,abandoned “µµF” as a way of designat-ing picofarads. Pretty generally wehave accepted the fact that a trillionthof a farad (10-12) is properly referred toas a pico. Thank goodness we havegraduated from referring to these de-vices as accumulators or condensers,which we did for decades (shades ofthe Leyden jar).

Another major problem with this lazydecimation habit is the fragility of thelittle dot required for decimals. Some,who are obviously afraid the little deci-mal will get lost in the nineteenth copyof the schematic, put another zero inthe lineup just to protect the puny littlething: as in 0.001µF. If you love redun-

dancy, help yourself. However, isn’t 1nFbetter? I submit it is not just better, it isa lot better.

It’s easier to think of these generallyused designators as three sets of threes.Larger ⇐ µ n p ⇒ Smaller

000 | 000| 000In this simple way, you never need

use a dratted decimal—that is, unlesswe can all agree on some of the largerunits of electrolytic capacitance forsmoothing low voltages such as a milli-farad: 1mF meaning 1000µF, but thatmight confuse everyone. However, weshould be able to manage a 1mF, realiz-ing that the lowercase “m” alwaysmeans milli or thousandths. Then, ofcourse, we could move on to 1cF, thecentifarad, or 100,000µF. This changewould be very helpful on power supplyschematics where the space around capacitors is usually limited. I’ll behappy if we can all agree to just startusing the nano regularly. Perhaps thisreluctance is part and parcel of the U.S.fear of that “foreign conspiracy” knownas the metric system, which is stan-dard in most of the civilized world

(to page 95)

2 audioXpress 8/01 www.audioXpress.com

QUANTIFICATION FOR ELECTRONICSLESS THAN ONE

TRILLIONTHS BILLIONTHS MILLIONTHS THOUSANDTHS HUNDREDTHS TENTHSPICO 10-12 NANO 10-9 MICRO 10-6 MILLI 10-3 CENTI 10-2 DECI 10-1

ampere µA mAfarad pF nF µF mFhenry µH mHhertzohmvolt µV mVwatt µW mW

ONE AND LARGERONES TENS 101 HUNDREDS 102 KILO 103 MEGA 106 GIGA 109

AFHHz kHz MHz GHzΩ kΩ MΩV kV MVW kW MW

Note: The discoverers of these units are lowercased referenced quantities. The letter representing them is capitalized,except for Georg Ohm whose unit is designated by the Greek letter Ω, omega.

C O N T E N T SVOLUME 32 NUMBER 8 AUGUST 2001

F E AT U R E S

SLIM JIMSHere’s an interesting exercise−for your muscles, as well as yourconstruction skills−to build a simple pair of full-range speakers.By Ken Ketler ................................................................................ 8A HIGH-QUALITY TWO-CHANNEL CHIP AMPA bargain chip−the LM3875−is at the heart of this quality IC-basedpower amp project−complete with circuit board layout−from thisaudio veteran.By Charles Hansen ....................................................................12THE CIRCLOTRON AMPIntrigued by the Circlotron amplifier as a solution to driving low-mutriodes, this author takes a closer look at this little-used circuit.By John L. Stewart.................................................................... 26BOX STUFFING––A FISTFUL OF FUZZTwo industry experts unsnarl the confusion surrounding cabinetstuffing.By Ron Horowitz and John Schaffer.................................... 30ALL ABOUT WIREThis audio veteran provides some practical information aboutchoosing the various wire types for building projects.By Peter Millett ................................................................34ELECTROSTATIC ODYSSEYYou’ll be proud to prominently display these elegant-looking,smooth-sounding speakers in your living room.By Gary Gendel..................................................................40A START-UP DELAY FOR VINTAGE AMPLIFIERSTry this “kinder, gentler” amp delay to extend the life of your components.By Alexander Rubli Kaiser ..................................................50RECREATING A 1948 CONSOLE AMP The secrets of this classic sound are revealed. By Larry Lisle ....................................................................54audioXpress (US ISSN 0004-7546) is published monthly, at $29.95 per year, $53.95 for two years.Canada add $12 per year; overseas rates $59.95 per year, $110 for two years; by Audio Amateur Inc.,Edward T. Dell, Jr., President, at 305 Union St., PO Box 876, Peterborough, NH 03458-0876. Period-icals postage paid at Peterborough, NH, and additional mailing offices.

POSTMASTER: Send address changes to: audioXpress, PO Box 876, Peterborough, NH 03458-0876.

page 12


page 34

R E V I E WS

ROKSAN CASPIAN CD PLAYERDoes this single-disc, professionally designed CD player from theUK deliver the goods?By Charles Hansen........................................................................ 58LISTENING CRITIQUEBy Betty Jane and Richard Honeycutt ..........................................60

MUSIC HALL MMF-2.1 AND GOLDRING CARTRIDGEThe LP sound is alive with this two-speed turntable.By Charles Hansen ........................................................................61

SOVTEK 6550 AND 2A3 PLATINUM MATCHEDTUBESTesting Sovtek tubes from New Sensor.By Charles Hansen........................................................................ 64

SONY CD PLAYERSHow do the CDP-XE400 and XE-500 single-play units from thisJapanese electronics giant stack up?By Gary Galo ................................................................................ 66

audioXpress August 2001 5

LEGAL NOTICEEach design published in audioXpress is theintellectual property of its author and is of-fered to readers for their personal use only.Any commercial use of such ideas or de-signs without prior written permission is aninfringement of the copyright protection ofthe work of each contributing author.

SUBSCRIPTION/CUSTOMERSERVICE INQUIRIESA one year subscription costs $29.95. Canadaplease add $12. Overseas rate is $59.95 peryear. To subscribe, renew or change addresswrite to the Circulation Department (PO Box876, Peterborough, NH 03458-0876) or tele-phone (603) 924-9464 or FAX (603) 924-9467for MC/Visa/Discover charge card orders.E-mail: [email protected] gift subscriptions please include gift recipi-ent’s name and your own, with remittance. Agift card will be sent.

EDITORIAL INQUIRIESSend editorial correspondence and manu-scripts to audioXpress, Editorial Dept., PO Box 876, Peterborough, NH 03458-0876.E-mail: [email protected]. No re-sponsibility is assumed for unsolicited manu-scripts. Include a self-addressed envelopewith return postage. The staff will not answertechnical queries by telephone.

CLASSIFIEDS & WEB LISTINGSContact Nancy Vernazzaro, Advertising Department, audioXpress, PO Box 876, Peterborough, NH 03458, 603-924-7292,FAX 603-924-6230, E-mail [email protected] in the USA. Copyright © 2001 by AudioAmateur Corporation. All rights reserved.

IN EVERY ISSUE

WEBSITE RESOURCESFind your favorite advertisers on-line .................................... 39CLASSIFIEDSAudio-related items forsale or wanted........................ 93AD INDEX ........................ 93

YARD SALEFree classifiedsfor subscribers ...................... 94

solid state tubes speakers

CO

DE Articles and letters identified with the following symbols:

page 66

page 64

Editor and PublisherEdward T. Dell, Jr.

Regular ContributorsEric Barbour Erno BorbelyRobert Bullock Richard CampbellJoseph D’Appolito Vance DickasonBill Fitzmaurice Gary GaloCharles Hansen G.R. KoonceNelson Pass Richard PierceReg Williamson

Vice PresidentKaren Hebert

Dennis Brisson Assistant PublisherMarianne Norris Editorial AssistantAmity Palmer Editorial AssistantTina Hoppock Graphics DirectorMelinda Taylor Production AssistantAmy Wyman Production AssistantLaurel Humphrey Marketing DirectorKelly Bennett Customer ServiceSharon Iozzo Customer Service

Advertising DepartmentAlden Sales, Inc.PO Box 271Rindge, NH 03461603-899-3010Fax: 603-899-2343E-mail: [email protected]

Rich Alden

Advertising Sales

Louise O’Sullivan

National Sales Manager603-899-5068

Nancy Vernazzaro

Advertising/Account Coordinator

T H E S TA F F

The peculiar evil ofsilencing the expression

of an opinion is, that it is

robbing the human race;posterity as well as the

existing generation; thosewho dissent from the

opinion, still more than those who hold it.

JOHN STUART MILL

D E PA RT M E N T S

EDITORIALOut, Out, Damned DotBy Edward T. Dell, Jr.....................................................................2AUDIO NEWSWhat’s new on the audio market ..................................................6BOOK REVIEWCunningham Radiotron ManualBy Scott Frankland ....................................................................68XPRESS MAILReaders speak out ....................................................................72CLASSIC CIRCUITRYPilot 264Courtesy of Lance Cochrane ......................................87ASK AXAdvice from the experts ..............................................88NEW CHIPS ON THE BLOCKNSC LM 4651/4652Linear Systems semiconductorsBy Charles Hansen ......................................................90GLASS SHARDChoosing Coupling Capacitor ValuesBy Michael Kornacker..................................................92TEST TRACKSReaders’ favorites to test audio systems By Fernando Garcia ....................................................96


NEW PUBLICATIONSAved/Generation Electronics is offering a capabilities brochure describing design andmanufacturing services for a wide range of custom cable assemblies, various types ofbattery packs and technologies, and interconnection components; for more information,visit www.aved.com. Focal Press (www.focalpress.com) announces the upcoming publi-cation of Modern Recording Techniques, Fifth Edition, a leading teaching guide, com-prehensive manual, and reference tool for those wishing to learn professional recording.Sams Technical Publishing (www.samswebsite.com) recently released their 2001 Annu-al Index. Sennheiser (860-434-9190) recently published The Handy Guide to EvolutionWireless Systems, a free, 32-page overview and application primer for Evolution wirelessmicrophone systems.

ELECTRONIC CAD FOR WIN95/98,2000 AND NT WORKSTATIONSAMS, Inc. announces the availability and shippingof Circuit Creator Windows 2000 and NT com-patible Computer Aided Engineering (CAE). Thisfourth-generation software integrates three essen-tial components (schematic capture, board design,and auto routing) in a single product. Advanced Mi-crocomputer Systems, Inc., 1460 SW 3rd Street,Pompano Beach, FL 33069, (954) 784-0900,(800) 972-3733, e-mail:[email protected], website: www.advancedmsinc.com.

6550EH VACUUM TUBEElectro-Harmonix added the 6550EH to the EHline of vacuum tubes. This tube offers 42W ofplate dissipation in tetrode operation and 44W

in triode. New features in-clude a ruggedized screenconstruction, phosphorbronze side rods for gridand screen, triple micasupports, unique “fourpillar” nickel steel siderod mica supports, Tung-Sol style “coke bottle”glass, heavier plate materi-al and more surface area

for increased cool-ing, and a newcathode coat-

ing formulation.Electro-Harmonix,

20 Cooper Square,New York, NY 10003,

(212) 529-0466, (800) 633-5477, FAX (212)529-0486, e-mail: [email protected], website:www.ehx.com.

AUDIOPHILE POWER AMPLIFIERSSMART Devices, Inc. introduces two uniqueaudio power amplifiers, directed at the high-end audiophile market, the 2X150VT Hybridamplifier (with a dual triode input) and themodel 2X150 (without the tube). These am-plifiers offer a patented MOSFET design andan extremely high damping factor, are conser-vatively rated at 150W per channel, and havea stainless steel top cover and steel bottomchassis for maximum magnetic shielding.Smart Devices, Inc., 5945 Peachtree CornersEast, Norcross, GA 30071, (800) 45-SMARTor (770) 449-6698, FAX (770)449-6728, e-mail:[email protected], website: www.smart-cinema.com.

Audio News

During my one-hour commuteto work every weekday, I haveplenty of time to just thinkabout stuff. One particular

morning, I had a revelation that hit melike a ton of Oreos: Egad, I’m out ofshape! Although we have a stationaryexercycle in our basement, why onGod’s green earth would I want to both-er with that? If I could listen to musicwhile pedaling, I just might be able totrick myself into a healthy dose of self-discipline.

The results certainly remain to beseen, but I decided to build a quick andeasy pair of loudspeakers using somefull-range drivers that were sittingaround in storage. I decided this mightbe an interesting exercise to help newspeaker builders get their hands dirtyand experienced folks might like theirtime-coherence. You won’t need muchwoodworking experience to completethis project, but you must love to listento music and be able to have fun! You donot need to wear a pulse-rate monitor.

If you ask me, full-rangers often get abad rap. I’ve heard for years about theirpoor high-frequency response, bad dis-persion characteristics, and plenty ofother reasons never to go near them.George Gershwin said it best when hewrote “It Ain’t Necessarily So.” Al-though I suppose it is possible that hewasn’t referring to full-range speakers,it does apply here.

For now, please forget all precon-

ceived notions. And to befair, remember how manyheadaches multi-driver sys-tems can cause with theircrossover systems and driv-er time-adjustment issues.Every configuration hassome trade-offs.

There is a small, butadamant, faction of people who swear bythe time-coherence and imaging of a sin-gle-driver configuration. They are oftenthe same folks who prefer low-poweredsingle-ended triode amplifiers and ultra-efficient horn loudspeaker designs. Ifmy little excursion into the full-rangeworld happens to pique your interest, I encourage you to visit the Single Driv-er Website at http://melhuish.org/audio/index.htm. I have found thisforum of people to be very helpful withtheir suggestions and willing to openlyshare their own results.

The enclosure for this project is sim-ply a piece of PVC pipe. When I de-signed these loudspeakers, I planned tohang them from the ceiling with theback-ends of the pipes aiming into thewall/ceiling junction. I believe thismakes Slim Jims good candidates for a

garage or basement workshop. Ofcourse, please feel free to modify and/orput this project wherever you’d like(Photo 1).

PREPARING THE INGREDIENTS • Get two Radio Shack 40-1197 drivers

(approximately $12 each).

• Cut two 26.5″ length pieces of 4″ di-ameter PVC pipe (enclosures).

• Cut two 2.75″ length pieces of 1.5″ di-ameter PVC pipe (bass reflex vents).

• Cut two 5″ × 5″ squares of any woodyou have around. Using a hole saw,cut a hole in the center of each

Get in shape—sonically and physically—with these 4″ driver speakers.

By Ken Ketler

Slim Jims: Get the Skinny on an Easy-to-Build Project

ABOUT THE AUTHORKen is a Technical Writer who enjoys many flavors ofmusic and audio equipment. He and his wife, Julie,often review consumer equipment for audioXpressmagazine and happily spend large quantities of timewith their son, Thomas.

PHOTO 1: The completed speaker in place.

PHOTO 2: Fitting speaker cables to the unit. PHOTO 3: Inside the pipe.


square of wood that will tightly fityour 1.5″ diameter PVC pipes. Also,drill a small hole in each piece thatwill fit whatever speaker cable youprefer to use (Photo 2). Please feelfree to use whatever type of connec-tors you like. I used cable simply be-cause my budget was thin, unlike mystomach.

• Cut two 23″ × 13″ pieces of fiberglassspeaker damping material (the yel-low stuff from Radio Shack is fine).

• Cut two 30″ length pieces of speakercable.

• Purchase a small can of polyurethane.

• Buy a roll of clay rope caulking.

• Obtain a tube of latex or siliconecaulking. I prefer latex because it is stiffer. Since I didn’t use screws to secure the driver to its enclosure,I thought that latex would have a better resistance to flexing duringloud musical passages, which wouldreduce efficiency and final soundquality.

GET IT TOGETHERFor each loudspeaker:

1. Insert the 1.5″ pipe in the center ofthe wood square so that it is flush atthe outside edge. Seal with the latex.

2. Run your zip cord through the drilledhole in the wood square so that thecable is long enough to stretch up thelength of the whole enclosure to thedriver. Again, seal with latex.

3. Glue (latex) the wood square to oneend of the enclosure pipe. Make surethat there are no air gaps betweenthe wood and the enclosure pipe.

4. Using latex caulking, glue the fiber-glass sheet to the inside perimeterdown the length of the 4″ pipe. Besure it is flush with the inner-wall so itdoesn’t obstruct the path between thedriver and the reflex vent (Photo 3).

5. Solder the speaker cable to the 40-1197 driver.

6. Glue (yep, latex) the driver to the openend of the enclosure pipe. Again,make sure there are no air leaks.

7. Splice the other end of the speakercable to a longer piece, which youconnect to your amplifier. This willmake it easier to move the Slim Jimslater (no huge pieces of cable hangfrom the enclosures). Don’t forget tocover your splices with heat-shrinktubing or tape to avoid short circuits.

Of course, you must allow the latexto dry in between each step. If you’dlike to decorate the loudspeaker, con-tact paper seems like a good cheap so-lution. I’ve used a funky granite-lookingpattern that...well,—certainly looks bet-ter than just plain PVC!

TAME THAT MIDRANGE!This is where the art in speaker buildingcan really kick in! The 40-1197 driverhas a rather bothersome mid-frequencypeak. This tends to make it sound likean AM radio speaker. It is, after all, a“budget” driver. There are three practi-cal ways that I know of to smooth outthis peak to make the 40-1197 a very“musical” speaker. I suggest fine-tuningthe system after it’s built so that you canhear the changes as you make them. Asa result, your changes will be based onyour own hearing within your ownacoustic environment.

The first and quickest solution tosmooth out this midrange peak is to usea graphic equalizer. You’ll also be ableto tweak your room response if it is par-ticularly troublesome. Another fairlysimple solution is to build a notch filterwith three simple components (Fig. 1):

R = 6.8ΩC = 4.7µFL = 0.8mH

This filter is often recommended onthe Single Driver Website , but I admitright here, right now, I haven’t tried itmyself.

The third solution is the most com-plicated, yet the most educational. Mod-ify the driver! After reading a great dealof opinions on the subject, here are themodifications that I chose to use on my40-1197 drivers (Photo 4):

1. Apply two coats of full-strengthpolyurethane to the front side of thedriver cone. This will help to dampvarious resonant modes and smoothout some of the “jagged edges.” Besure not to get any on the center dustcap or surround of the driver.

2. The basket of the 40-1197 is astamped steel design. It can oftenhave a bell-like resonance that youcan easily damp. A few wraps of clayrope caulking around the magnet/basket junction does a great job of de-resonating the frame of this driver.

3. I have heard many different opinionsregarding this third step, but I havetried it and am pleased with the re-sults. With a hot soldering iron, verycarefully burn a ³⁄₁₆″ diameter hole inthe center dust cap of each driver.This will further smooth out themidrange response by stopping vibra-tion breakup modes that can trulyruin the midrange quality of this driv-er. Again, do this with the system as-sembled, so you can fine-tune the sizeof the hole “by ear.”

The clearest explanation I haveheard was a simple analogy to “drilling


PHOTO 4: Modifying the drivers.

6.8 Ω

4.7 µF

0.8 mH

FIGURE 1: Filter circuit.

a hole at the end of a crack to keep itfrom spreading.” In fact, some modelsin the Fostex FE-AV Series of full-range drivers have a conspicuous-look-ing hole right in the center of theirdustcaps! Hmm, maybe it’s not such acrazy idea.

RESULTSI hung the Slim Jims (using string)from the ceiling beams in my cellar sothat the drivers are angled down to-ward me, the pedaler, while the portend is firing right into the cornerwhere the ceiling meets the wall. Ifyou are thinking of hanging them as Idid, put dabs of superglue on theknots in the string so they don’t comeuntied (Photo 1).

Overall, they sound amazingly good,especially for little 4″ drivers! Thestereo imaging is excellent and thebass is clean although somewhat limit-ed. The upper frequency responsesounds great up to about 15kHz anddrops off around 16.5kHz.

Now, I can hear you hecklers in theback row. Yes, a cone driver begins tobeam its signal like a flashlight nearthe frequency at which wavelengthequals speaker diameter (in this case,somewhere around 3.5kHz). I may beoverly optimistic, but I think this is ac-tually an asset. In my concrete base-ment (a very reflective acoustic environ-ment), most of the high-frequency sig-nal arrives at the listening positionwithout being corrupted by early reflec-tions. Great for musical clarity andstereo imaging!

The designed frequency responsedrops off at 81Hz and the enclosure/

port is tuned around 84Hz. Firing theports into the corner gives a smoothbass rise right around cut-off. When Iswept the loudspeakers with a tonegenerator, their in-room output wasstrong to around 80Hz and continued toprovide useful output down to about62Hz, but then dropped off like a stone.Although I don’t have a sophisticatedaudio test lab, I relied on the funny-looking microphones that are attachedto my head and my Radio Shack soundlevel meter.

According to my computer design(using BassBox 5.1 software), the“group delay” at the tuning frequencyis approximately 8ms (Fig. 2). Thismeans that the bass response is quickand punchy. I think this is because Iam not asking the driver to do morethan it wants to—80Hz is right in the 1197s’ ballpark. If you try to forcethis driver to work much below this, itwill pump like a sweaty guy on an ex-ercise bike and its group delay/tran-sient response will go to heck in ahand basket.

If you are more of a bass fiend, con-sider adding a subwoofer or using larg-er-diameter full-range drivers in a simi-lar arrangement. But for now, try build-ing yourself a pair of Slim Jims. They’requick and cheap, kind of ugly, but sur-prisingly musical-sounding!

As far as the “other” results are con-cerned, I have actually been draggingmyself out of bed every morning at 5:30and cranking away to the best tunes Ican find in the dark. After a few songs, Iusually start jammin’ like a rock star.Who’d have thought that playing airguitar could be so good for you?


FIGURE 2: Design response.

Please visit our web page for more

specifications. www.madisound.com

FOSTEX

FE108 $82.00 Each

FE168 $134.00 Each

FE208 $168.00 Each

MADISOUND SPEAKER COMPONENTS, INC.8608 UNIVERSITY GREEN

P.O. BOX 44283

MADISON, WI 53744-4283 U.S.A.TEL: 608-831-3433 FAX: 608-831-3771

e-mail: [email protected]

Web Page: http:/www.madisound.com

Madisound is pleased to offer the Fostex

Sigma Series of full range transducers.

These speakers are ideal for the purist who

doesn’t want a speaker encumbered by caps

and coils. The relatively flat impedance

and high efficiency will lend these speakers

for use with tube amplifiers.

• 8” (9” OD)

• Fs 45 Hz

• 8 ohm

• Qms 4.03

• Qes 0.23

• Qts 0.21

• Vas 36.9 ltrs

• Mmd 12.0 g

• 96.5 dB

• 4” (5” OD)

• Fs 80 Hz

• 8 ohm

• Qms 2.21

• Qes 0.22

• Qts 0.20

• Vas 7.07 ltrs

• Mmd 2.7 g

• 92 dB

• 6.5” (7.5” OD)

• Fs 60 Hz

• 8 ohm

• Qms 4.53

• Qes 0.40

• Qts 0.37

• Vas 15.3 ltrs

• Mmd 6.5 g

• 92 dB

This project is a simple 25Wpc(watts per channel) power am-plifier based on the NationalSemiconductor (NSC) LM3875T

amplifier IC. I use it to power the rearambience speakers of my audio system.The L+R and L−R ambient informationis extracted from the stereo signalusing the Audio Amateur QuadPod1.While I built the QuadPod into my am-plifier chassis, the project as presentedhere is just as applicable to a two-chan-nel amplifier project alone (Photo 1).

The 3875 has appeared in Audio Am-ateur publications before2, but it was asingle supply subwoofer amp from Fig.2 of the LM3875 data sheet. I prefer adual-supply design, which dispenseswith the large electrolytic output cou-pling capacitor and a number ofbias components. I also madechanges to the Zobel RC to matchmy speakers, changed the gaincapacitor from 10µF to 39µF to ex-tend the bass (my −3dB point is4Hz versus 16Hz for data sheetFig. 2), and changed the input net-work. Total parts cost is about$180 (without the QuadPod).

HOW IT WORKSThe schematic diagram for theaudio amplifier is shown in Fig. 1,along with the LM3875 packageoutline and its five pin connec-tions (the other six pins have noconnection). The parts list is inTable 1. One of the nice advan-tages of toroidal power trans-former T1 is the dual secondarywinding.

To keep the power supply-in-duced noise to a minimum, I full-

wave-rectified both windings with CR1and CR2 to ±30V DC no load, or about±27V DC at 25W. Capacitors C1–C4snub the rectification noise at CR1,while C12–C15 do the same for CR2. Ca-pacitors C5–C8 filter the rectified DCthat is applied to power amplifier ICsU101 and U201. Film caps C9 and C10filter RF noise from the DC supply rails.

In addition, the supply rails to thetwo ICs are locally bypassed with 10µFtantalum caps as recommended by theapplication notes. R1 limits current to

power-on indicator LED 1, and fuses F1,F2, and F3 provide fault protection.Going from a single rectifier bridge totwo bridges reduced the 180Hz power-supply component from −79dB to −91dB, taking a big bite out of the “N”portion of THD+N.

Each channel of audio amplificationis provided by a National Semiconduc-tor LM3875T audio power amplifier IC.This device is of much higher audioquality than the power amplifier ICstypically used for car audio and com-

A High-Quality Two-Channel Chip Amp


FIGURE 1: Power amplifierschematic and LM3875 outline.

This IC-based rear-channel amp has several uses to improve the quality

of your sound system. By Charles Hansen

PHOTO 1: Front view of completed unit.

A-1967-1

OnlyRight Channelshown.

puter speakers. It is housed in a plastic11-pin metal-tab package. A data sheet

for the LM3875T is available at the Na-tional Semiconductor website as a PDF

file (www.national.com). I suggest youdownload it so you can avail yourselfof the application notes, data, and testcircuits.

I will describe the left channel sinceboth are identical. The line-level audiosignal is applied to DC blocking capaci-tor C101 and R101 through input jackJ101. R102 and C102 reduce the gain athigh frequencies to avoid quasi-satura-tion output transistor oscillation, andhelp suppress external RF noise. R102also limits the input current that maybe passed through to the output ofU101 after power-down, once the supplyrails fall below ±1.5V DC.

U101 is configured as a non-invertingamplifier. The mid-frequency gain is setto 21 (26dB) by R104 and R103 (Av = 1 +R104/R103). C105 produces a LF pole at4Hz and ensures unity gain at DC. R105and C106 produce a HF pole to roll offthe gain above 170kHz. R105 is neededbecause the LM3875 is stable at or aboveclosed-loop gains of ten or greater.

The amplifier output is coupled tothe load through R108 and a parallel L-R network consisting of L101 and R106.


FIGURE 2: LM3875 equivalent schematic. National Semiconductor Corp. andreprinted with permission of National Semiconductor Corporation. NSC reservesthe right at any time to change said circuitry without notice. A-1967-2

The impedance of L101 increases athigh frequency, so that R106 can decou-ple a capacitive load and reduce the Qof the resulting series resonant circuit.I wound L101 with 18-gauge wire so itsresistance is negligible. However, afterinitial tests with a load of 8Ω in parallelwith 2µF produced oscillation near125kHz, and enough thermal distressthat the chip protection circuits shutdown, I added R108 to further decouplethe IC from capacitive loads.

To achieve ideal pole-zero compensa-tion with this inside-the-loop resistor,you could add another cap from outputto inverting input. However, this capaci-tor would need to be fairly large, so Isettled on some harmless frequency

FIGURE 3:Heatsink detailcross-section.

FIGURE 5: Stuffing guide.

A-1967-3

A-1967-5

FIGURE 4: Copper side of circuit board. (100%)


A-1967-4

PHOTO 2: Amplifier circuit board close-up.

peaking at 46kHz. This manifests itselfas several cycles of damped ringingduring square-wave testing. If you haveno intention of driving capacitive loadssuch as electrostatic speakers or “diffi-cult” crossovers, leave out the 0R22 re-sistor and enjoy the added efficiency.

The Zobel network, R107 and C107,

produces an output stage pole thathelps prevent high-frequency oscilla-tion with real speaker loads. You mayneed to adjust these values to suit yourown speakers, or use the default valuesin the LM3875 data sheet.

The LM3875 is capable of delivering56W of continuous power to an 8Ω load


FIGURE 6: Chassis interior layout.

TABLE 1PARTS LIST

DESIGNATION VALUE, DESCRIPTION VENDER P/NC1–C4, C12–C15 10nF 100V V-stacked film DK P4713-NDC5–C8 2200µF 35V aluminum HFQ DK P10335-NDC9, C10 100nF 50V 2% PP film DK P3104-NDC11 4n7 400V film DK P1066-NDC101, C201 4µ7 100V Mylar DK EF1475-NDC102, C202 100pF 5% NPO ceramic DK P4456A-NDC103, C203, C104, C204 10µF 35V Tantalum DK P2065-NDC105, C205 39µF 35V aluminum HFQ DK P5727-NDC106, C206 47pF 5% NPO ceramic DK P4452A-NDC107, C207 22nF 50V PP film DK P3223-ND CR1, CR2 (see text) 25A 100V bridge DK MB251WMS-NDF1 1A SB fuseF2, F3 4A SB fuseJ1 IEC filter receptacleJ101, J201 Phono jack, gold (pair) PE 091-1120J102, J202, J103, J203 5-way binding post (pair) PE 091-1140L101, L201 700nF air-core See textLED1 Red LED indicatorR1 3k3 ½W carbon resistor DK 3.3KH-NDR101, R201 100k ¼W 1% metal film DK 100KXBK-NDR102, R202 221Ω ¼W 1% metal film DK 221XBK-NDR103, R203 1k00 ¼W 1% metal film DK 1.00KXBK-NDR104, R105, R204, R205 20k0 ¼W 1% metal film DK 20.0KXBK-NDR106, R206, R107, R207 10Ω 1W 5% metal oxide DK P10W-1BK-NDR108, R208 (see text) 0R22 5W wire wound PE 015-.22S1 SPST toggle switch, 10AT1 20+20V AC, 4A toroidal Plitron 057027201U101, U102 LM3875T power amplifier IC DK LM3875T-ND

Fuse clip, single (3) DK 3536-NDHeatsinks (2) DK 294-1076-NDHeatsink insulator pad, K10 (2) DK BER118-NDTO-220 mounting kit Jameco 34121Chassis Sescom MC-19ARCA jack insulators DK SC1144-ND

and SC1145-ND

A-1967-6

with less than 0.1% THD+N over theaudio band. This requires supply railsof ±35V DC. Since the device can func-tion with supply rails ranging from±10V DC to ±42V DC, I chose to design amore modest 25Wpc (8Ω) amplifierusing ±27V DC rails.

If you prefer more power, you willneed a larger transformer, with a highersecondary voltage. Higher current (donot exceed 4A) will also require a largerwire size for the speakers and power-supply wiring. Use the same wire gaugeas that of your transformer secondaryleads. You will also need to increase thevoltage rating of the polarized caps. Thetantalums should not see more than 78%of their rated voltage, and the aluminumcaps should be limited to 80% of ratedvoltage at maximum audio output.

You can also add a stereo volumecontrol between the input jacks and theamplifier board. I used a dual 20k audiopot, a small diameter Alps control I hadon hand. Mouser sells 1″ diameter dualaudio pots (the part number is 313-2420-value); you can use 10k or 50k (20k isnot available).

LM3875 DESCRIPTION AND PERFORMANCEThe NSC LM3875 is one of three Over-ture amplifiers (the 56W LM3875, the56W+Mute LM3876, and the 68W+MuteLM3886). DigiKey sells these devices asrated for 40W, 40W+Mute, and 60W+Mute, respectively. Overture amplifiersare all based on Bob Widlar’s classicLM12 design3, and there are good ap-plication notes available on the NSCwebsite.

Monolithic power amplifiers such asthe LM3875 use a quasi-complementaryoutput stage as shown in the equivalentschematic (Fig. 2). Even the comple-mentary bipolar process is not yet ableto produce matched-complementaryNPN and PNP high-power transistorson the same chip. A monolithic PNPoutput device (whether lateral or verti-cal) has lower performance than theNPN device.

To compensate for this design com-promise4, the LM3875 uses a compos-ite “power PNP” made from a lateralPNP and the usual NPN power de-vices, with some careful engineeringfor much lower Class-AB crossover

distortion. The input stages are de-signed for low noise (the input is run“richer”). Since all the devices are onone piece of silicon, the thermaltracking is excellent, ensuring a sta-ble output-stage bias point.

Undervoltage protection allows thepower-supply voltages to approachtheir full levels before the amplifier isturned on. This prevents any DC out-put spikes. A similar function occurson shutdown, where the output of theLM3875 is brought to ground beforethe power supplies decay. Overvoltage

protection limits the output current to4A peak, while providing output volt-age clamping.

SPiKe (self peak instantaneous tem-perature, Kelvin°) protection limits theoutput array safe operating area (SOA)voltage-power-current profile. This in-volves small temperature-sensor diodesdistributed all over the die and allaround in the power transistors, so anyhot spots cause it to shut down, ratherthan go into thermal runaway.

Thermal protection prevents long-term thermal stress. The LM3875 will

FIGURE 7: Chassis front and rear views.


PHOTO 3: Interior view.

PHOTO 4: Rear view of unit.

A-1967-7

shut down if the die temperaturereaches 165°C, and resume below155°C. A sustained fault will cause theamplifier to cycle between the thermalshutdown limits, greatly reducing thethermal stress. Momentary thermal

transients, such as a short-circuit atthe output to ground or to either sup-ply, or back emf from a reactive load,can cause the die temperature to rise100°C in less than 100µs. This also re-sults in a safe shutdown.

HEATSINK DESIGN The methodology for the heatsink de-sign is provided on the LM3875 datasheet, pages 10 and 11. Based on theformula provided, each IC dissipates10.7W with 25W audio output into 8Ω. Ifyou use 4Ω speakers, the IC dissipationwill increase to 25.5W, and you mayneed larger heatsinks with a maximumθSA (thermal resistance sink to ambi-ent) of 4.8°C/W. I decided to use afinned heatsink mounted on each IC(IERC p/n 7-404 with 3°C/W—see partslist). I made the heatsinks part of thePC board assembly.

I attached the bottom of eachheatsink to the chassis with a 6 × ½″sheet-metal screw to provide a goodthermal path and to take the mechani-cal stress off the ICs (Fig. 3). Photo 2 isa close-up of my prototype amplifier PCboard. I made some changes based ontesting, so it looks different than thefinal layout in this article.

Do not mount the heatsinks ontoU101 and U201 until after all the com-ponents are stuffed and all the wiring iscomplete (more on this later). This pre-


FIGURE 8: Rear panel lettering. A-1967-8

4470 Avenue ThibaultSt-Hubert, QC J3Y 7T9Canada

SOLEN INC. Tel: 450.656.2759

Fax: 450.443.4949

Email: [email protected]

WEB: http://www.solen.ca

[ y

COMPONENTS:

SOLEN HEPTA-LITZ AND STANDARDINDUCTORS ANDCAPACITORS - THECHOICE OF MANYHIGH-END SPEAKERMANUFACTURERS.

HARDWARE:

POWER RESISTORS, L-PADS, CABLE,ABSORBING ANDDAMPING MATERIALS,GOLD SPEAKERTERMINALS, GOLDBANANA PLUGS ANDBINDING POSTS, GRILLFASTENERS, PORTTUBES AND TRIMRINGS, PAN HEADSCREWS, SPIKES AND TEE NUTS WITHALLEN HEAD BOLTSAND PLENTY MORE...

Solen crossover components - used by the most discriminating

loudspeaker manufacturers.

SOLEN HEPTA-LITZ INDUCTORSAir Cored Inductors, Litz-Wire Perfect Lay Hexagonal WindingValues from .10 mH to 30 mHWire Size from 1.3 mm (16AWG) to 2.6 mm (10 AWG) 7 Strands

SOLEN STANDARD INDUCTORSAir Cored Inductors, Solid Wire Perfect Lay Hexagonal Winding

Values from .10 mH to 30mHWire Size from 0.8 mm (20AWG) to 2.6 mm (10 AWG)

SOLEN FAST CAPACITORSFast Capacitors, Metalized PolypropyleneValues from 0.10 µF to 330 µFVoltage Rating: 630, 400, 250 VDC

CROSSOVER AND SPEAKER PARTSMetalized Polyester Capacitors, 1.0 µF to 47 µF, 160 VDC, Non PolarElectrolytic Capacitor, 22 µF to 330 µF, 100 VDC, Power Resistors 10 W, 1.0 Ω to 82 Ω, 8 Ω L-Pads plus all the hardware and suppliesto complete any speaker project.

C A L L T O O R D E R T H E S O L E N C R O S S OV E RC O M P O N E N T C ATA L O G F O R $ 8.00 PACKAGEDWITH T H E C O M P R E H E N S I V E S O L E N S P E A K E RD R I V E R C O M P O N E N T S C ATA L O G .

vents stress or damage to the amplifierICs and insulators while you are han-dling the PC board.

CIRCUIT BOARD LAYOUTI designed an etched, single-sided cir-cuit board for this project. The proto-type board was furnished by one of theaudioXpress suppliers (Atlas CircuitsCompany, 1500 Old Lake Rd., PO Box892, Lincolnton, NC 28092). The copperside of the PC board is shown in Fig. 4.The board uses two wire jumpers tocomplete some of the circuits.

If you use the prototype board suppli-er, the holes will be drilled with a singledrill size. You must re-drill a number ofthese holes to accommodate wires andthe larger component leads of someparts. Use the drill sizes in this table:

C5–C8, C101, C201, #56 drillR108, R208LM3875s, R1 #54 drillL101, L201; ¹⁄₁₆″ drillwire and jumper padsMounting holes ¹⁄₈″ drill

The stuffing guide is shown in Fig. 5.Be sure to install both 20-gauge wirejumpers, using sleeving to preventshorts to other components (or use in-sulated hookup wire). JMP1 and JMP2indicate these jumpers.

I designed inductors L101 and L201using a formula in the Reference Datafor Radio Engineers5. Wrap nine turnsof 18-gauge wire around a ½″ diameterwood dowel. Space the turns as evenlyas possible, so the finished length isalso ½″. Dress the free ends of the leadsso they fit into the PC board holes.

In keeping with good assembly prac-tice, install the least sensitive partsfirst, followed by the more sensitiveparts. Begin by soldering the passiveparts (inductors, resistors, then capaci-tors). Install the LM3875s last. Double-check the orientation of the polarizedcomponents.

CONSTRUCTIONI used a 17 × 7 × 3.5″ Sescom MC-19Achassis for my project, which has

enough room for the QuadPod board.For the amplifier alone, you can use theMC-25A (12.75 × 7 × 3.5″). Figure 6


TABLE 2CROSSTALK

FREQUENCY R TO L L TO R100Hz −60dB −60dB1kHz −57dB −52dB10kHz −39dB −35dB20kHz −32dB −30dB

FIGURE 11: THD+N versus frequency. A-1967-11

FIGURE 9: Wiring diagram.

FIGURE 10: Frequency response. A-1967-10

A-1967-9

shows the dimensioned interior of theamplifier chassis layout. All the wiringon the power-supply side is twisted tominimize magnetic fields, which couldinduce power-supply noise or hum.

I drilled a pattern of ¹⁄₈″ holes aboveand below the heatsinks to promote

convective cooling. One hole must lineup with the center slot in the bottom ofeach heatsink, on the center-line of thehole in the LM3875. The rest of theholes are evenly spaced ¼″ apart fromeach heatsink mounting hole. The slotin the heatsink can accommodate a fair

amount of error toward or away fromthe PC board.

The interior view of my amplifier(Photo 3) includes the QuadPod circuitboard, switches, jacks, and controls inmy particular amplifier. I mounted thefuse holders on a 2″ × 3.25″ aluminumL-shaped bracket so they are easily ac-cessible after removing the chassis endplate. I mounted CR2 on the inside ofthe rear panel.

I suggest this location because itgives more room for the wiring. Howev-er, feel free to rearrange the compo-nents as you see fit. You may even pre-fer to locate your DC supply filter capsat the rectifiers, which will reducepower-supply noise even further.

The front and rear views, with dimen-sions for chassis-mounted components,are shown in Fig. 7. I decided to paintthe front panel flat black and use a min-imum of lettering (Photo 3). If you buildit without the QuadPod board, youdon’t need any lettering, since the frontpanel will have only a power switch andLED (self-explanatory).

All the detailed information for my


FIGURE 12: THD+N versus output power.A-1967-12

amplifier is on the rear panel (Photo 4).Figure 8 shows the lettering designa-tions for my rear panel. I broke it intotwo sections since it is too long for a sin-gle 8½ × 11 sheet. I made a full-size copyof this lettering on drafting appliquéfilm, which is an adhesive-backed trans-parent plastic (see parts list).

The photocopy lettering is black, so Ileft the rear panel plain aluminum.Apply the lettered film to the panel inthe proper locations. Cut through thefilm at the drilled holes and then sprayit with two coats of clear polyurethanefor protection. Then you can mount thecomponents.

Mount the four snubber caps on eachbridge rectifier. I took the easy way out,

using just the four 10nF capacitorsrather than calculating the optimum C-RC snubber6. You may also care to trylow-noise, soft-recovery diodes, as rec-ommended by Gary Galo7. I tappedpower for the QuadPod board from the±30V DC supply and stepped it down toabout ±15V DC with series resistors (onthe standoffs near CR1).

All circuit-board mounting holes havesufficient clearance for a 4-40 screw withflat washer. I used four ³⁄₈″ long 4-40tapped aluminum hex spacers to mountthe board to the chassis. The speakerposts should be exactly ¾″ apart, withthe terminals holes aligned vertically.

An easy way to do this is to mountthe jacks in place finger tight. Plug in a

dual banana plug to set the ¾″ distance.Bend a piece of 14-gauge solid barewire (a stripped piece of Romex is fine)into a “U” shape so the legs are also ¾″apart. Insert this wire through the holesin both adjacent connectors and tight-en the mounting hardware.

Figure 9 shows the wiring to parts lo-cated off the circuit board. After in-stalling the mechanical parts in thechassis, place the board near its finallocation and then wire to the switch,LED, jacks, and volume control if youdecide to add one.

I used AWG-20 for the power supplyand speaker wiring—two AWG-20 wiresin parallel (equivalent to AWG-17) forthe amplifier power ground and AWG-22 for the signal wiring. The easiest wayto position the front and rear panels forwiring is to place them in the rails onthe bottom plate and hold the threepanels together with large rubber bandsat each end.

The amplifier uses a single chassisstar ground point, located near the fusebracket. I ran separate ground wiresfrom CR1, CR2, each PC board, thespeaker (−) posts, and the RCA jacks tothis point. Also connect the ground lugof the IEC connector J1 here as a safetyground.

I used insulators on the RCA jacks toisolate them from the chassis. Useshielded wire from the input jacks tothe circuit board, with the shieldgrounded only at the jacks. The pur-pose of all these single-ended groundand shield connections is to preventground loops and noise currents on theshields, which could couple into thesignal. Finally, after all wiring is com-plete, install the board in its final loca-tion. I used nylon wire ties to make thewire harnesses neater.

Since I also installed the QuadPodboard in my chassis, there are addi-tional twisted-pairs of wiring involved.Only one end of the ground lead ineach twisted-pair is connected to aground, at the jacks, or a circuit board.Again, this is to eliminate any ground-loop current.

The rubber feet that come with thechassis are kind of skimpy, so I pur-chased ¾″ square vinyl bumpers at ahome-improvement center. I used fiveof them, with one under the heavy


ms0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0

V

-10

-8

-6

-4

-2

0

2

4

6V

-1.2

-0.8

-0.4

0.0

0.4

0.8

1.2

1.6

2.0

FIGURE 13: Residual distortion, 1kHz.

kHz

0.0 0.2 0.4 0.6 0.8 1.0 1.2

dB

-100

-90

-80

-70

-60

-50

-40

-30

-20

-10

0

FIGURE 14: Spectrum of 50Hz sine wave.

A-1967-13

A-1967-14

transformer so the bottom plate wouldnot sag.

HEATSINK INSTALLATIONOnce all the other construction is com-plete, you can install the heatsinks ontothe LM3875 ICs. I used a TO-220 hard-

ware mounting kit (see parts list) to in-sulate the heatsink from the IC. I didnot use the mica insulator furnishedwith the TO-220 heatsink kit, replacingit with a K10 heatsink insulating pad(the tab of the 11-pin LM3875T packageis connected to the V-supply).

Alternatively, you can use theLM3875TF package, whose mountingtab is electrically isolated from the IC.These devices are harder to find thanthe LM3875T, however. DigiKey doesnot carry them. You should be able toorder them directly from NationalSemiconductor (www.national.com).

K10 is a thermally conductive Kap-ton polyimide coated with boron nitrideresin. This material is thermally superi-or to mica, silicone elastomer and ce-ramic, and almost as good as berylliumoxide (which is brittle, and its dust ispoisonous). In addition, the K10 doesnot need messy heatsink grease, sinceit is an elastomer that conforms well tothe metal surfaces.

The heatsinks in the parts list do nothave any mounting holes, so I drilled a¹⁄₈″ hole in each heatsink about 1¼″ upfrom the bottom (see following) andcentered side-to-side. You must careful-ly de-burr this hole on the side facingthe IC to prevent damage and a possi-ble short-circuit to the K10 insulator. Iused a countersink for this purpose.

This hole must place the bottom of


kHz

0 5 10 15 20

dB

-100

-90

-80

-70

-60

-50

-40

-30

-20

-10

0

FIGURE 15: CCIF intermodulation response. A-1967-15

the heatsink flush with the bottom ofthe chassis, so it is best to mount the PCboard in position on its spacers andmark the center of the hole in the IC tabon the heatsink. Use this mark as the lo-cation for the hole in the heatsink. I

also added a 4-40 flat washer under thelock washer and nut furnished with theheatsink kit (Fig. 3).

Lift the bottom center fin pin withneedle-nose pliers, insert a ½″ long #6sheet-metal screw through a hole in thechassis bottom plate, and screw it intothe heatsink slot. Be careful not to putany force on the IC while mounting theheatsinks. The sheet-metal screw pro-duces small metal chips as it cuts itsthreads, so be sure to turn the chassisupside down and dump out these chips,as well as any stray bits of solder andwire strands.

USING THE AMPLIFIERAs I mentioned in the beginning of thisarticle, I use the LM3875 amplifier formy rear ambience speakers, and theQuadPod board drives it in my particu-lar amplifier. The volume control setsthe rear ambience level. The rear-chan-nel outputs of the QuadPod are con-nected through switch S2 near the op-tional volume control to the line inputsof the LM3875 amplifier. The other sideof this switch selects the external input,which bypasses the QuadPod and vol-ume control.

LISTEN TO THE MUSICI did my listening audition using the ex-ternal input jacks, placing the 25WpcLM3875 amplifier below my 80WpcNAD-214 front-channel stereo poweramplifier. The NAD amplifier has a

higher gain than the LM3875 (30.2dBversus 26.4dB into 8Ω), so I used a pas-sive line stage equipped with inputpass-through jacks to reduce the NAD’sinput signal so the levels were identicalat 1kHz. This eliminated the “louder isbetter” psychological factor. Thus, I wasthen able to easily switch the speakerplugs from one amp to the other.

I used a number of CDs, including theHFN/RR Test Disc III (HFN020) that au-dioXpress has standardized for all theirequipment reviewers. I listened with thefull-range speaker system, as well aswith just my satellite speakers alone.Rear-channel ambient information is al-most devoid of bass, but is rich in themid and upper audio frequencies. I alsolet it drive the subwoofer alone.

The LM3875 sounded a bit brittle oredgy at first in the treble, then becamemore transparent as it warmed up. Ingeneral, the NAD has a bit more sound-stage and better imaging than theLM3875. It also has better bass exten-sion. Of course, this compares a 4A in-tegrated circuit with the paralleled 15A130W discrete complementary bipolarpairs in the NAD (I use the NAD withits soft-clipping circuit turned off). TheNAD’s low-frequency rolloff point isalso lower.

The bass with the LM3875 never gottubby or boomy, it was just a bit on thethin side, even with the satellites alone.(My speakers are not the most efficientat 86dB/2.83V/m, or maybe it’s that extra0R22Ω resistor I added?) Acoustic basswas sufficiently woody and resonant,and kick drums had a lively pop. It justdidn’t dig down to the lowest tones on afive-string electric bass as well.

The massed vocals in Jerusalem(Parry) were a bit less defined than withthe NAD. This selection is a difficulttest for any amplifier. The trumpets andstrings in Vivaldi’s Trumpet Concertoin C were very clear and bright. Theonly deficiency I could find vis-à-vis theNAD was a bit less definition in theharpsichord in the far right side of thesoundstage. The male voice and indi-vidual instruments in Peter and theWolf (Prokofiev) were very nicely repro-duced by the LM3875.

Two tracks on Trudy Desmond’s Tai-lor Made CD, “I’m Shadowing You” and“Make Someone Happy” (Jazz Alliance


SOURCESDK == DigiKey Corp.701 Brooks Ave. SouthThief River Falls, MN 56701-06771-800-344-4539www.digikey.com

Jameco Electronics1355 Shoreway Rd.Belmont, CA 94002-41001-800-831-4242www.jameco.com

Old Colony Sound Lab305 Union St.PO Box 876Peterborough, NH 03458-0876603-924-6371www.audioXpress.com

Mouser Electronics958 N. MainMansfield, TX 760631-800-346-6873www.mouser.com

PE == Parts Express1-800-338-0531 www.parts-express.com

Plitron Manufacturing8-601 Magnetic Dr.Toronto, OntarioCanada M3J 3J21-800-754-8766www.plitron.com

Sescom, Inc.517 Main St.PO Box 839Wellsville, KS 660921-800-634-3457www.sescom.com

kHz

0 5 10 15 20

dB

-100

-90

-80

-70

-60

-50

-40

-30

-20

-10

0

FIGURE 16: MIM intermodulation response. A-1967-16

TJA-10015), have a trombone and cellolocated closely together in the center ofthe mix. They interplay with each otherin almost the same register, and some-times with the upper notes of theacoustic bass. It’s amazing how muchalike these instruments can sound. TheLM3875 had no problem separatingthem, providing adequate space aroundeach instrument.

Next I listened to the big band in theAnne Hampton Calloway CD, To EllaWith Love (Sin-Drome SD-8933). Whilethe imaging and soundstage were againa bit less than with the NAD, the instru-ments were well defined. You could eas-ily tell an alto sax from a tenor, from abaritone.

The initial transients of cymbals, tri-angles, and chimes were not as sharpas with the NAD, but they did remainclear and bright as their notes decayedinto the low noise floor. The complexorchestration sounded full and natural,and her voice had a nice presentation.

The LM3875 is on a par with the bet-ter consumer audio equipment on themarket. Not as detailed as budget high-end discrete amplification from Adcom,Creek, NAD, Parasound, or Rotel, butcertainly respectable. At about $9 each,the LM3875 is one of the genuine bar-gains—in both cost and performance—in the audio IC marketplace.

MEASUREMENTSI operated the LM3875 amplifier at 2Winto 8Ω for one hour. The initial 0.03%THD reading dropped to 0.011% at theend of this run-in period. The distortionwas essentially the same for each chan-nel, so the right channel is presentedhere. There is a very faint thump duringpower-up and no noise at all at shut-down. The IC brings itself to life oncethe supply rails exceed ±7.3V DC.

The LM3875 amplifier does not in-vert polarity. Input impedance was100k at 1kHz. The gain at 2.83V RMSoutput into 4Ω and 8Ω loads was 26.3dBand 26.4dB, respectively. The output im-pedance at 1kHz was 0.072Ω, increas-ing slightly to 0.078Ω at 20kHz.

The frequency response (Fig. 10) waswithin ±1dB from 8Hz to 41kHz, at anoutput of 2.83V RMS at 1kHz into 8Ω.When I connected a load of 8Ω paral-leled with a 2µF cap (a test of compati-


T-Series contemporary styled horn systems . These systemssound truly lifelike. They achieve remarkable performance

from low or high powered amplification.

CLASSIC AUDIO REPRODUCTIONS

Established 1986 • Ph. (810) 229-5191Fax 810-229-5969, Email: [email protected]

Web: www.classicaudiorepro.com

Contact:

bility with electrostatic speakers), thegain became extended at the high fre-quencies, peaking at 46kHz (dashedline). As I mentioned earlier, you canleave out R108 and R208 and enjoy theextra efficiency.

The IHF load, which simulates aloudspeaker impedance peak at 50Hz,produced a 0.2dB higher response atthis frequency than the 8Ω resistiveload alone. The LM3875 amplifier willbe insensitive to variations in speakerimpedance with frequency.

Output hum and noise (input short-ed) measured 1.3mV, with −14mV DCoffset. Crosstalk performance was limit-ed by the close proximity of the twoaudio channels near S2 (Table 2).

THD+N versus frequency is shown inFig. 11 for the loads indicated at theright side of the graph. During distor-tion testing, I engaged the test-set80kHz low-pass filter to limit the out-of-band noise.

Figure 12 shows THD+N versus out-put power for various loads and frequen-cies (in the order of the list on the rightof the graph, when aligned with the 20Wvertical line). There was absolutely nostrain right up to the point of maximumpower. The amplifier reached its 1% clip-ping point at 32W for all loads. With the20Hz 4Ω load waveform on the scope, Icould see the SPiKe protection activat-ing intermittently.

I also plotted the 1kHz product of themulti-tone intermodulation (MIM) 9kHz+ 10.05kHz + 20kHz test signal versusoutput power (dashed line). This gives abetter indication of the LM3875’s non-linear response, since it is a closer ap-proximation to music than a sine wave.

The distortion residual waveform for10W into 8Ω at 1kHz is shown in Fig.

13. The upper waveform is the amplifieroutput signal, and the lower waveformis the monitor output (after the THDtest-set notch filter), not to scale. Thisdistortion residual signal just barelyshows the third harmonic, overlaidwith noise. THD+N at this test point is avery low 0.005%.

The spectrum of a 50Hz sine wave at10W into 8Ω is shown in Fig. 14, fromzero to 1.3kHz. The THD+N here mea-sures 0.0055%. The second, third, fourth,and fifth harmonics measure −100dB, −94dB, −104dB, and −92dB, respectively.Low-level power-supply rectification ar-tifacts are also present at 120Hz, 180Hz,and 240Hz, all below −90dB.

Figure 15 shows the amplifier outputspectrum reproducing a combined19kHz + 20kHz CCIF intermodulationdistortion (IMD) signal at 12V pp into8Ω. The 1kHz IMD product is 0.012%.Repeating the test with a multi-toneIMD signal (9kHz + 10.05kHz + 20kHz,shown in Fig. 16) resulted in a 1kHzproduct of 0.011%. (I plotted MIM distor-tion versus output in Fig. 12.) This fineperformance did not change when I re-duced the load to 4Ω.

A 2.5V p-p square wave into 8Ω at40Hz showed the expected LF tilt. The1kHz square wave was just about per-fect, while the edge of the 10kHz squarewave was slightly rounded. When I con-nected 2µF in parallel with the 8Ω load,the leading edge of the 10kHz squarewave showed underdamped ringing atthe 46kHz peak measured in the fre-quency-response test.


The British specialistsin tube amplifiers andpre-amplifier kits, loudspeaker kits andrelated publicationsVisit our informative website:www.worldaudiodesign.co.ukEnter our HD83 competition on-line

KiT88 integrated amplifier kit

World Audio Publishing Ltd.12a Spring Gardens.

Newport Pagnell.Milton Keynes.

MK16 0EE. England

tel/fax: 00 44 1908 218836e-mail:[email protected]

Kel84 integrated amplifier kit

Series II modular pre-amplifier kit

300b PSE monobloc kit

REFERENCES1. The QuadPod project is presented in The Joy ofAudio Electronics, C. Hansen, Chapter 2, Old Colony,BKAA52.

2. “A Basic 50W Stereo System, Part 3: A Solid-StateAmplifier,” M. Gustafson, GA 4/00, pp. 42–51.

3. My thanks to Bob Pease, Staff Scientist at NSC, forhis help in tracing the history of the LM3875 designand for answering my technical questions.

4. “Quasi-Complementary Output Stage,” pp. 172–173,High-Power Audio Amplifier Construction Manual, G.Randy Slone, Old Colony, BKMH29. “Quasi-Comple-mentary Outputs,” pp. 111–113, Audio Power AmplifierDesign Handbook, Douglas Self, Old Colony, BKB43.

5. Reference Data for Radio Engineers, Sixth Edition,Howard Sams, Fig. 1, p. 6-1.

6. “Calculating Optimum Snubbers,” J. Hagerman, AE1/98, pp. 26–31.

7. “Upgrading Monarchy’s A/D Converters and Ampli-fiers,” G. Galo, AE 5/00, pp. 8–11.

VISIT OUR WEBSITE ATwww.audioXpress.com

OR E-MAIL US [email protected]

Aresponse to my articles in GA2/99 and 3/99 covering a differ-ent way of driving low-mu tri-odes got me thinking—about

using a Circlotron amplifier. I havenever had experience with this configu-ration and was curious as to why itwould be used.

CIRCUIT FEATURESThe Circlotron has some important ad-vantages. The power tubes are connect-ed in such a way to drive the load in par-allel rather than in series, as they wouldin an ordinary push-pull circuit (Fig. 1).

For a given power output in watts,one-half the voltage at twice the currentwill produce similar results. If you dothe arithmetic with Ohm’s Law (R =E/I), you get the following result:

R (load) = 0.5E / 2I = (¼) #R

This would let you use an outputmatching transformer having one-quar-ter the impedance normally used in apush-pull amplifier. That transformerwill produce good results much easierthan a high-impedance unit.

If you apply this to an amplifierusing the 6AS7/6080 family, you woulduse a matching transformer of 1kΩrather than 4kΩ. Since the load is in thecathode circuit, the resulting output im-pedance (not the same as the matchingimpedance) will also be very low. In acathode follower, output impedance isreduced by the factor:

1 / µ + 1

Again using the 6AS7/6080, the pub-lished plate impedance is 280Ω, giving

you a source impedance of 93Ω as acathode follower. Since you have a pairof these driving the load in parallel, theresulting source impedance becomes46.5Ω.

The damping factor (DF) is the loadimpedance divided by the source im-pedance. In this example, it is 1000/

46.5, or 21.5, which is a very high resultif you take into account that you havenot applied overall feedback. However,that is not the complete picture.

The output transformer will havesome resistance in both its primary andsecondary. That will somewhat in-crease the source impedance as seenby the load. In a practical circuit youcould expect a damping factor of about10 without external feedback. Other fac-tors influencing the DF factor will bewhether you use triodes or pentodesand what kind of driving scheme is ap-

As a follow-up to a discussion of driving low-mu triodes, this author

investigates the Circlotron amp configuration. By John L. Stewart

The Circlotron Amp


FIGURE 1: Circlotron power stage.

FIGURE 2: Circlotron power and driver stages.

plied. In recent work, I’ve found thefinal DF could be as low as two withoutexternal feedback.

Another advantage of this circuit isthat it is completely symmetrical. Youmight counter that so is an ordinarypush-pull circuit, and it is. However,you can match the Circlotron to a rela-tively low output impedance at close-to-ground potential.

You now have the possibility of usinga number of power tubes in a push-pullparallel combination to drive evenlower impedance loads. That makes itpossible to drive a loudspeaker load di-rectly without the help of an outputtransformer. You’ve avoided the asym-metry associated with the series push-pull (Futterman)1 system.

WHY DON’T WE ALL USE IT?With all these advantages, you wouldthink that everyone would build vacu-um tube amplifiers using this greattopology. However, there are someproblems.

First of all, you have probably noticedby now that I have shown two high-volt-age power supplies in the simplifiedschematic. That’s not a mistake. If youwere to run a pair of the 6AS7/6080s inthis circuit, you would need two of the375V power supplies electrically isolat-ed from each other (Fig. 2).

Another problem is how to provideenough driving voltage to the outputstage. After all, it is connected as a cath-ode follower and has no voltage gain.The power gain is the same as it wouldbe if the load were connected in theplate circuit. In practice you will need adrive of more than 2× the expected out-put voltage at the cathodes.

The driver stage is normally assistedby bootstrapping. Figure 3A is a simu-lation of a bootstrapped circuit usingsoftware by Electronic Workbench. Iused a 12AT7 in the simulation to pro-vide drive to the output tubes. It is con-nected as a differential amplifier withits cathode resistor returned to a 150Vnegative power supply, showing that adrive of 413.4V is required grid to gridof the output tubes to get 16W. That is11.39V in 8Ω. The oscilloscope connect-ed to one of the output tube grid leadshas just started to show some clipping(Fig. 3B).


FIGURE 3A: Basic Circlotron for 6AS7/6080 with bootstrapping.

FIGURE 3B: Scope display of Fig. 3A.

Figure 4A shows the same circuitsimulation, but this time the bootstrap-ping is unhooked. The 158.0V of signaldeveloped between the output tubecathodes in the bootstrapped circuit isnot available to assist in driving theoutput grids. There is much clipping(Fig. 4B).

To get results similar to those in thebootstrapped circuit, you would proba-bly need a 500V or more power supplyfor the driver. This is a good exampleof transferring the problems in an am-plifier from the output stage to thedriver.

BOOTSTRAPPINGSome audiophiles are concerned thatbootstrapping, because it is positivefeedback, will cause problems for theircircuit. They may be surprised to dis-cover that one of the benefits of the cir-cuit is the same as that found in themu-follower. The bootstrap connectionincreases the load impedance seen bythe driver stage in proportion to gainin the output stage. That has the sameeffect in lowering second harmonic

distortion in the driver as it does in themu-follower.

In this example of the Circlotron, thedriver stage uses 27k loads. Because ofthe bootstrapping the load they will seeis 1.62 times (44k) that. In the McIn-tosh the driver sees about 4× the loadresistors used. I measured the effect ofthe bootstrapping in my version of itsuse and found one-half of the requireddrive was provided by the bootstrap. Inthat case the load resistors for the driv-er were 27k and would appear to be 54kto the driver.

The most important advantage of thebootstrapped driver is that it has amuch larger output voltage capability.Without the bootstrap you would needto build a special high-voltage driver.That is why it is used by McIntosh and

(to page 87)


REFERENCESAudio Anthology #5, Realistic Audio Engineering Phi-losophy, Norman Crowhurst

Electronic Workbench, Interactive Image Technolo-gies, 111 Peter Street, #801, Toronto, Ontario, CanadaM5V 2H1, http://www.interactiv.com.

FIGURE 4: Basic Circlotron for 6AS7/6080 without bootstrapping.

FIGURE 4B: Scope display of Fig. 4A.

Distributor Lundahl highquality double C core tube transformers

LL1627SE/PP 6C33/ 4 pcs EL34LL1623SE/PP 300B/2A3/KT88LL1620SE/PP 845/211 application

SE in 40/60/80/90/120/160/240 mA 125$

These transformers are beautifully winded with 4 primary sections and 8 sec-ondarys they can stand 4 kV and handle 250watts PP or 50watts SE

LL 1663/1664 SE/PP 5K/ 8 3K/ 8ohm 100$/p

Beautifull black pot for trannys 35us$/p

Mnf of kits specialty monoblokKit 300B/2A3 999us$Kit 845SE 2499us$Kit 845PP 2000us$Kit KT88/6550 750us$

Vellemankits price on request CheapPaper in oil caps: available4, 10, 15, 32µF all 630 volt 4 µF 850 volts 17, 5us$/p4, 6.3, 20, 50µF / 2kV

Krone tubes: identical to new modelsVV300, VV302, VV52BXStarting from 200us$ per matched pair

Dale 1W & 5W 1% non magnetic preci-sion resistors from 0, 5us$/p

Chokes 20H/80mA or 7H/160mA25us$ 30H/20mA 18, 5us$

Round 0-150mA, 0-500volt meter 20us$Current regulating diodes 0,5-5mA5us$

Also lots of speaker designs and partsAudax Focal Peerless ACR Thiel

We accept major credit cards

Main address: headquarter

Boomgaardstr 264 / 2600 Antwerp belgium Tel 011 323 281 3244

Address USA: 23900 Chagrin blvdCleveland OHIO 44122Tel 216 470 3167 (message box)Email [email protected] www.diyparadiso.com

With the exception of speak-er cables, perhaps no otheraudio component carriesmore baggage than cabinet

stuffing. Part of the confusion stemsfrom a misunderstanding of various re-quirements for different and diverseapplications. Some materials workwell in most cases, while others pro-vide more psychological comfort thanacoustical benefit.

Likewise, costs, manufacturability,and stability must be considered. Sort-ing wisdom from witchcraft starts withunderstanding the various benefits ofbox stuffing, which can vary with en-closure type. When the correct type ofstuffing is selected and properly ap-plied, it can slow down the speed ofsound, increase the acoustic compli-ance, and attenuate the midbandwhile enhancing the low frequencies(Figs. 1 and 2).

WHAT’S GOING ON BEHIND THATSPEAKER CONE?About two years ago, while working ona project at the Dai-Ichi speaker plantin Manila, our associate, Mike Klasco,was asked to comment on the tweeterimpulse-response printouts from anewly constructed tweeter test box. Itwas apparent that the ragged, extendeddecay time could not possibly be attrib-uted to the tweeter, a balanced-driveunit of advanced construction. Mikeknew the test mike had superior im-pulse response, as does the Bell Labs

SYSid test and mea-surement system. So,new test box’s acrylic-fluff “anechoic stuff-ing” was suspect.

Noticing the pinkbuilding insulationcovering the ceiling ofthe plant, Mike had theacrylic box stuffing re-placed with that. After installing thefiberglass in the test box and re-mea-suring the tweeter, Mike saw that thelong decay time from the previousplots was gone. It had merely been anartifact of the acrylic fluff’s poormidrange absorption!Imagining that the simulated response

of his design is adequate, the rookiespeaker engineer, armed only with alow-frequency modeling program,might be tempted to omit box stuffingaltogether. After all, no one sees thisstuff anyway! Veteran engineers knowbetter. Midrange roughness and im-pulse response signatures result whenyou omit enclosure stuffing.

If half the energy coming off a speak-er’s cone is emanating from its back-side, the “back wave” energy must gosomewhere. Though enclosures suchas the bass reflex utilize the low-fre-quency portion of the “back wave” com-ponent constructively, enclosures suchas the infinite baffle just wish it wouldall go away. In any case, box modesmust be suppressed, a necessity that isincreasingly overlooked.

At Menlo Scientific we are often con-tracted to do audio design for con-sumer electronics. A recent undertak-ing was a stand-alone Internet radio (nocomputer required), from which the

best possible sound quality per dollarwas desired. Although we’d specifiedbox stuffing in the design, we caughtthe Chinese subcontract factory leavingit out. While the cost savings was negli-gible, the performance degradation wasdramatic! The omission of box stuffing in less-ex-

pensive, compact audio products is adisturbing trend, indicating a clear mis-understanding of the importance of thiscomponent. While enclosure stuffing isgenerally understood to influence thelow-end response, its effects onmidrange clarity aren’t as widely appre-ciated. Especially in the case of smallsatellite enclosures (with wooferscrossed over high into the midrange), alot of midrange energy is dumped fromthe back of the cone.

THE FUNCTION OF BOX STUFFINGFOR VARIOUS TYPES OF ENCLOSURESSo, which stuffing is best for whichapplication? Let’s look at enclosure design requirements, survey the de-signer’s options, and discuss price andperformance.

SEALED ENCLOSURESAbsorptive stuffing in a sealed boxserves a number of functions. It in-creases the acoustical compliance, low-

The function of stuffing in your speaker cabinet depends

on application and box design. By Ron Horowitz and

John Schaffer

Box Stuffing–A Fistful of Fuzz


ABOUT THE AUTHORSRon Horowitz and John Schaffer are employed atMenlo Scientific, Ltd., which serves as a consultant toa number of loudspeaker firms.

PHOTO 1: AR’s famous “Bee Butt” midrange.

ers the woofer’s in-box resonance, addsresistive damping, and lowers the sys-tem’s Q. Absorption at higher frequen-cies suppresses box modes, smoothingupper frequency and transient respons-es. To a point, increasing stuffing densi-ty enhances all three effects. Since stuffing costs can become signif-

icant with large acoustic suspensiondesigns, it’s left to the engineer to justi-fy the cost. All too often, out of sight is

out of mind as far as the corporate“bean counters” are concerned. On theother hand, accountants don’t knowanything about capacitors and induc-tors, either, and I understand a fewspeakers have been made with those! Itmust be made known that adequatedamping is essential to optimal systemperformance, so stuffing is not a com-ponent to be omitted at the whim of acorporate know-nothing.

PORTED AND PASSIVE RADIATOR ENCLOSURESGenerally, ported boxes are merelylined with an absorptive boundary layerto help dissipate midrange energy andsuppress box modes. Following this ap-proach, low-frequency absorption isnegligible. Occasionally, one portion ofa ported enclosure is lined, while an-other portion is stuffed.

With this approach, the acoustical


TD–15 15” LF Driver

Many other models available...Complete specs and

product information online at www.lambdacoustics.com

UNITY HORN

The entirely new Unity multi-driver

approach to HF horn design yields

super low distortion, extreme high-

efficiency, and a natural presenta-

tion quality that has to be heard to

be believed.

With the patent-pending Unity

design, traditional horn colorations

are no longer an issue! Visit our

website for detailed technical info

and application notes.

Speakers always were the hardest

part of building a good tube audio system…

PO Box 971, Melbourne, FL 32902

321-751-9750 • 321-751-9751 fax

96 dB@1W, 8 ohms

20 Hz-4kHz

Copper Faraday motor

Pro-grade paper cone

Explosive dynamics!

20mm+ Xmax!

Handcrafted in the USAWe’re making it easy.

110 dB/1W, 8 nominal

300-22000 Hz +/- 3 dB

LAMBDA ACOUSTICS

FIGURE 1: Frequency response without stuffing (solid) and withstuffing (dash).

FIGURE 2: Impedance without stuffing (solid) and with stuffing(dash).

compliance of the enclosure can be in-creased, though you may observe low-frequency damping effects. Regard-less, unlike the obsolete aperiodic en-closure type, stuffing should not befoulthe port and upset the port tuning.Since the largest possible port area isdesired to keep port velocities low, itappears that modern, linear-compli-ance passive-radiator systems mighthave a slight advantage here, especial-ly for two-way systems using a highcrossover point.

BANDPASSDesigners of bandpass enclosuresoften omit absorptive materials alto-gether, not to mention crossovers, pre-suming the acoustical filter nature ofthe design will suffice. While this expe-dient has advantages in low-cost de-signs, nothing precludes the judiciousapplication of box stuffing in a band-pass design. Ported sections can belined, whereas compliance and damp-ing effects should be considered insealed bandpass sections.

ACOUSTIC LABYRINTHS AND TRANSMISSION LINESThe function of stuffing in a transmis-sion line is two-fold: it lowers the speedof sound, and it absorbs midrange en-ergy. By slowing the transmissionspeed, a line can be mechanically short-ened, while retaining the same tuningfrequency. Ideally, the stuffing wouldpass low-frequency energy, which canbe used to enhance low-frequency out-put while suppressing box modes atupper frequencies. A boundary layer ofstuffing lining the walls provides thisfunction, though relatively little slowingof the transmission speed occurs, ascompared to a fully stuffed line.

BASS HORNSThe back chamber of most bass hornscan be treated like a sealed box, provid-ing loading below the horn cut-off. (Inthe horn’s bandpass range, the rearchamber has other interactions.) In hisPataxial (coaxial horns) loudspeaker, Dr.Eugene Patronis of Georgia Tech linedthe walls of the low-frequency horn withSonex, and, while developing hyperbolic(high Q) bass horns in the 1970s,Richard Long found that an effectivemeans of damping resonances was toplace a little bit of stuffing in the horn’sthroat. Of course, if the air column in along, slowly expanding hyperbolic hornis stiffer than the diaphragm of its driv-er, the best place to put the acousticalstuffing may be in the listener’s ears!

TWEETERS AND MIDRANGE DRIVERSIn order to smooth the frequency andtime responses of dome-type mid-range and tweeter units, the space be-hind the dome is often filled withstuffing. With this application, in ad-dition to acoustical performance, youshould consider mechanical and ther-mal issues. The stuffing will do moreharm than good if it shifts out of posi-tion, or worse yet, if it melts or burnsunder high-power operation. Fortu-nately, because the volume of stuffingused is so small, material cost is ofminor importance.

Though currently out of vogue, resis-tive damping of dome-type midrangedrivers (achieved by positioning stuff-ing in front of the dome using an ex-panded metal cover) has been used in

a number of commercially successfulproducts. The Acoustic Research “bee-butt” midrange (Photo 1) drivers (so-called because they resembled a bee’sbutt), as used on the AR3 and the morerecent AR 303, are examples of this ap-proach. A remarkable smoothing of the response is the primary be-nefit, and cosmetic and mechanicalconcerns are the primary drawbacks. This method works well with conemidrange units with B&W’s Nautilusbeing a shining example.

TYPES OF STUFFING CHOICESThere are two main categories of stuff-ing, foams and fibers. Each type has ad-vantages and disadvantages.

FOAMSOpen-cell polyether and polyesterfoams have similar properties, in whichfoam density and cell structure are re-lated to acoustical performance. Theprimary advantage of foam is ease ofhandling. Foams can be cut to order(hot wire or blade) in fixed, non-friablegeometries, making for ease of ware-housing, inventory, handling, assem-bly, and so forth.

Foams tend to lack low-frequency ab-sorption compared to fibers, though,and are of limited benefit in sealed en-closures at low frequencies. They candeteriorate when exposed to UV, fun-gus, ozone, and the like, but these don’ttend to be issues in sealed or passive ra-diator systems, where foam surroundsmay go first! Small, die-cut foam but-tons are often used under the di-aphragms of dome tweeters andmidrange drivers, where a non-friable,stable geometry is desired.

FIBERSThe major fiber groups used inacoustics include acrylics, wools, andglass fibers. Fiber length, texture, diam-eter, flexibility, and packing density arethe major acoustical variables for agiven fiber type. Optimal fiber geome-tries give best results, though such at-tention to detail increases costs overgeneric types.

Wool is an expensive fiber. On theother hand, when properly applied, itdoes have excellent acoustical charac-teristics. This is the “cost no object”


INSTOCKNOW!

•• MORE PAGES

•• MORE GRAPHS

•• MORE REFERENCES

•• BRAND NEW CHAPTERS

•• DETAILED EXPLORATIONS

BKAA60$39.95

Shipping and handling add $7.25, Canadaadd $11.00

Overseas surface mail add $10.25, air mailadd $22.25.

Order Your Copy TODAYby calling 1-888-924-9465

or order on-line atwww.audioXpress.com

Old Colony Sound Laboratory, PO Box 876 Dept. X1, Peterborough, NH 03458-0876 USA

Phone: 603-924-9464 Fax: 603-924-9467 E-mail: [email protected]

choice for the ultimate audiophile, JonRisch of Peavey Electronics. While the long-fiber wool of the female

sheep of the Outer Hebrides is thehands-down cult favorite (sheared inlate autumn after a hard rain), it hasmany practical limitations. The fibersbunch up, it’s difficult to get consistentpacking density and volume, it’s dimen-sionally unstable, and so on. As a prac-tical stuffing material, it’s a non-starter.Besides, who wants to make cedar-linedspeaker enclosures?

Fiberglass is a low- to medium-pricedfiber with good to excellent perfor-mance. Fiberglass, specified in thick-ness and density, is generally used pre-cut in sheet form. While semi-rigidfiberglass is available, it generally lacksthe non-friability of foam. Glass fibersare fairly brittle, especially when largecross-section fibers are present, and,unlike wool or acrylic fibers, are proneto fracturing.

As anyone who’s ever worked withfiberglass can attest, its primary disad-vantage is adverse worker reactionssuch as itching and allergic reaction.

One of the better low-cost, “reduceditch” fiberglass types available in Asiais Owens Corning’s TIW-20. This brown-ish fiber serves double duty as an oveninsulation! In the US, Owens Corning’s"Think Pink" thermal insulation issometimes used, as is its yellow coun-terpart, which can be ordered in vari-ous densities and semi-rigid sheets.

Acrylics are medium-priced fiberswith good acoustical performance, de-pending largely on fiber geometry andpacking density. Unlike fiberglass,acrylic fibers are stable and non-aller-genic, with no special handling issues.Acrylics come in sheets for lining, or inbulk for filling, and occasionally insmall buttons of acrylic fiber for back-of-dome applications.

Select Sound Acoustic Core is a rela-tively new fiber technology from OwensCorning. Acoustic Core is an engineeredglass fiber product that specifically ad-dresses the issues of high acousticalperformance. Select Sound fibers weredesigned using two different fiber typeswrapped around each other with a ran-dom twist and a curved fiber shape.

After Mr. Risch’s explanation of thelong-fiber wool, you could almost de-scribe Acoustic Core as a synthetic ver-sion of the wool. The fiber diameterswere selected to afford a “non-stinging,”reduced-itch fiber. Likewise, the resinwas selected to minimize allergic reaction.

Select Sound has superior acousticalcharacteristics at a fraction of the cost oflong-fiber wool. It is an engineered prod-uct, so it’s consistent in its critical pa-rameters and availability, unlike wool,which was designed originally to keepsheep warm. Select Sound is fairly re-silient and springy, too, so it doesn’ttend to bunch up as much as long-fiberwool, and it’s available as a white, un-faced, loose fiber in banded, 275 lb bales.

CONCLUSIONOften regarded with confusion or indif-ference, box stuffing is an essential, ifmisunderstood, component. Under-standing various applications, availablematerials, and necessary functions isthe first step toward superior productdesign.


State-of-the-art speaker enclosure designsoftware used around the globe by profes-sionals & amateurs for worldclass results.

Also available: BassBox Lite, a streamlined version ofBassBox Pro without a driver database for only $69.Visit our website at www.ht–audio.com for details.

Versatile passive circuit design software forcrossover networks, LP, BP, HP filters, driver

impedance EQ, L-pads and notch filters.

Includesthe

WORLD'SLARGEST

DriverDatabase

Design Loudspeaker Boxes Design Crossover Networks

Harris Technologies, Inc.Post Office Box 622

Edwardsburg, MI 49112-0622 U.S.A.

• Easy-to-use Design Wizard.• Enormous driver database.• Box types: closed, vented, B6 vented, bandpass and PR.• Imports acoustical data.• Nine performance graphs.

• Beautiful 364-page manual.• Plot vent "pipe" resonance.• Box drawings & parts lists.• Driver & PR test procedure.• Multi-driver/isobaric boxes.Price: only $129 (plus S&H).

• Design 2-way & 3-way passive crossover networks, separate LP, BP, HP filters and L-pads. • 1st, 2nd, 3rd and 4th-orders. • Driver impedance EQ (Zobel). • Enormous driver database. • Four performance graphs.

• Beautiful 243-page manual.• Schematics & parts lists.• Include box response.• Estimate ESR & DCR.• Parallel-series calculator.• Notch filter design tool.Price: only $99 (plus S&H).

Typical shipping & handling (S&H) fees for BassBox Pro & X•over Pro are: $9 in the U.S. (UPS), $11 to Canada (Airmail) & $35 elsewhere (G-Priority).

Copyright © 2001 by Harris Technologies, Inc. All rights reserved worldwide. BassBox is a trademark of Harris Tech. Other trademarks belong to their respective companies. Harris Tech reserves the right to make changes without notice.

Tel: 616-641-5924Fax: 616-641-5738

Email: sales@ht–audio.com

TM

All Harris Techsoftware is design-ed for Windows 95,98, Me, 2000, NT4.

Save $29 andbuy BOTH BassBoxPro and X•over Pro

for only $199.

For moreinformation

visit us on theworldwide web at

www.ht–audio.com.

On the surface, the function ofwire seems pretty simple: getelectrons from point A topoint B. But there are many

choices to be made when wiring audioequipment. A multitude of differentwire types, sizes, and materials is avail-able to the builder.

THE SINGLE CONDUCTORSingle-conductor wire is what most peo-ple think of as “normal” wire. The sin-gle metal conductor can be made of onepiece of metal called “solid” wire, orfrom many smaller wires wrapped to-gether, which is called “stranded” wire.Single-conductor wire may be baremetal (un-insulated) or covered in anonconductive material. (I discuss thematerials used for the conductor andinsulation in more detail later.)

Electrically, there is not much differ-ence between solid and stranded wire.The main difference is mechanical.Solid wire is, well, solid, while strandedwire is more flexible. Photo 1 showssolid and stranded wires with the insu-lation removed.

In old-fashioned point-to-pointwiring, like that used in tube equip-ment, solid wire is generally used. It iseasier to terminate to solder lugs (likethose on tube sockets) since it doesn’tfray out into a million little wires, atleast one of which will never go into

the hole you choose. Solid wire alsoholds its shape, so you can dress it intoposition and it will stay there.

On the other hand, stranded wire ismore flexible, especially in larger sizes.It also survives being bent back andforth without breaking, so any wiringthat involves motion should be donewith stranded wire, which is also usedwhen the wire is to be terminated intocrimp or insulation displacement (IDC)connections. Solid wire doesn’t general-ly make a reliable connection this wayexcept in very small wire sizes and withconnectors specially designed for it(such as telephone connectors).

A typical piece of audio equipmentmay use both solid and stranded wire.For example, AC power wiring is usual-ly done with stranded wire, since it ismore flexible in the larger size re-quired for AC wiring, and can be termi-nated with crimp terminals. Point-to-point wiring, such as between tube

This article discusses some of the different types of wire commonly

used in audio equipment, and provides some practical advice about

when to use each type. By Pete Millett

All About Wire


ABOUT THE AUTHORPete Millett started playing with tubes as a kid, and hasbeen an electronic design engineer for about 20 years.His experience ranges from consumer electronics tocomputer hardware and system design. He also runsWheatfield Audio, a small manufacturer of tube head-phone amplifiers. Pete can be contacted by e-mail [email protected].

TABLE 1COMMON WIRE SIZES AND THEIR CHARACTERISTICS

AWG DIA. AREA RESISTANCE RESISTANCE MAX CURRENT,(INCHES) (SQ. INCHES) (MILLIOHMS/ (MILLIOHMS/ AMPS (BARE

FOOT, COPPER FOOT, SILVER COPPER WIRE)20°°C) 20°°C)

4 0.2043 0.03278 0.2485 0.229 1996 0.162 0.02061 0.3952 0.364 1258 0.1285 0.01297 0.6281 0.579 7910 0.1019 0.00816 0.9988 0.921 49.612 0.0808 0.00513 1.59 1.46 31.214 0.0641 0.00323 2.52 2.33 19.616 0.0508 0.00203 4.02 3.7 12.318 0.0403 0.00128 6.39 5.89 7.7520 0.032 0.000804 10.1 9.34 4.8922 0.0253 0.000503 16.2 14.9 3.0424 0.0201 0.000317 25.7 23.7 1.9326 0.0159 0.000199 41 37.8 1.228 0.0126 0.000125 65.3 60.2 0.76130 0.01 0.0000785 103 95.6 0.47732 0.008 0.0000503 162 149 0.304

PHOTO 1: Solid (left) and stranded wirewith the insulation removed.

PHOTO 2: Different wire sizes. From left:12AWG, 16AWG, 24AWG, 30AWG.

sockets or terminal strips, may usesolid insulated wire.

For under-chassis wiring inside atypical audio project, the choice be-tween solid and stranded wire is mostlya matter of builder preference. Eitherone works fine.

WIRE SIZEWire is manufactured in sizes as thinas a hair to several inches in diameter.The size affects its electrical character-istics, such as resistance and ability topass large currents, as well as its me-chanical characteristics. Smaller wirehas a higher electrical resistance, andvery small wire is difficult to work with.

At the other extreme, largewires take up space, are dif-ficult to bend and fit intochassis, and are expensive.

Wire size is commonlyspecified with a numbercalled “gauge.” “AmericanWire Gauge,” or “AWG,” isthe standard normallyused for electronic wiring.The larger the number, the

smaller the wire. Table 1 shows the di-ameter and characteristics of a rangeof wire sizes that covers most of whatis used in wiring normal electronicequipment. Note that commonly avail-able wire sizes are only in even num-bers; the odd-numbered sizes do exist,but are not commonly found. Photo 2shows differently sized wires for comparison.

Two factors influence what size ofwire you use inside a piece of electron-ic equipment. Electrically, a wire needsto be large enough so that its resistanceis low enough not to influence the cir-cuit or be heated significantly by thepower dissipated in the resistance. Me-

chanically, you need a wire that is nei-ther so small that it’s fragile and hardto terminate, nor so big that it is diffi-cult to bend into shape.

For low-level signals, where the cur-rent is measured in milliamps, the wiresize is chosen more for ease of use thanfor electrical characteristics. Usually, asize between 22AWG and 28AWG ischosen somewhat arbitrarily.

For power wiring (both power-supplyand high-level signals like speakerwiring), you must choose the wire sizebased on the amount of current thatmust be passed. There are many rules-of-thumb for determining the currentcapacity of a given wire size—somebased on temperature rise, others onvoltage drop. Generally, it is best to beconservative and use a wire size that isa couple of sizes bigger than what is ab-solutely necessary.

The maximum currents listed inTable 1 are based on temperature riseof bare wire. Consider these absolutemaximums for wiring in audio equip-ment, and it would be best to use wiretwo sizes larger than this minimum.


TABLE 2COMMON AUDIO EQUIPMENT APPLICATIONS

OF DIFFERENT WIRE SIZES

WIRE SIZE (AWG) APPLICATION24–26 Low-level audio signals, power wiring <500mA20–22 Power wiring <1.5A16–18 AC line wiring (usually stranded) <5A,

speaker wiring <40W, power wiring <5A12–14 AC line wiring (usually stranded) <10A,

speaker wiring >40W, power wiring <10A

CT101 key specificationsGain (selectable) 0, 6 or 12 dB

25 MHzSlew rate (at 0dB gain) 500 V/uSS/N ratio (IHF A) 112 dBTHD 0.0002 %Output resistance 0.1 ohmChannel matching ± 0.05 dBPCB dimensions: 100 x 34 mm

3.97 x 1.35 "

General attenuator specificationsNumber of steps: 24Bandwidth (10kOhm): 50 MHzTHD: 0.0001 %Attenuation accuracy: ±0.05 dBChannel matching: ±0.05 dBMechanical life, min. 25,000 cycles

Fax: (+66) 2 260 6071 E-mail: [email protected]

gwith a stereo CT1 attenuator added.

CT100 key specificationsGain (selectable): 40 to 80 dBRIAA eq. deviation: ± 0.05 dB S/N ratio (40/80dB gain): 98/71 dB THD: 0.0003 %Output resistance: 0.1 ohmChannel separation: 120 dBBandwidth: 2 MHzPCB dimensions: 105 x 63 mm

4.17 x 2.5 "

CT2 6-gangvolume control for A/V Audio

Table 2 lists some common wiresizes used for hook-up wire in audioequipment.

CONDUCTOR MATERIALSAlmost any metal can be used to con-struct wire, as can combinations ofmore than one metal. For example,wire designed to be strung betweentelephone poles sometimes uses asteel core for strength, with copper lay-ered around it for good electrical con-ductivity. In the wiring of electronicequipment, though, the electrical char-acteristics of the material are more im-portant than mechanical characteris-tics, so they determine what metal isused.

Of the commonly available metals,the best electrical conductor is silver,followed by copper and gold. Gold is fartoo expensive to consider for electronicwiring—you’d need to be a very wealthyperson (not to mention one in need ofpsychiatric help) to wire a project usinggold wire!

Silver wire is sometimes used inhigh-end audio projects. Although it isa better conductor than copper, its re-sistance is only about 8% less than cop-per, which is much less expensive.Even though from an engineering per-spective there is no reason to use silveras a wiring material, many audiophilesswear by it (see sidebar).

So, copper is the material of choice

for almost all electronic wiring. Howev-er, there are different types of copper aswell!

Pure copper is a very soft metal, so itis often alloyed with other metals toalter its mechanical properties. There isalso always some level of impurities leftin the metal, so even “pure” copper hassome amount of other metals—and non-metallic materials—in it. As far as theelectrical properties are concerned,purer copper is better.

There has been much discussionabout “ultra-pure” and “oxygen-free”copper being needed to wire audioequipment. Certainly, impurities suchas copper oxides are detrimental to theelectrical properties of the wire, butthey are very difficult to quantify.

Copper’s problem is that it reactswith air and forms an oxide layer on itssurface, which turns the shiny coppercolor to a dull brown. Copper oxide isnot a good conductor, and it can makeconnections and solder joints work

poorly. Because of this, most wire usedin electronics has another metal—usu-ally either tin or silver—plated on theouter surface of the wire. This platingdramatically improves the solderabili-ty of the wire and reduces the chanceof a connection becoming open overtime. Unless you have a specific reasonto use unplated wire, you should al-ways use wire that has tin or silverplating.

INSULATIONMost wire used in electronic equipmenthas a coating over the metal to insulateit, or to prevent it from making electri-cal contact with other wires, a chassis,or you!

There are applications that use unin-sulated wire. Short connections, suchas between adjacent lugs on a tubesocket, are often done with uninsulatedwire, which is fine, as long as the con-nection is short enough that there is nodanger of the wire bending and makingcontact where it’s not supposed to.Likewise, grounded connections, suchas bussing together the grounded sideof a bunch of phono jacks, are oftendone with uninsulated wire, which iscommonly called “bus” wire.

Many different materials are used toinsulate electronic wire. The two basicparameters to consider in an insulationmaterial are its breakdown voltage(how high a voltage the insulation canwithstand before failing), and its tem-perature rating (how hot the insulationcan become before it melts or other-wise breaks down).

Various plastic and rubber materialshave been used for hook-up wire inelectronic equipment over the years.Today, for electronic wiring, the mostcommonly available insulation materi-als are PVC (polyvinyl chloride) andTeflon . For special applications requir-


BLACK MAGICSome audiophiles seem to have a nearly neurotic obsession about wire. The science andengineering says, within reason, wire is wire. Yet, there is a proliferation of exotic materi-als and technologies: from silver hook-up wire and silver-wound transformers to long-grain ultra-pure oxygen-free copper wire with some exotic polymer insulation. So, what’sup with this?

Though I am an engineer by profession, I’ve learned not to dismiss all of this “blackmagic” as hallucination and marketing hype. As much as engineers and scientists hate toadmit it, there are things that can be heard but not measured.

Can I hear the difference between identical tube amps—one wired with silver hook-upwire at $20 per foot, and one with Radio Shack copper wire? Nope. But I’m not quite will-ing to say that nobody can. If there’s one thing I’ve learned in this business, it’s to keep anopen mind.

Things become even more complicated when you start looking at speaker wire and in-terconnect cables, which can cost you anywhere from a couple of bucks up to thousandsof dollars. There are certainly measurable and audible differences between cables, but, tomy ear, they don’t necessarily relate to the price tag.

I’ve heard so much discussion about how solid wire sounds so much better than strand-ed wire (something about “strand interaction?”), and how PVC insulated wire sounds bet-ter than Teflon (and vice-versa), and how you just have to use silver solder, that it makesmy head spin.

So how do you resolve all this? Well, you don’t. Build and buy what sounds good andmakes sense to you. Experiment and make your own decisions—don’t be swayed by themagazines and the marketeers.—PM

TABLE 3COMMON UL AND MIL WIRE TYPES

WIRE TYPE INSULATION APPLICATION

UL1007 PVC, 300V, 80°C, 0.016″ General use hook-up wireUL1015 PVC, 600V, 105°C, 0.031″ Higher voltages and temperatures—thicker insulation UL1429 Irradiated PVC, 600V, 105°C, Thinner, tougher insulation, resistant to solderingMIL-W-16878/1 0.010″UL1213 Teflon , 600V, 200°C, 0.010″ High reliability, high temperaturesMIL-W-16874/4MIL-W-16878/5 Teflon , 1000V, 200°C, 0.014″ High voltage, high temperatureMIL-W-76B PVC, 1000V, 80°C, 0.016″ High voltage

ing very high voltages, temperatures, orflexibility, there are many other materi-als available as well, but they are sel-dom used by the hobbyist.

The most common type of electronicwire, usually called “hook-up” wire, isnormally insulated with PVC, which isa good—not great—electrical insulator,and is inexpensive. It’s easy to workwith because it strips easily. Its biggestdownside is that it doesn’t withstandhigh temperatures well. It melts easilywhen soldering, so you must be careful(and quick) when soldering connec-tions. It comes in different thicknesses,which are rated for different voltages.Common PVC hook-up wire is rated for300 or 600V, which is adequate foreverything except some tube circuits.

The other type of wire used often inaudio equipment is Teflon insulatedwire. Teflon is a better electrical insu-lator than PVC and can withstandmuch higher temperatures. It is alsovery resistant to flame, and for thesereasons Teflon insulation has alwaysbeen the preferred material in high-reli-ability and military applications.

Teflon has a distinctive slippery feel toit, and is usually shiny, as opposed toPVC, which is duller.

Teflon insulated wire has a coupleof disadvantages. It tends to be expen-sive (though in the quantity you wouldneed for a typical audio project, the costis probably not too great a deterrent),and it is also harder to strip the insula-tion from the wire. Unless you use agood, sharp wire stripper, Teflon tendsnot to cut cleanly and to draw out inlong stringy pieces.

So, which one should you use? It re-ally comes down to personal prefer-ence. I like the fact that Teflon doesn’tmelt easily, so you can solder a connec-tion without worrying about the insula-tion, and you can run the insulationright up to the solder joint. But it’s hardto find Teflon wire in any type otherthan the silver-plated, stranded wirecommonly used in military applica-tions—if you prefer solid wire, it may bea bit more difficult to find.

Table 3 lists several commonly avail-able types of insulated wire and theirmilitary (MIL) and UL designations.

MULTI-CONDUCTOR CABLES AND SHIELDED CABLEMore than one insulated conductor canbe enclosed inside a common jacket toform a multiconductor cable. There aremany different types of cables, manyused in constructing audio equipment.

Shielded Wire and Coaxial CablesA single conductor can be surroundedby braided wires or metal foil to form acoaxial, shielded cable, which is usedin many audio interconnect cables. Thesignal travels through the center con-ductor, with the outer shield acting asthe return or ground conductor. Thegrounded outer shield helps preventnoise from being coupled into theaudio signal traveling within.

You should always use shielded cablefor very low-level audio signals, such asthose from a microphone or phono car-tridge, to prevent noise or hum frombeing coupled into the signal from near-by power cords, transformers, or othercomponents. You should also use it in-side equipment, such as in an amplifierchassis, if the signal cable must be rout-


ed near a power transformer or other po-tential source of noise.

When you use shielded cable, youneed to consider that there is signifi-cant capacitance between the centerconductor and the (normally grounded)shield. This can cause problems in sen-sitive circuits, such as phono cartridgeinputs. This capacitance can causedegradation of the high-frequency re-sponse of your system. For this reason,it is best to use only cable that is specif-ically designed for audio use, and toavoid cables designed for digital orradio-frequency used in audio circuits.

Braided shielded cables can be diffi-cult to terminate, since the shields areusually made up of many small wires.The best approach is to try to unbraid allthe tiny wires, and then twist them to-gether and tin them with solder beforetrying to connect them to anything (Fig.1). You can use heat-shrink tubing overthe twisted shield wires and over theend of the cable jacket for a neat appear-ance and to prevent stray shield wiresfrom causing a short circuit (Fig. 2).

Foil-shielded cables normally have awire running along the inside of thefoil, called a “drain” wire. When termi-nating this type of cable, you simply cutoff the foil shield and terminate thedrain wire as you would any other wire.

Another common technique is to use

a shielded cable with two conductorsinside. At one end (usually the outputof the cable) the shield is connected to-gether with one of the conductors toground, and at the other end the shieldis left unconnected. This prevents theflow of current through the shield,which can sometimes introduce noiseinto the signal inside.

Photo 3 shows some different shield-ed wires designed for audio use.

Multi-Conductor Cable and Twisted-Pair CableA multiconductor cable is made up ofseveral single conductors randomly ori-ented inside a common jacket. Twisted-pair cable is similar, except the wiresare arranged in pairs that are twistedaround each other. Multi-conductor ca-bles may also have an overall shieldaround all the wires, as describedabove.

Multiconductor cables are not oftenused in audio equipment, except in pro-

sound reinforcement and studio applica-tions, where many signals must be trans-ported. They are sometimes used be-tween components, such as between anamplifier and a separate power supply.

Twisted pairs are sometimes used—with or without a jacket surrounding thepair of wires—inside equipment. Twist-ing a signal wire together with a return(or grounded) wire helps reject somenoise, though it is not as effective as ashielded wire. Conversely, wires carry-ing AC power, like filament connec-tions, are often twisted together to pro-vide cancellation of the magnetic fieldemanating from the wires. This reducesthe coupling of AC hum into surround-ing signal circuits. [Tight twists can beachieved by clamping one end of a pairin a bench vise, twisting and solderingthe other ends of the pair together, andtwisting them with a bent nail in thechuck of an electric drill.—Ed.]


FIGURE 1: Cable termination (from Heathkit booklet #595-294).

FIGURE 2: Cable modification (fromHeathkit booklet #595-186).

PHOTO 3: Shielded cables for audio.


Adirehttp://www.adireaudio.com

All Electronicswww.allelectronics.com

Audio Consultinghttp://www.audio-consulting.ch

Divergent Technologywww.divertech.com

Linear Integratedwww.linearsystems.com

Marchand Electronicswww.marchandelec.com

McFeelyswww.mcfeelys.com

Solen, Inc.http://www.solen.ca

Speaker Citywww.speakercity.com

The Tube Storewww.thetubestore.com

http://www.plitron.com

http://www.partsexpress.com WE

BSI

TERE

SO

UR

CE

SADVERTISE YOUR WEBSITE HERE!ON THE AUDIOXPRESS WEBSITE RESOURCE PAGE

SEE BELOW FOR MORE INFORMATION

A great web and print combination−promoteyour website to the readers of audioXpress andlink them to your own website on-line at audioXpress.com!

ADVERTISERS: CONTACT NANCY AT 800-524-9464 or 603-924-7292 to find out more infor-mation about this great opportunity! Or, [email protected].

WEBSITE RESOURCE LISTINGS ARE ONLY AVAIL-ABLE TO PRINT ADVERTISERS IN audioXpressMAGAZINE. DON’T BE LEFT OUT!

8

WE

BSI

TERE

SO

UR

CE

S

http://www.tubesandmore.com http://www.atma-sphere.com http://www.netspace.or.jpl/~eifl/eifl/export/export2.htm

http://www.2000apex.com

http://www.upscaleaudio.comhttp://www.tubes.ruhttp://www.rell.com/audio/index_Aud.asp

http://www.sonicfrontiers.com/TPC

Audiencehttp://www.audience-av.com

Audio Advancementshttp://www.audioadvancements.com

Audio Consultinghttp://www.audio-consulting.ch

Cary Audio Designhttp://www.audioelectronicsupply.com

DH Labshttp://www.silversonic.com

Divergent Technologyhttp://www.divertech.com

http://www.plitron.com

Electronics for thosewho prefer Music

http://www.perpetualtechnologies.com

Danish Audio Connecthttp://www.DACT.comStepped audio attenuators, input selectors, phono stage modules,linear amplifier modules, and more.

Hagerman Technologyhttp://www.hagtech.comInnovative vacuum tube products.Vacu-Trace Vacuum Tube Curve TracerKit version available soon.

AUDIO AMATEUR INC, PO BOX 876, 305 UNION ST, PETERBOROUGH, NH 03458 USA

TOLL-FREE: 800-524-9464 INTL: 603-924-7292 FAX: 603-924-9467

E-MAIL: [email protected]

THE AUDIOXPRESS

WEBSITE RESOURCE PAGE

When my wife Dani decidedwe needed new furniturefor our living room, she in-sisted my speakers had to

go. I had designed and built these manyyears ago. They had a marvelous soundbut weren’t much to look at. Since Danirequired that any speaker must blendin with the room’s new décor, I neededto find a replacement (Photo 1).

This was a golden opportunity for meto acquire a pair of electrostatics. At thetime, the commercial offerings werelacking, so I merged some ideas of myown with those found in The Electrosta-tic Loudspeaker Design Cookbook byRoger Sanders, and a beauty of sightand sound was born.

Many years of listening and tweak-ing had made me comfortable with thesound of my favorite home-brew design.These three-way, time-adjusted speak-ers were supported about 10″ off thefloor by metal legs. The result lookedlike potbellied robots (Photo 2), and thebust of Beethoven on top of one bol-stered the illusion.

I had built several other speakersover the years, trying side-firingwoofers, phase alignment, and otherconcepts, but none sounded as good asthese robots. They provided me with lis-tening enjoyment for over 18 yearswithout fail, except when I had to re-place the woofers after they succumbedto speaker rot.

ESL SEARCHMy first encounter with electrostaticscame in the late ’70s when I heard theDayton Wright Mk III at the AudioGuild in Hohokus, N.J., a place that Isorely miss. The Audio Guild had animpressive showroom spread acrossvarious rooms of a big stone mansion.There were systems set up in the liv-ing room, the sunroom, the poolroom,and the den. The electrostatics weredisplayed prominently in the livingroom.

Because the Audio Guild was such aunique place, it was frequently visitedby some of the big names in audio. Ihad the privilege of talking to MarkWright, Ed Nakamichi, Nelson Pass,and others. The Audio Guild also hadsome very interesting clients.

One day I was sitting on the couch,relaxing, listening, sipping a cup of cof-fee, and tasting the most fabulousBlack Forest cake I’ve ever had(brought in by an Audio Guild client, agourmet pastry chef), when my breathwas taken away. I could no longer tell Iwas listening to speakers and all Iheard was clear, open, effortless music.I had to have them but, as a college stu-dent, I settled for a pair of Stax electro-static headphones. I hoped to own apair of these speakers some day.

That day had finally arrived. Initially,I did not intend to build speakers, andset out on a quest to find a commercialelectrostatic product. A local SoundLabdealer had only the full-range Auras,which I knew would be too large andexpensive, but I went with my wife tolook at them anyway.

Listening brought back those oldfeelings I had in Hohokus. Dani just

kept staring silently. She finally brokethe silence to tell me that they seemedtoo big and to ask how much they were.The response seemed to put her into astate of shock. I forgot to mention that Iwasn’t buying these $27,000 speakers.She wasn’t amused!

After that we traveled around andlistened to various hybrid designs, butthey didn’t have the sound I expected,and some even sounded awful. I justcouldn’t shelve my trusty robots forany of them.

I also scoured the Internet trying tofind more electrostatic speakers, whenI came across Barry Waldron of theElectrostatic Loudspeaker Design eX-change on an audio newsgroup. Wehad an animated e-mail discussion forthe next few months, and I was per-

This article describes the construction of a hybrid electrostatic/transmis-

sion line speaker with highlights of one audiophile’s lifelong journey for

quality sound. By Gary Gendel

Electrostatic Odyssey


PHOTO 1: The finished speakers.

ABOUT THE AUTHORGary Gendel has BSEE, MSEE, MSCS, and MSBEdegrees from NJIT. He holds three patents and severalawards, including the David Sarnoff Outstanding Tech-nical Achievement Award and two Emmys. He is VP atGenashor Corp. and is a researcher for Sarnoff Corpo-ration.

suaded to build speakers instead. I feltlike a novice since it had been so longsince I had designed any speakers oraudio equipment. I had been hooked onaudio since I built my first transistorradio at the age of eight.

When I went to engineering college,I wished to design audio equipment. Inmy junior year, I produced a high-quali-ty amplifier using the cascode transis-tor configuration. In my senior year, Ibuilt a digital-switching amplifier basedon the brand new VFET technology.

As a student, I worked for a localacoustic testing lab and consultingfirm. After graduation, I didn’t receiveany response from the resumes I sentto audio companies, so I took a job de-signing integrated circuits. Having thebackground, but no recent experience,I was both excited and nervous abouttrying to pick up where I left off. I pur-chased several books on loudspeakerdesign and dug in.

DIY DECISIONGiven the limited wall space, I startedwith the “Compact” design from Roger

Sanders’ book, in which the woofer isloaded into a transmission line bentinto an “L” to hold the electrostatic pan-els. I calculated that the design coulduse only about 25″ of wall space. Origi-nally, I planned to produce the electro-static grids with closely spaced weldingrods. My biggest concern was how theywould look.

Just about that time, Barry started tooffer for sale assembled panels thatseemed to fit the bill. These used perfo-rated metal panels for the grids andwere about 14 × 42″ overall. I was hop-ing for something more on the order of70″ high, so I could provide an overallspeaker height around 7′ that wouldperform well with the listener in a seat-ed or standing position. I wrestled withthis problem and finally decided to usetwo panels stacked vertically, since Ihad room to spare with 9′ ceilings.

I chose the 12″ Dynaudio 30W54woofers in the transmission lines. Inorder to provide solid support for thepanels, I made a complete framearound the electrostatic section. I locat-ed an active crossover unit designed


PHOTO 2: The time-adjusted design that served me for many years. Functional, but ugly.

specifically for hybrid electrostatic de-signs from Acoustic Instruments, Inc.This unit came with full documentationand enabled me to tailor the crossoverpoints, equalization, and gain of the cir-cuits for my design. The company is nolonger in business, so an alternativesource is required for future designs.

I produced a 3D model of the cabinet(Figs. 1a and b) in AutoCAD. I chose toangle the cabinet 60° to produce a paral-lelogram for three reasons. The firstwas to reduce by several inches the lin-ear wall space required by the speakers.

The second was to provide a beam-splitter for the acoustic back wave. Icalculated that the cabinet would de-flect approximately two-thirds of the re-flected wave when placed against thewall. I added a slight backward slopeto the front to move back the center ofgravity for stability. As an added bene-fit, with walls that are not parallel thecabinet’s resonance characteristicsshould improve.

Third, the sonically dead regionwhere the two vertical panels meetwould be focused away from the headof a listener in either a standing or sitting position. The software modelwas needed to show Dani what I had in mind.

ASSEMBLYI ordered panels, step-up transformers,and high-voltage power supplies fromBarry. The first pair of panels came gen-erously surrounded by bubble-wrap in acustom crate. I shipped back the emptycontainer and was promptly reim-bursed for the shipping charges.

The other pair came in a sandwichconsisting of the panels individuallycovered in bubble-wrap placed betweenplywood plates and wrapped with card-board. These panels had a gentle butcomplex warp when they arrived. Theyseemed to be fine after mounting, but Iwarned Barry about the problem. He as-sured me that future panels would beshipped using crates.

Each panel had one perforated sheetwider than the other and centered, soabout ¾″ extended on the left and rightsides. This made the panels very easy tomount using screws. I routed a channelaround the mounting frame so that thesmaller perforated grid would be re-cessed, hiding the panel edges fromview. I made the routing deep enoughso that there was ample room to keepthe panels from becoming compressed.

The wider grids were also slightlyshorter in length where the wires areattached, which gave easy access to the connection leads. The perforations of

the front and back grids were aligned,giving them a semitransparent look. Inoted only a couple of minor problemswith the design.

First, the larger grids were not cutthe same. The panels were assembledwith a significant difference in the posi-tion of one grid to the other. This wasprobably due to alignment of the perfo-rations, but had the effect of the flangevarying by up to ¼″ from panel to panel.

The second was a large (more than1″) nonperforated area at the top andbottom of the panel. Because I wasplanning to stack the panels vertically,it meant it would be “sonically dead” atthe 2″ region where the panels abutted.It should be possible to perforate moreof the panel without sacrificing panelstrength. The slope of the front panel inmy design helped direct this dead re-gion away from the ears of a listener.Overall, the look of each panel is excel-lent and the construction is good.

I was forced to increase some of myoriginal dimensions to accommodate


PHOTO 3: The small speaker cabinet mock-up.

FIGURE 1a: Blueprints for the new design (using AutoCAD 2000).

FIGURE 1b: 3D computer-renderedmodel (generated from Fig. 1a).

the variances in the delivered panels. Ineeded at least another inch vertically,but ended up adding 3″ and an inch tothe width to accommodate any possibleproblems due to woodworking toler-ances. I built a roughly one-third scalemodel out of scrap lumber (Photo 3).This helped uncover several construc-tion issues and ensured that the unitwould be stable.

Using this model, I determined thatthe horizontal transmission lineboard above the woofer should beattached only after mounting theelectronics, wiring the woofer andpanels, and stuffing the transmis-sion line. Otherwise I would needan access opening (Photo 4). Thismodel was instrumental in reduc-ing Dani’s skepticism about my de-sign while I was constructing theactual cabinets.

FINISH LINEI used MDF for the cabinet, androuted the woofer opening to flush-mount the woofers. I assembled thecabinets with a glue-and-clamp ap-

proach, determining that additionalbracing was not required. The finalpanel for the top of the woofer wasmade to fit snug. Sealing this later withsilicon caulking made an air-tight, vi-bration-free, joint for easy removal tomake repairs or modifications.

Once the glue dried, I applied amaple veneer and finished with a redoak stain, distressed by black flecks of

paint, and added several coats ofpolyurethane. A worried Dani cried,“They’re huge!” I reassured her the bestI could and kept working.

Drill small holes in the transmissionline adjacent to the wiring area on thepanels and feed the connecting wiresthrough these holes and down thetransmission line. Apply beads of sili-


PHOTO 5: Wiring of the adjacent panels.

PHOTO 4: The woofer section of the TL withthe top plate removed prior to final assembly.

LANGREX SUPPLIES LTDDISTRIBUTORS OF ELECTRONIC VALVES, TUBES & SEMICONDUCTORS AND I.C.S.

1 MAYO ROAD, CROYDON, SURREY, ENGLAND CR0 2QP24 HOUR EXPRESS MAIL ORDER SERVICE ON STOCK ITEMS

E-MAIL: [email protected]

A SELECTION OF OUR STOCKS OF NEW ORIGINAL VALVES/TUBES MANY OTHER BRANDS AVAILABLE

MANY OTHER BRANDS AVAILABLEThese are a selection from our stock of over 6,000 types. Please call or FAX for

an immediate quotation on any types not listed. We are one of the largest distributors of valves in the UK. Same day dispatch. Visa/Mastercard acceptable.

Air Post/ Packing (Please Enquire). Obsolete types are our specialty.

5R4GY RCA 7.505U4GB SYLVANIA 10.005Y3WGT SYLVANIA 3.506BX7GT GE 7.506FQ7 SYLVANIA 7.506L6GC G.E. 17.506L6WGB SYLVANIA 10.006SL7GT USA 5.006SN7GT USA 7.506V6GT BRIMAR 6.0012AX7WA SYLVANIA 6.0012BH7 BRIMAR 12.0012BY7A G.E. 7.00211/VT4C G.E. 85.00807 HYTRON 7.505687WB ECG 6.006080 RCA 10.006146B G.E. 15.006550A JAN G.E. 25.006973 RCA 15.007027A G.E. 25.007581A SYLVANIA 15.00

PHONE44-208-684-1166

FAX44-208-684-3056

LANGREX SUPPLIES LTD

ECC81 RFT 3.00ECC82 RFT 5.00ECC83 RFT 6.00ECC83 EI 3.00ECC85 RFT 5.00ECC88 BRIMAR 6.00ECC88 SIEMENS 7.50ECC88 MULLARD 7.50ECL82 MULLARD 5.00ECL86 SIEMENS 7.50EF86 USSR 5.00EF86 MULLARD 15.00EL34 EI 5.00EL34G SOVTEK 5.00EL84 USSR 3.00EL519 EI 7.50EZ80 MULLARD 5.00EZ81 MULLARD 7.50GZ33/37 MULLARD 15.00PL509 MULLARD 10.00UCH81 MULLARD 3.00UCL82 MULLARD 2.00

A2900/CV6091 G.E.C. 17.50E82CC SIEMENS 7.50E83CC TESLA 7.50E88CC/CV2492 BRIMAR 8.50ECC81/6201 G.E. 5.00ECC81/CV4024 MULLARD 6.00ECC81/M8162 MULLARD 7.50ECC81/6201 G. PIN MULLARD 8.50ECC82/CV4003 MULLARD 7.50ECC82/M8136 MULLARD 10.00ECC83/CV4004 MULLARD 20.00

SOCKETSB7G CHASSIS 0.60B9A CHASSIS 1.00OCTAL CHASSIS 1.00OCTAL MAZDA 2.00LOCTAL B8G CHASSIS 2.50

SCREENING CANSALL SIZES 1.00

STANDARD TYPES AMERICAN TYPES SPECIAL QUALITY TYPES

con caulking to hold the wires andkeep them spaced apart to prevent arc-ing. Mount the panels so the connec-tion wire leads are near each other atthe center (Photo 5) and affix them with#4 sheet-metal screws every 4″ in theside. I soldered and insulated the con-nections with heat-shrinkable tubingand stuffed each transmission line with3½ lbs of polyester fiberfill, reducingthe density linearly starting from thewoofers and ending about 6″ from theend of the transmission line (Photo 6).

Once the lines were stuffed, Imounted the high-voltage supply andstep-up transformers in the space be-hind the first baffle directly behindthe woofer (Photo 7). I installed five-way binding posts for external con-nections and tapped the “removable”panel into place and secured it usinga silicon caulk along all inside seamsand along the rear baffle. Finally, Icaulked and mounted the woofers. Atop screw also helped secure the “re-movable” panel.

Using a signal generator and cali-brated microphone, I adjusted the ac-tive crossover. I was finally ready to lis-ten and enjoy.

LISTENING CRITIQUEI started by playing a few pieces. Justlike my headphones, they had a warmsound with incredible depth and clari-ty. I listened to a portion of my collec-tion covering rock, jazz, and classicalon both vinyl and CD. Since I had neverlistened to a transmission-line woofer


TABLE 1PARTS LIST

QUANTITY DESCRIPTION SUPPLIER4 ESL panels ESL Information eXchange2 Audio step-up transformers ESL Information eXchange2 High-voltage power supplies ESL Information eXchange1 Active crossover/equalizer Acoustic Instruments, Inc.2 Dynaudio 30W54 Parts Express2 pair 5-way binding posts Parts Express4 4 × 8′ sheets MDF Local lumber supply4 4 × 8′ sheets maple veneer Local lumber supply1 quart Red oak stain Local hardware supply1 pint Black enamel (for distressing) Local hardware supply1 quart Polyurethane clear coat Local hardware supply1 cartridge Silicon sealer Local hardware supply1 box #4 × ½″ sheet-metal screws Local hardware supply

PHOTO 6: Thetransmission-lineopening fullystuffed.

NEW!

ELECTRONIC CIRCUIT ACTION SERIES: AMPLIFIER CIRCUITS

by Thomas M Adams

New CD-ROM reproduction of this very in-formative book. Using four-color diagrams,Adams helps you visualize the actions thattake place inside amplifier circuits. In-cludes circuit fundamentals and analysisof basic AF and RF circuits. Using non-mathematical and logical explanations, hediscusses the movements shown by colorin the diagrams and then discusses thesemovements and explains their significancein overall circuit operation. Questions areincluded to reinforce your understandingas you progress through the text. This sec-ond edition published 1966, originally136pp., CD-ROM published 2001.

CDAA1 ....................................$16.95plus $6.50 shipping in US & Canada ($7.00 other surface; $11.25 other air)

Old Colony Sound Lab, PO Box 876 Dept. X1, Peterborough, NH 03458-0876 USA Phone: 603-924-9464 Fax: 603-924-9467E-mail: [email protected] ORDER ON-LINE AT www.audioXpress.com

To order call 1-888-924-9465 or e-mail [email protected]

before, I was thrilled to discover theirclean, tight bass.

The speakers looked much smallerin place against the wall and within afew weeks they were hardly noticeable.Dani even shows them off to guests as agreat conversation piece.

It’s been more than a year since thatfirst listening session. Dani and I haveenjoyed many hours listening togetherand also had the pleasure of showingthem off to numerous people.

I suggest that my visitors bring somemusic that they are very familiar with.One brought a CD that he had been lis-tening to continuously for severalweeks. He commented, “I thought Iknew this CD inside and out, but I’mhearing instruments that I never knewwere there.”

Another time, listening to BuddySpicher’s acoustic jazz rendition of“Georgia on My Mind,” a friend ex-claimed, “I could swear the musiciansare right in front of me!” I saw thatsame look in his eyes that I had yearsbefore when I was sipping coffee, eat-ing cake, and listening to the beautyof music from a pair of electrostaticspeakers.


SOURCESESL Information eXchange2820 Miller WayPlacerville, CA [email protected]/eslinfoNote: Barry included replacement resistors for thehigh-voltage supplies to allow them to charge two panels each.

Parts Express725 Pleasant Valley Dr.Springsboro, OH 45066-1158513-734-3000sales@partsexpress.comwww.partsexpress.comNote: Parts Express included a mounting kit includingscrews and sealant with the Dynaudio woofers.

REFERENCES“An Electrostatic Speaker System,” David P. Hermey-er, TAA 3 and 4/73.

“The Sanders Electrostatic Speaker,” Roger R.Sanders, TAA 4/75 and 1/76.

“A High Efficiency Electrostatic Loudspeaker System,”David P. Hermeyer, TAA 2 and 3/77.

Electrostatic Loudspeaker Design and Construction,Ronald Wagner, Audio Amateur Publications, Inc.,1993.

The Electrostatic Loudspeaker Design Cookbook,Roger R. Sanders, Audio Amateur Publications, Inc.,1995. Note: This last one contains the No-compromiseHybrid and the Compact/Integrated ESL/TL designsthat I merged together for the basis for the new design.

PHOTO 7: An overhead view showing thewoofer area (stuffed), the rear woofer deflector, and the electronics area.

Ihave read dozens of articles and re-viewed many recommendationsabout adding a retard circuit via arelay-controlled timer, a rectifier

tube, or even a “standby” switch toavoid voltage jolts to cold components.All these solutions can work. But whatif you own a vintage amp? There simplyis no space to add a transformer, a tube socket, or relay. Drilling a hole into the

chassis so as to add a standby switch isnot an option. Most collectors wouldnot purchase an original amplifier ifmajor modifications had taken place.

My proposed circuit (Fig. 1) is asmall PCB with only two terminals thatgo in series to the HV. Attaching thePCB to the amplifier can return it to itsoriginal state by reconnecting the cableyou unsoldered. There are no connec-tions to ground, no relay, no switch,and no cable runs (Photo 1).

The circuit contains an SCR that actsas the switch, a 555 IC timer, and otherassociated components. Once the SCRswitches on, all components on the cir-cuit vanish from the electrical path.The drop voltage of the SCR is less than1V. The voltage of the associated com-ponents is taken from a shunting cas-cade of zener diodes.

INSTALLATIONBe careful, for this project involveshigh voltage; electric shock can befatal. Before installing, always unplugthe unit from the AC socket and fullydischarge the electrolytic capacitorswith a 10kΩ resistor to the ground.Then wait a few seconds before tryingto handle the amplifier. Always checkthe voltage before soldering or unsol-dering any cable.

Locate the positive lead of the mainrectifier and the first electrolytic capaci-tor. You will notice several other con-nections that lead to the amp: one goesto each OPT, and a resistor goes to thenext HV stage. The connections to thePCB go from one side to the junction ofthe positive side of the rectifier and thefirst electrolytic capacitor; the otherconnection goes from the remaining ca-bles that were connected to that point.

A Start-Up Delay for Vintage Amplifiers

ABOUT THE AUTHORAlex Rubli is a Macintosh computer consultant based inCholula, Mexico. He worked as chief engineer/designerat Pierdant Electronica, the German DUAL audio com-pany representative in Mexico back in the ’80s, whensome of his designs, such as turntables, solid state am-plifiers, and speakers, hit the Mexican market. For per-sonal reasons, he moved out of this business and wentinto computing science. His interest in tube audio re-emerged some years ago. In his laboratory, it isusual to see high-end Macintosh computers testinglaboratory cards and software, next to a pair of vacuumtube amplifiers waiting to be restored.

FIGURE 1: Circuit schematic.

To OPT`s

cut here

FIGURE 2: Connecting the PCB.

To OPT`s

PCB

FIGURE 3: PCB placement.

PHOTO 1: Finished circuit.


This circuit offers a different solution to delay the start-up of classic

amps. By Alexander Rubli Kaiser

G-1731-1

G-1731-2 G-1731-3

See Fig. 2; notice thedirection of the arrowon the PCB.

The circuit is intend-ed for use with an HVof 350-550V. If youhave an amplifier witha lower HV, try by-passing ZD2. The volt-age across R1 shouldbe 45V or more.

THE LEDWhen the amplifier ispowered on, the LEDwill turn on until thetimer fires up andtakes the SCR to the“on” state. After thisoccurs, the LED willturn off.

The timing circuitwill fire up the SCR inabout 45 seconds ifthe tubes are cold. Ifthe tubes are hot, thetime is reduced aboutten seconds. The nec-

essary current for this circuit originatesprimarily from the idle current in theoutput tubes. If you are troubleshootingthe amp and remove the tubes, the SCRwill not activate, and you need to by-pass the circuit.

The PCB (Fig. 3) measures 3″ × 1¹⁄₈″.The completely assembled PCB is avail-able from the author at [email protected]. You can find further information athttp://www.pobox.com/~rubli.


FIGURE 4A: Copper side. (100%)

FIGURE 4B: Component side. (100%)

FIGURE 4C: Soldering mask. (100%)

G-1731-4A

G-1731-4B

G-1731-4C

VISIT OUR WEBSITE AT

www.audioXpress.com

OR

E-MAIL US [email protected]

In 1948 my parents purchased oneof the last great radios, a Westing-house AM/FM-Phonograph con-sole, Model H-169, for $400. This

was the top of the line, a no compromise“build it the best we can” radio that wasat the tail end of a long, long period ofdevelopment. The engineers knew thequestions and had the answers. I don’tplay the Westinghouse much anymore,but when I do I’m amazed at the wonder-ful sound that comes out of this 50-year-old classic. I recently set out to seewhether I could find out why, and buildone that sounds the same.

A MARVEL OF QUALITYThe first thing you notice about theWestinghouse is the quality. The cabi-net work is superb. There’s a saying,“they don’t build them like that any-more.” In this case, I’m not sure theycan. The quality is internal also. In 50years the output 6L6s have been re-placed twice, as well as, I believe, twoother tubes. That’s it! No capacitors, noresistors, no transformers, nothing.And there’s not a trace of hum.

Electronically, the Westinghouseuses push-pull beam tetrodes in the output without feedback. D.T.N.Williamson would be spinning in hisgrave but the fact remains that bigtubes—run at normal listening levelsunder a watt of average output—don’tproduce much distortion, and the West-inghouse has some unique circuit fea-tures to minimize even that.

You’ll notice in the circuit diagramthat there is a 5000pF capacitor fromthe plate of each output tube to ground.When I tried it in my version, there wasa severe rolloff of the high frequencies—

over 4dB at 10kHz and 8dB at 20kHz.But the Westinghouse has great treble!What I believe happens is that the out-put stage, which produces the most har-monic distortion, is purposely designedto roll off the highs, while the circuitrybetween the triode voltage amplifierand driver is used to boost them.

For example, say there’s a 2kHz noteat the input of the amplifier. The volt-age amplifier is flat at this frequencyand sends it on to the output stage,where a small harmonic at 6kHz is pro-duced, as well as a smaller one at10kHz. The harmonics are rolled offand are not as noticeable.

Suppose, instead, you put a 6kHzand a 10kHz note or overtone that’spart of the music at the input. These areemphasized in the voltage amplifier,but rolled off in the output stage, so theoutput is reasonably flat with lower dis-tortion. At least that’s my theory, andit’s the way the Westinghouse sounds. Iprefer my amps to be flat, but this maybe a technique worth exploring.

TWO-TAP VOLUME CONTROLAnother feature of the Westinghouse isthe volume control with two taps thatare connected to circuitry that booststhe bass at low volume levels. This,combined with the open-back-speakersub-cabinet, helps give the Westing-house a velvety bass at low listeninglevels. Poorly designed, an open-backspeaker can sound boomy, but theWestinghouse is an integrated system,

Before super-low-distortion amplifiers, there was the classic sound of the

Westinghouse console, which you can re-capture in your living room

today. By Larry Lisle

Recreating a 1948 Console Amp

PHOTO 1: The amplifier enclosure.


PHOTO 2: This is a basic push-pull amplifier, but it sounds very niceat quiet listening levels.

and all of the sections work together.I’d love to talk to the designers!

The driver circuit is unusual, featur-ing a 6SN7 cathode-coupled phase split-ter and a center-tapped interstage trans-

former. I’ve never seen both in anyother amplifier. The cathode coupledphase splitter has the advantage oversome other designs of canceling outhum from both the high voltage and fil-

ament supplies. It also prevents currentfrom flowing in the primary of the inter-stage transformer, which would de-grade its bandwidth.

The transformer, on the other hand,offers excellent balance at all frequen-cies, as well as low resistance in theinput circuit if the grids of the poweramplifier tubes are driven positive. It’sanother indication of the “spare no ex-pense” mentality that went into theWestinghouse.

There are other examples of this attention to quality, such as the sepa-rate output transformers for eachspeaker, but now I’ll describe my at-tempt to duplicate the sound usingmodern components.

USING AVAILABLE PARTSI didn’t attempt to build an exact dupli-cate of the Westinghouse since I alreadyhad one in the living room. Besides,some of the parts, such as the two-tapvolume control, the paper capacitors,the field-coil speaker, and so on, aren’tavailable. Instead I concentrated on thesound and on using parts that you couldfind without too much trouble.

I chose 807s as the output tubes be-cause they’re very similar to 6L6s andfine-quality NOS tubes are easy to findat low prices. I used an output trans-former purchased at a flea market thatwas almost certainly rated for 6L6s.There are better transformers available,


FIGURE 1: The secrets of the old console re-vealed. There are a lot of clever tricks in thiscircuit. See text for details.

but I thought if I could make this onesound good, you could be confident intrying other transformers.

The screen and plate connection ofeach output tube is brought out to aseparate terminal, so if you decide youreally don’t like the beam-tetrodesound, you can easily reconnect the807s as triodes. Otherwise, the outputstage is routine, with screwdriver-ad-justable controls for the total bias andbalance between the tubes. The capaci-tors from cathodes to ground are to by-pass odd order harmonics.

In the driver stage I used a trans-former from Antique Electronic Supply.There are much better interstage trans-formers, but again, if I could get thisone to perform, the others will soundeven better.

DRIVER STAGEAfter considerable experimenting, I de-cided on a pair of triode-connected 3Q4sin parallel as the driver. The reasoninghere was to get the lowest possible plateresistance in the driver stage, which isaround 1700Ω with parallel 3Q4s, to-gether with low distortion, hum, and

other noise. It was interesting to seethat the formulas for transformer band-width really work in trying tubes withdifferent plate resistance. The 3S4s areeven better, but their gain is a bit lower. I used parallel feed to the primary wind-

ing of the interstage transformer to keepunbalanced DC from limiting its perfor-

mance. Dropping from the 300V sourcedown to about 60V for the drivers allowsyou to use a big resistor and consequent-ly lowering the distortion, so don’t put abypass capacitor at the junction of the10,000Ω resistor and the voltage divider.

As a result of all this, the amplifierdroops about 1.6dB at 20Hz and falls off


PHOTO 3: The old Console. This Westinghouse H-169 from 1948was one of the last of the great radios.

3Q4/V1

3Q4/V2

2W

2W

10W10W

450V DC

2W10W

470V DC

10W

10W

117V AC

1 15 5

ALL GROUNDCONNECTIONSTO THIS POINT

+300V@100mA

2

2

3

3

3

6

4

4

4

571

2W

35V

35V

807V4

J1INPUT

807V3

C1

1µF

R1470KΩ

BT26V

V1, V22x3Q4

R210KΩ

R625KΩ

C2

47µF

R720KΩ

T1 R3150Ω

R4100Ω R5

50Ω

V3807

V4807

C3470µF

C4470µF

R8100Ω

R9100Ω

1

1

7

7

5

5

BT11.5V

M1

0-3VDC

T2C5

47µF

T3

6.3V AC

FIGURE 2: Despite its simplicity, this circuitsounds very good at normal listening levels.The parallel 3Q4s have an exceptionally lowplate resistance, and even an inexpensive inter-stage transformer can be made to work well.

a barely perceptible amount at 20kHz,with a slight bump of about 0.6dB at13kHz. THD runs about 2.5% at 10Wand is under 1% at 1W.

The 807s are capable of much higheroutput, of course, if you raise the sourcevoltage. (Voltage ratings of the capaci-tors may need to be increased.) But thepresent setup gives enough headroomfor transients and peaks. 300V is also acomfortable limit for everyday use.

THE ENCLOSURE I built the amplifier on a 24″ by 8″ (nomi-nal) piece of red oak. I like this style of

construction for its sound, the way itkeeps heat localized, and its simplicity.The cover is made of a combination ofpoplar and birch. The meter on the frontmeasures filament voltage for the 3Q4s,and the battery clip for a single alkaline“D” cell is on the back. Battery life isvery good. Battery tubes were designedto operate over a wide range, and 3Q4swill sound fine down to 1.1V and lower.

The front panel drops down onhinges to access the controls for bias-ing and balancing the output tubes.The front and back panels are a littledistance away from the baseboard toallow air to enter at the bottom of thebox and exit through vents located atthe top of the cover on the back side.The driver stage is built in a small alu-minum box. I used battery bias be-cause I like the sound.

You should be able to rotate the inter-stage transformer for minimum humpickup, so leave the wires long anddon’t box it in with other parts. Thehum actually comes from fields in theroom, not from circuitry in the amplifi-er or power supply, which you shouldlocate a little distance away. If youmove the amplifier to a new location,take a minute to rotate the transformerfor minimum output with no signal.You can make this amplifier very quiet.

POWER SUPPLYYou can use any power supply capable

of delivering 300V at 100mA. Don’t everwork on this or any other amplifierwithout unplugging every power cordand discharging the positive terminalof every capacitor to ground. You canlocate the filament transformer withthe power supply if you wish, but wirescarrying the filament current should betwisted together and shielded from thewire carrying the high voltage. Switch-ing arrangements and connectors canbe chosen to suit the builder.

THE SOUNDSo how does it sound? It’s hard to de-scribe the sound of an amplifier inwords, but I’ll try. It’s not completelytransparent, but almost. It doesn’tsound liquid, as can a good single-ended triode amp. There’s just a bit ofan edge right at the threshold of per-ception. It doesn’t sound good on elec-tronic drum machines, but with big-band or other music using acoustic in-struments it sounds bouncy. If I had tochoose one word I’d say it sounds“happy.” It certainly is not dull.

This amplifier is a good starting pointif you wish to go further in the directionof push-pull amps. You can change the807s to triode operation, add some nega-tive feedback, try different transformersand input tubes, and so on.

As for me, I’m going down to my lis-tening room and put on some TommyDorsey.


TABLE 1PARTS LIST

B1 1.5V alkaline “D” cellB2 6V battery (4 “AA” cells)C1 1µFC2, C5 47µF, 450V DCC3, C4 470µF, 35V DCM1 0–3V meterR1 470kΩ, 2WR2 10kΩ, 2WR3 150Ω pot, 5WR4 100Ω, 5WR5 50Ω pot, 5WR6 25k, 10WR7 20k, 10WR8, R9 100Ω, 2WT1 Interstage transformer, single plate to push-

pull grids, 1:1/1 or 1:1.5/1.5 turns ratio prima-ry to secondary. Antique Electronic Supply P-T20A14 or equivalent (6221 S.Maple Ave. Tempe, AZ 85283, 480-820-5411)

T2 Output transformer, 5000Ω plate to plate to voice coil Hammond P-T1645 or equivalent

T3 117V to 6.3V at 2 amps or more transformer V1, V2 3Q4 vacuum tubesV3, V4 807 vacuum tubes

Roksan, 30 St. Peter’s Rd., Hunt-ingdon, UK PE18 7DB, www.roksan.co.uk, [email protected] by May Audio Marketing,2150 Liberty Drive Unit 7, NiagraFalls, NY 14304, 716-283-4434.$1750. Dimensions: 43.5cm W ×33cm D × 8cm H; net weight: 9.5kg, 21 lbs. Limited two-year war-ranty.

T he Roksan Caspian is one ofthe Caspian family of prod-ucts, which includes a power

amplifier, integrated amplifier, andtuner. Photo 1 shows the frontpanel, with its combination front-loading door and fluorescent dis-play. This panel swings down, andthe CD loading drawer slides outwith a push of the open/close but-ton. Three push-button controlsflank each side of the display, withplay/pause buttons on each side.

CONSTRUCTIONThe rugged chassis is black-

painted steel with a thick alu-minum front panel, and thedoor/display is also aluminum. Thedisplay itself is covered with a thin

plastic film that gives it a ghostlygreen out-of-focus look. The user’smanual did not say to remove it.Track information and total timeare displayed for about 10 secondsafter loading a CD, but the unitdoes not display index marks.

The optional remote controlsupplied with the review unit(with the all-too-common tinybuttons) can also operate theCaspian-series Tuner and Inte-grated Amplifier. Making quicktrack changes is not easy, due tothe nonstandard number padwith the buttons arranged in twovertical columns. The CD-only re-mote control (not supplied) hasthe more familiar touch-tonetelephone layout.

The rear panel (Photo 2) has theswitched IEC power receptacle, twopairs of high-quality gold-platedTeflon -insulated RCA analogoutput jacks, and a 75Ω BNC con-nector for the S/PDIF coax digitaloutput. An optical output is notprovided.

The unit is furnished with apower cord, but not with analog ordigital interconnects. The powercan be connected for 120V or 240Vmains, and a line fuse is located ina drawer in the IEC receptacle. Thethird pin of the AC receptacle isconnected to the chassis.

TOPOLOGYA schematic was not supplied

with the unit, and I did not at-tempt to open the unitfor inspection. The Caspi-an uses a Philips TDA1305T BCC-DAC2 withcontinuous calibration.The unit does not per-form HDCD decoding.

The Caspian informa-tion sheet describes itspatented Laser Environ-ment Enhancer Light,which is a green laserthat bathes the CD ingreen light. The En-hancer is supposed to

produce an effect similar to paint-ing the edges of your CDs withgreen paint, such as Audioprism’sCD StopLight.

PRELIMINARY TESTUnfortunately, the Caspian ar-

rived inoperative. The frontdoor/display panel was floppedopen and the loading tray was notfully retracted. When I powered itup and pressed the open/closebutton, the tray retracted, madesome noises and opened up again.The unit appeared to be a (proba-bly well-used) review sample. I re-turned the unit to May Audio Mar-keting, and they promptly fur-nished a new replacement.

MEASUREMENTSI operated the Caspian for one

hour with a loud music CD beforemaking any tests. I performed alltests using the CBS Labs CD-1 andPierre Verany test CDs (both avail-able from Old Colony Sound Lab).The output impedance at 1kHzwas 127Ω, and channel balancewas better than 0.1dB.

The frequency response for theCaspian was within ±0.68dB from17Hz-20kHz. The 0dBFS outputwas 2.12V RMS at 1kHz, or 0.51dBhigher than the CD standard of 2V


PHOTO 2: CD player rear view.

PHOTO 1:CD player front view.

Roksan Caspian CD PlayerReviewed by Charles HansenListening Tests by Betty Jane & Richard Honeycutt

RMS. Crosstalk between channelsto 16kHz was excellent and wasbelow the noise floor of my testequipment.

THD+N versus frequency resultsare shown in Fig. 1. Normal prac-tice when testing devices with dig-ital-analog converter outputs is toengage a 22kHz six-pole, low-passfilter to remove out-of-band noise.My HP-339A distortion test set hastwo three-pole LP filters, at 80kHzand 30kHz.

With both filters engaged, the

combined effect is a slightly bet-ter than three-pole LP filter at28kHz. As such, my THD+N read-ings are somewhat higher thanthey would be if they were mea-sured with the specified LP six-pole filter at half the 44.1kHzdigital sampling rate. (When I re-peated the THD measurementswith just the 80kHz LP filter en-gaged, the THD+N was 0.02%higher across the graph.) Thespikes in THD above 2kHz are theresult of noise and distortion

cross-products with the 44.1kHzsampling frequency and are notunusual in digital equipment.

The spectrum of a 50Hz sinewave at 0dBFS is shown in Fig. 2,from DC to 1.3kHz. The calculatedTHD+N was 0.0025%. There wereno significant harmonics and onlyextremely low-level noise artifacts.

Figure 3 shows the spectrum of response to equal level 19kHzand 20kHz signal, each at −6dBFS, from DC to 20.8kHz. The1kHz intermodulation differenceproduct measured −70.5dB, whilethe 18kHz and 21kHz productsmeasured −38dB. The spectrumanalyzer FFT was still processingthe data, and the IMD spectrumwas decreasing when the 30-sec-ond CD test track ran out. The ac-tual IMD is probably lower than itappears.

Figure 4 shows the reproductionof an undithered 1kHz sine waveat −90.31dBFS. At this level, the

signal consisted of ±1 bit of data,producing two different voltagelevels that are symmetrical aboutthe horizontal axis (time). Thesediscrete voltage steps were not ob-vious, probably due to out-of-bandhigh frequency noise. The ringingafter each vertical transition canbe explained by the Gibbs Phe-nomenon1.

The Caspian ignored defects onthe Pierre Verany Test CD#2 out totrack 32, which has a 1.25mm longsection of blank data. At track 33(1.5mm defect), the unit emitteda series of audible clicks at eachpass through the defect. The uniteasily met the Red Book require-ment of 0.2mm max.


TABLE 1MEASURED PERFORMANCE

PARAMETER MANUFACTURER’S RATING MEASURED RESULTSOutput 2V RMS 2.12V RMSFrequency response 20Hz–20kHz ±1dB 8Hz–20kHz ±0.68dBTotal harmonic distortion 0.003% (1kHz) 0.012% (see text) IMD-CCIF (19 + 20kHz) N/S −70.5dB, 1kHzOutput impedance N/S 127ΩSignal to noise ratio 105dB (IHF-A)Channel separation 100dB (1kHz) unmeasurablePower requirements <15W

kHz

0.0 0.2 0.4 0.6 0.8 1.0 1.2

dB

-100

-90

-80

-70

-60

-50

-40

-30

-20

-10

0

kHz

0 5 10 15 20

dB

-100

-90

-80

-70

-60

-50

-40

-30

-20

-10

0

FIGURE 1: THD++N versus output power.

FIGURE 2: Spectrum of 50Hz sine wave.

FIGURE 3: Spectrum of 19kHz ++ 20kHz intermodulation.

REFERENCES1. Arfken, G. Gibbs Phenomenon,“Mathematical Methods for Physicists,”3rd Ed., pp. 783–787, Academic Press,1985.

ms0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0

V

-0.5

-0.4

-0.3

-0.2

-0.1

0.0

0.1

0.2

0.3

0.4

0.5

FIGURE 4: Undithered 1kHz sine wave at −90.31dBFS.


Reviewed by Betty Jane and Richard Honeycutt

It’s usually the little things...The elegant front panel.The solid (and heavy) feel. The smooth, silent glideof the door as it opens and closes. The fact that it isa single-CD player, not a changer. The gold platingon the RCA connectors. The heavy rocker switch onthe back panel. The heavy, removable power cord.And, of course, the name. (We’re Narnia fans fromway back!) By the time you actually start to listen tothe Roksan Caspian, you’re already impressed by thequality.

On the other hand, it is this very sort of first im-pression that makes it doubly important to be onyour guard, especially when not testing double-blindagainst some standard. So between the Caspian’sarrival in our home and our actual listening session,I did some thinking about just what sort of sonic dif-ferences you could reasonably expect among profes-sionally designed CD players. My mind went back tothe horrid first impression I had of compact discs.

A friend of mine had purchased an early CD playerand several “Best of the Beach” or some such CDs.Those of us who were in radio back in the ’60s used tocomment that 45 RPM records were pressed on a mix-ture of chicken manure and sawdust. This first CDsounded even worse. Obviously the engineer thoughtyou could overdrive a CD master by 10 or 15dB, just ashe always overdrove 45 RPM record masters. The dis-tortion was hideous.

Fast-forward to the ’90s. By now, the requirementsof proper digital recording were well understood by theprofessionals. Overdriving simply did not occur. Butlistening to a slow fadeout, you could sometimes heargraininess, especially near the “silent” end of the fade.This seemed to vary with the CD player.

Such graininess could arise from improper resistortolerances in the D/A converters, causing amplitudeerrors. And since all CD players have analog sections,the usual bugaboos of hum, noise, high-order harmon-ic distortion, IM distortion, and slight frequency-re-sponse aberrations could rear their heads. As a result,the sound could be a bit thin or edgy, or inner voicescould be somewhat muddy.

All that said, I must admit that I own a fairly low-priced five-CD changer, and have been generallypleased with its performance. In fact, my cassetterecorder cost over twice what my CD player did. Butnow you know what we were looking for in our auditionof the Caspian.

We listened to the following selections from Hi-FiNews and Record Review’s CD Test Disk III, with com-ments as appropriate.

TEST 1Track 2—Jerusalem/ParryBJH: The performers sounded very close to the listen-er. Because of the recording venue being outdoors, Icould not tell much about the original acoustic envi-ronment.RH: The outdoor recording, reproduced in our fairlyacoustically dead living room, sounded cramped to me.Not the best environment for listening to this recording!

TEST 2Track 4—Trumpet Concerto in C/Vivaldi

BJH: Excellent presentation! Very natural.RH: The trumpets (which are very difficult to properlyrecord) are superb. See note below on a further test.

TEST 3Tracks 5/6—Peter and the Wolf (narr)/ProkofievBJH: Not quite the presence of the Vivaldi, but thesonority and locations of the various instruments arevery well-defined.RH: The narrator, especially, could have had a bitmore presence, but the soundstage was excellent.

TEST 4Track 7—Welcome, Welcome/PurcellBJH: Generally good ensemble. However, the harpsi-chord’s position seemed to roam according to theregister in which it was played: lower register to theleft, upper register to the right. The apparent changein position was twice the length of the typical instru-ment, so the instrument’s orientation could not ex-plain this effect.RH: The overall sound would be wonderful for a Ro-mantic composition, but the room sounds too largefor a Purcell chamber piece. But then the equipmentaccurately conveys the size of the room. Singerssounded somewhat blanketed and lacking the inti-macy that should characterize this type of music.But I believe this is all in the recording.

TEST 5Track 10—Corkhill (Piece 2)BJH: Crisp, lots of presence.RH: The silent sections are truly silent. The full dy-namic range of the piece (pp to f, I would guess) isreproduced effortlessly.

TEST 6Track 14—Rio Napo RSS DemoBJH: Presence and stereo effect are exceptional.RH: Fine studio recording.

Notice that by far most of our comments on thesetracks did not relate to the ways any professionally de-signed CD player might conceivably affect the music;rather, they pertained to the recordings. Consequently,the accompanying chart of sonic characteristics rat-ings, based on these tests, should be taken with agrain of salt. So, to focus on any possible graininess,edginess, hum, noise, or distortion, and thereby givean honest and useful rating of the sound of the Caspi-an, we made...

FURTHER TESTSFirst, we played the Vival-di back on our own CDplayer, A/B-ing the two.The Caspian definitelyproduces a rounder, moreperfect trumpet sound. MyCD player has a slightedginess on that sound.As a trumpet player my-self, I (Richard) definitelyconsidered the Caspian todeliver superior reproduc-tion in this test.

Second, we listened to Chadwick’s “Noel” from Sym-phonic Sketches, on the CD HDCD Sample, Vol. 2 (Ref-erence Recordings RR-905CD). This silken choral per-formance would easily reveal any artifacts in frequencyresponse as a deviation from the absolutely perfectbalance and sonority of the choir, and the wide dynam-ic range provides a fine test of low-level imperfectionsin the D/A converter. No flaws in the Caspian werefound.

Third, we listened to Chesnokov’s “SpaséñiyeSodélal” (“Salvation is Created”), also from the HDCDSample CD. On this one you could hear the colors ofthe singers’ dental fillings. No distortion could hidefrom this test.

Finally, we auditioned several tracks from MichaelCard and John Michael Talbot’s Brother to Brother,which contains some of the best-recorded blends ofnylon-string guitar, electric bass, percussion, and malevoice we have heard. Again, the Caspian performedflawlessly.

CONCLUSIONWe had only one minor concern: the operation of theremote control was not exactly as we expected. In ourexperience, when you press a button on the remotecorresponding to a track number, the CD player im-mediately begins playing that track. On this one, youmust first stop the current selection, then select thenew track, then press play. Of course, we had nomanual with the unit, so there may be a simplermode, but we did not find it.

We were not told the price of the Caspian; perhapsthat’s a good thing. But if you, like myself (Richard)prefer a single-disk CD player, I don’t see how youcould do better than the Caspian. Even the remote-control issue is minor, because I seldom use a remotewith a CD player anyway. And the appearance, con-struction, and sonic performance are superb.

CRITIQUE

SONIC CHARACTERISTICS RATINGS

BJHRH

Presence

BJHRH

Stereophonic Effect

BJHRH

Soundstaging

BJHRH

Ambience

1 2 3 4 5 6 7 8 9 10

ABOUT THE AUTHORSRichard Honeycutt is an accomplished musicianand author. He is Chief Engineer at ElectroacousticDevelopment Company and EDC Sound Services,where he is engaged in acoustical and electroa-coustical consulting and design. Betty Jane Honey-cutt is a professional composer and musician witha BA in Music from Greensboro College. She andRichard have three grown children. In her sparetime, Betty Jane helps operate EDC Sound Ser-vices and also teaches literacy and English as asecond language.

Music Hall Audio, 108 Station Rd.,Great Neck, NY 11023, 516-487-3663,www.musichallaudio.com. $299 US.Supplied by Music Direct, 800-449-8333,www.amusicdirect.com. Dimensions:16″ W × 4″ H × 13″ D (including dustcover); weight: 17 lbs.

The Music Hall MMF-2.1 is a single-play two-speed turntable equipped witha factory-installed Goldring Elan mov-ing-magnet (MM) cartridge and dustcover. Photo 1 shows a front view of theMMF-2.1. The 4.5 lb platter is belt-dri-ven by a shielded synchronous-reac-tance motor that has two step-crownedpulleys, providing speeds of 33¹⁄₃ and 45rpm (a 45 rpm spindle adapter is sup-plied). The motor is mounted to theplinth by an elastomer isolator.

You can change speeds by removingthe platter and shifting the flat rubberdrive belt from one pulley step to an-other. A plastic speed change key issupplied with the unit so you do nothave to touch the belt with your hands,thus avoiding oily contamination(MHA also supplies a spare drive belt).The aluminum platter, equipped with afelt record mat, sits on a die moldedplastic sub-platter. The tonearmlift/cueing lever has a nice damped ac-tion that sets the stylus gently on therecord.

FEATURESThe straight tonearm measures 9″ frompivot to stylus and has a dual bearinggimbal mount. The cartridge is mount-ed at a 25° angle in the headshell. Twoopposing hex-key screws on the coun-terweight shaft allow for vertical track-ing alignment (VTA) adjustment, and ahex-key wrench is supplied.

The counterweight has an elastomerisolation insert and threads onto thecounterweight shaft for tracking force(TF) adjustments. The TF range is 1.25 to 5.25 grams, with a cartridge

weight of 3gm. Music Hall supplies aslotted ½ × ¾″ sheet-brass weight thatcan be installed under very lightweightcartridges.

Anti-skating is implemented bymeans of a small cylindrical weight.You place its thin monofilament line inone of three notches in the anti-skatingrod attached to the rear of the tone-arm. I found that this is best done with tweezers. The weight then loopsthrough an anti-skating lever that di-rects the monofilament line away fromthe tonearm and allows the weight tohang vertically above the plinth. Notch2 is used for the Elan.

The tonearm wiring is hard-wired toa pair of shielded interconnects withgold-plated RCA phono connectors. Anintegral turntable ground wire is in-cluded with this cable pair to allow youto ground the MMF-2.1 to your phonopreamp chassis.

The standard Goldring Elan car-tridge is premounted to the tonearm,with VTA preadjusted at the factory.The Elan has a 15-micron spherical dia-

mond stylus. No performance curvewas supplied, as it is with the more ex-pensive cartridges. My Shure V15 TypeV came with a 20-page booklet, withcurves for frequency response andtrackability versus recording velocity.

If you prefer, you can install a differ-ent cartridge. Music Hall supplies aplastic mounting pattern that you setonto the platter spindle. The patternhas three reference points that allowyou to adjust the cartridge mountingscrews so that the stylus-to-pivot andcartridge/headshell alignment are inaccordance with specified dimensions.

I found the tracking force very easyto set up, even without a stylus forcegauge. You simply remove the stylusguard and adjust the counterweight sothe cartridge just floats at the record-playing level. Then you hold the coun-terweight in place and turn the forcedial ring until the top of the scale readszero. Finally, turn the counterweight tothe recommended TF reading (1.7gmfor the Elan). When I checked the TFwith my stylus force gauge, it read

Product ReviewMusic Hall MMF-2.1 and Goldring CartridgeReviewed by Charles Hansen

PHOTO 1:MMF-2.1 front view.


1.8gm, so the counterweight scale ap-pears to be sufficiently accurate.

LISTENING TESTS I used a number of LPs for my listeningaudition. I enjoy jazz and classicalmusic, but I also played some popularmusic (well, popular from the’60s–’80s). I used the phono preamp inmy Control Preamplifier with Head-Room Module (AE 6/97, p. 8), a Para-sound HCA-1000A power amp, andNHT SuperOne’s and SW2P subwoofer.

The MMF and Goldring combinationhas a nice, warm midrange, with a slightly rolled-off treble that com-plements jazz very well. I don’t believeit has a really extended deep-bass response when compared to the Shure

V15 type V, but it is no slouch either. It acquits itself quite nicely for a $60cartridge.

The table and cartridge had no trou-ble with the Shure Audio ObstacleCourse test LP, or the high velocitytracks on the Stereo Review SR12Stereo Test Record. The massed stringsand choral voices were clear and well-defined—much more so than CDs of thesame label and title. The NutcrackerSuite is one of my favorite classicalrecordings, and I have yet to find a CDthat comes close to the Fritz Reiner/Chicago Symphony Orchestra LP (RCARed Seal VCS-7100).

The MMF-2.1 brought out all the finedetails in instruments such as thebrushed cymbals, triangles, acoustic

guitar, strings, and the woody tone ofthe acoustic bass. I think it is a realprice/performance bargain.

MEASUREMENTSEd Dell asked me to take as many ob-jective measurements as I could. I havetwo records with test-tone tracks:Stereo Review’s SR12 Stereo TestRecord (1969) and Hi-Fi News & RecordReview’s Test Record (HFN-001, 1996).Many of these test tracks are designedfor subjective listening evaluation, but Iselected those I believed were best suit-ed for measurement.

I measured the frequency responsefor the MMF-2.1 (Fig. 1) at the phonopreamp output jacks. I established 0dBras 400mV using the 1kHz referencetrack on the SR12 test record. The over-all response was determined by measur-ing the output voltage for the 19 warbletracks covering 20kHz down to 20Hz.

HF response rolled off gradually ateach end of the audio spectrum, but re-mained within the ±3dB specificationfor the Elan. Music Hall recommends acartridge loading of 150–400pF, 47kΩ.My preamp load is 100pF and 47k5.

Channel separation for the six testtracks is shown in Table 1. The resultswere essentially the same for the R-Land L-R tests. Goldring specifies −20dBfor the Elan at 1kHz, and I think that isprobably within my measurement errorrange.

THD+N measurements proved diffi-cult, since the test tones all have over 5%THD+N on the test records I used. Onelow-distortion 3kHz test tone for usewith a wow-flutter meter measured 0.8%THD. Since there is no indication of therecorded THD+N of this test signal, itmay or may not represent the actual per-formance of the phono cartridge.

My LED strobe showed the platterspeed to be very slightly on the fastside, with one of the test bars on thestrobe disk moving one bar width

TABLE 1CHANNEL SEPARATION

FREQUENCY SEPARATION dB6.4kHz −13.63.2kHz −16.71.3kHz −18.4800Hz −19.2400Hz −21.0

ms0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0

mV

-200

-160

-120

-80

-40

0

40

80

120

160

200

Music Hall MMF 2.1, Goldring Elan1kHz Square Wave

FIGURE 1: Frequency response.

ms0 5 10 15 20 25 30 35 40 45 50

mV

-200

-160

-120

-80

-40

0

40

80

120

160

200

Music Hall MMF 2.1, Goldring Elan1kHz Tone-Burst

FIGURE 2: 1kHz square-wave response.


clockwise in 30 seconds. This is of noreal-world consequence. There was noflutter or wow that I could observe withthe strobe, which is synchronized to the60Hz power line.

Figure 2 shows the cartridge re-

sponse to a 1kHz square-wave testtrack. The illustration in the test book-let shows a slight tilt and cupping inthe “ideal” as-recorded waveform, so itdoes not represent a perfect squarewave shape (I wonder whether any cut-

ter stylus can make an instantaneous90° turn). The Elan shows a morerolled-off wave-shape, with a slight os-cillation at the leading edges.

Figure 3 shows the response to a1kHz tone burst. The Elan achievedvery accurate results here, with no evi-dence of response dips, resonance, orspurious response for any tone burst inthe sweep from 500Hz to 10kHz.

Figure 4 shows the MMF-2.1/GoldringElan output spectrum reproducing acombined 300Hz + 4kHz intermodula-tion distortion (IMD) signal. The 4kHzsignal is recorded at 7.5cm/s (−50dB rel),and the 300Hz signal is recorded at9cm/s (−25dB rel). The 4.3kHz IMD prod-uct is −67dBr, and the 3.7kHz product is−61dBr. There are additional responsepeaks at or below −70dBr.

Interestingly, the SR12 test bookletstates that the IMD tones are 400Hz and4kHz. The spectrum analyzer and mycomparison with a 300Hz oscillator toneshow the lower frequency to be 300Hz.

The manufacturer’s specificationsand measured results are shown inTable 2.


TABLE 2MEASURED PERFORMANCE

PARAMETER MANUFACTURER’S RATING MEASURED RESULTSFrequency response 20Hz–20kHz ±3dB 20Hz–20kHz±2.91dBSeparation 20dB, 1kHz 18.4dB, 1.3kHzTracking force range 1.5–3.0gm, 1.7gm nom.Recommended loading 150–400pF, 47kΩCompliance 16µm/mNMM output 5mV, 1kHz, 5cm/sWow and flutter 0.15%, 33-¹⁄₃, DIN 45-507Rumble −70dB, DIN 45-539-BSpeed error ±0.9%Power requirements 2VA

ms0 5 10 15 20 25 30 35 40 45 50

mV

-200

-160

-120

-80

-40

0

40

80

120

160

200

Music Hall MMF 2.1, Goldring Elan1kHz Tone-Burst

FIGURE 3: 1kHz tone burst.

kHz

0 1 2 3 4 5

dB

-120

-110

-100

-90

-80

-70

-60

-50

-40

-30

-20

Music Hall MMF 2.1, Goldring ElanIntermodulation Spectrum 300Hz + 4kHz

FIGURE 4: Spectrum of 300Hz + 4kHz intermodulation.

Web ShoppingMade Easier!Find Out Who’s On-Line… Check Out Our Advertisers& Their Websites!

Look in one place for your favorite sources ofaudio parts, componentsand more!

Also find links to these advertisers in the Marketplace section of our website,

www.audioXpress.com

Product ReviewSovtek 6550 and 2A3 Platinum Matched TubesReviewed by Charles Hansen

Sovtek 2A3 Power Triode, Sovtek 6550Beam Power Tube, New Sensor Corp.,20 Cooper Square, 4th Floor, New York, NY 10003, (212) 529-0466,www.newsensor.com.

INTRODUCTIONI previously reviewed the Sovtek 2A3 inthe 5/00 issue of Glass Audio (“2A3Sovteks,” pp. 46–48). New Sensor fur-nished two more “Platinum Matched”2A3s along with four 6550s for this eval-uation. As before, these tests were strict-ly objective measurements. I did not in-stall them in an amplifier for any subjec-tive testing.

2A3 TEST DATAAt 5.375″ , the tubes exactly met themaximum height of the RCA ST16 Out-line 51. The maximum diameter of theSovtek tube was 2.0″, just under theST16 maximum of 2.063″. The filamentcurrents for both tubes were within thespecified 2.5A maximum at a filamentvoltage of 2.5V AC 60Hz.

I measured the dynamic characteris-tics using the Audiomatica Sofia vacu-um tube curve tracer. Figures 1 and 2are the plate curves for samples 1 and2, respectively.

The variations in plate current at afixed plate voltage of 250V DC and a grid voltage of −48V DC is shown in Table 1. I have also listed the Gm, µ,RP, and P at the selected test points forinformation.

I also ran the Sofia matching pro-gram, and initially the two 2A3smatched with an ERR of 116. This wassomewhat high for matched tubes, so Iran the dynamic test data and thematch program two more times.

ERR is a dimensionless figure ofmerit representing the smallest root-sum-square differences between themeasured values of plate current andplate voltage over the specified range ofgrid-bias values1. The lower the num-ber, the better the match. An ERR ofless than 100 is a very good match.

The second trial resulted in an ERR of103, and the third trial produced an ERRof 98. However, you should note that I PHOTO 1: The Sovtek 6550.


FIGURE 1: Plate curves for Sovtek 2A3 sample 1.

TABLE 1COMPARISON OF 2A3 PLATE CURRENTS AT

Vak = 250V DC, Vg = −48V DC

2A3 IP GM µµ RP P (WATTS) SAMPLE (mA RMS) (mA/V) (kΩΩ)1 27.7 6.34 6.49 1.02 6.942 29.2 6.03 6.34 1.05 7.33

FIGURE 2: Plate curves for Sovtek 2A3 sample 2.

measured an ERR of 88 with two un-matched 2A3s in the earlier 5/00 GA test.

6550 TEST DATAAt 4.47″, the Sovtek 6550 tube was with-in the 4.75″ maximum height of theST16 Octal outline. The maximum di-ameter of the tube, at 1.72″, was underthe maximum of 2.063″. The filamentcurrents for the tubes were within thespecified 1.8A maximum at a filamentvoltage of 6.3V AC 60Hz.

I measured the dynamic characteris-tics of the four “Platinum Matched”tubes using the Audiomatica Sofia in6550 ultralinear mode. Figures 3–6 arethe plate curves for 6550 samples 1through 4, respectively.

The variations in plate current at afixed plate voltage of 450V DC and a gridvoltage of −49V DC is shown in Table 2. Ihave listed the Gm, µ, RP, and P at the se-lected test points for information.

I ran the Sofia matching program for

the four 6550s, and sample numbers 3and 4 matched with a very low ERR of28. The remaining two tubes matchedwith an ERR of 209. The ERR spread forall four tubes (Quartet mode) was 99.


REFERENCES1. Don Jenkins, “The Sofia,” V&T News 1/99, pp.11–19.

TABLE 2COMPARISON OF PLATE CURRENTS AT

Vak = 450V DC, Vg = −49V DC

6550 IP GM µµ RP P (WATTS)SAMPLE (mA RMS) (mA/V) (kΩΩ)1 53.9 6.68 16.6 2.48 24.32 46.2 6.09 17.0 2.78 20.83 46.7 6.33 16.7 2.64 21.04 44.9 6.48 17.1 2.64 20.2

FIGURE 3: Plate curves for Sovtek 6550 sample 1. FIGURE 4: Plate curves for Sovtek 6550 sample 2.

FIGURE 5: Plate curves for Sovtek 6550 sample 3. FIGURE 6: Plate curves for Sovtek 6550 sample 4.

V I S I T O U RW E B S I T E

or E-mail us at [email protected]

Sony CDP-XE400 and CDP-XE500 CDplayers. Sony Electronics Inc., 1 SonyDrive, Park Ridge, NJ 07656, 800-222-7669 or 941-768-7669, FAX 941-768-7790,www.sel.sony.com. CDP-XE400: $160;CDP-XE500: $180. Warranty: one yearparts and labor.

Sony’s CDP-XE400 and CDP-XE500are the least expensive single-play CDplayers from the Japanese electronicsgiant. Inexpensive single-play CD play-ers are becoming increasingly difficultto find, since most manufacturers arepushing CD changers.

The CDP-XE400 and CDP-XE500 play-ers are identical except for a few basicfeatures. The CDP-XE400 is a bare-bones player—no remote, no digital out-put. For $20 more, the CDP-XE500 addsa Sony Remote Commander infrared re-mote control, a Toslink digital output,and digital volume control operatedfrom the remote. The digital volumecontrol affects both the line outputsand the headphone jack.

The CDP-XE400 chassis contains thereceiver for the infrared remote control,so full remote capability (including thedigital volume control) is possible withthe remotes supplied with many Sonyreceivers. You may also control theCDP-XE400 with a “universal remote,”with one caveat: Sony’s track-selectionsystem does not allow direct entry oftracks higher than ten—all Sony re-motes contain a “greater than ten” but-ton (>10) for this purpose. If your uni-versal remote does not have this but-ton, it may not be compatible withthese players. Among the incompatibleremotes is the Radio Shack 15-1994 re-mote I reviewed in aX 3/01.

FEATURESThe Sony CDP-XE400 and CDP-XE500players include a full array of standardfeatures. Play modes include programplay, in which you can program up to24 tracks on a disc to play in any order

you choose, and shuffle play, in whichthe player randomly selects the pro-gram order. The repeat feature can re-peat an entire disc, track, or selectedprogram, or begin another shuffle ofthe entire disc. The edit/time fade but-ton allows you to produce a program tofit any standard cassette tape length.You can also program the player to fadeout at a specific time on the CD.

The peak search feature quickly lo-cates the highest recorded level on aCD, allowing you to set the level on acassette deck at that spot. When you in-sert a disc in the player, the total num-ber of tracks and total playing time aredisplayed. If the disc is not played afterinsertion, the displayed informationdisappears after a few seconds (the dis-play reads “0” tracks).

Pressing the time button restoresthe total time and track informationfor about three seconds. The time andtrack display is always active whenthe player is in the play mode. Thetime button also allows you to togglebetween elapsed time, remainingtrack time, and total remaining timeon the disc.

The music scan feature plays the firstten seconds of each track—when youhear the track you are looking for, sim-ply press the play button to play thistrack from the beginning. You canchange the playing time of each tracksample to 20 or 30 seconds by togglingthe music scan button.

These players do not contain a key-pad for track selection—you select indi-vidual tracks using a rotary control lo-cated in the upper right-hand corner ofthe front panel. Sony calls this controlthe AMS (“Automatic Music Sensor”).Direct keypad entry is supported on theremote supplied with the CDP-XE500,or the remotes supplied with Sony re-ceivers. Both players come with stan-dard double-arrow search buttons. Youcan conduct the search silently, moni-toring the time and track display, bypressing the pause button.

INTERNALS Sony uses what it call a “Hybrid PulseD/A Converter” in these players. Thischip bears Sony number CXD8567AMand is a combination 8× oversamplingdigital filter and bitstream-type D/A

Product ReviewSony CDP-XE400 and CDP-XE500 CD PlayersReviewed by Gary Galo


TABLE 1MANUFACTURER’S SPECIFICATIONS

(IDENTICAL FOR BOTH MODELS)Frequency response: 2Hz–20kHz, ±0.5dBTotal harmonic distortion: 0.0045%Signal-to-noise ratio: 100dBDynamic range: 98dBChannel separation: 95dBLine output voltage: 2V with 50kΩ loadRecommended load impedance: >10kΩHeadphone output: 10mW/32ΩPower consumption: 10WDimensions: 17 × 3¾ × 11 ⁵₈″ (430 × 95 × 295mm)Weight: 6 lbs, 10 oz (3.0 kg)Supplied accessories:

Stereo audio patch cord (both models)Remote control and (2) AA batteries (CDP-XE500 only)

PHOTO 1: The Sony CDP-XE500 CD player with its supplied remote. The lower-cost CDP-XE400 does not come with the remote control, but the internal infrared remote receiver al-lows you to use it with the remotes supplied with most Sony receivers, as well as manyuniversal remotes.

converter with analog voltage outputs.The analog circuitry is about what I’vecome to expect in low-cost players.Each channel uses a JRC NJM4558 opamp for analog filtering and outputbuffering, and a Mitsubishi M5218ALfor the headphone amplifier. TheM5218AL will sink ±50mA of load cur-rent, making it suitable for headphoneamplifiers.

The CDP-XE400 and CDP-XE500 play-ers are surprisingly well-built, particu-larly considering the price. The chassisand cover are all-metal, and the front-panel controls have a solid operationalfeel. The mechanism is quite sturdy,and the CD tray operates smoothly. Theisolation feet are mounted with thescrew holes off-center to help diffuse vi-bration modes.

Both players bear the designation“Made in China” and, in case youhaven’t noticed, the quality of Chinesemanufacturing has improved consider-ably in recent years. Overall, these play-ers exhibit none of the chintzy look andfeel so prevalent in inexpensive con-sumer electronics.

PERFORMANCE At these prices, the sonic performanceof the CDP-XE400 and CDP-XE500 can’tbe beat—these players are as good asany I’ve heard for under $200, and bet-ter than most in this price range. Al-though Toslink is not the best digital in-terface, the CDP-XE500 makes a re-spectable entry-level CD transportwhen used with an outboard digital

processor. I recommend a high-qualityToslink interconnect, such as the Kim-ber OPT1 I reviewed in AE 4/99.

Looking for an under-$1000 systemfor your son or daughter? Try matingthe CDP-XE500 with Sony’s STR-DE445receiver (street price $250) and a pair ofBoston Acoustics CR-9 loudspeakers($420 per pair; www.bostonacoustics.com). You can operate the Sony receiv-er in the two-channel stereo mode or asa five-channel Dolby surround unit(with additional speakers, of course).

The receiver has both Toslink opticaland S/PDIF coaxial digital inputs, soyou can digitally link the CDP-XE500CD player to the receiver, which alsocomes with a Sony universal remote tooperate the CD player. I assure readersthat I haven’t abandoned perfectionist,high-end audio, but I do find it remark-able what $820 will buy these days, ifyou pick your components carefully.

Over the past two years we at TheCrane School of Music, SUNY Potsdam,have purchased around 20 of theseplayers for use in the faculty studiosand music library. About six monthsago I installed a dozen CDP-XE400 play-ers in library listening carrels, wherestudents tend to give any piece of elec-tronic equipment a real workout. So far,they are holding up extremely well. Mydealer told me that several local radiostations are using these players, withsimilarly positive reports on reliability.

These players typically sell for about$30 below the manufacturer’s suggestedretail price, whether through mail order

or a local retail dealer. If there is nodealer near you, try J&R Music World(www.jandr.com). At these prices theSony CDP-XE400 and CDP-XE500 are agreat bargain, and are my first choicesamong under-$200 CD players.


PHOTO 2: Inside view of the CDP-XE500 (the CDP-XE400 is virtually identical). The buildquality belies the low cost of these players.

We Want Your Feedback!audioXpress encourages readerfeedback in the form of letters,

queries, and comments.

Send correspondence to:

Editorial Dept.PO Box 876Peterborough, NH · 03458-0876

Phone: (603) 924-9464Fax: (603) 924-9467

Website: www.audioXpress.comE-mail: [email protected]

audioXpress reserves the right to edit letters for length and clarity. Letters should be brief

and to the point.

Book ReviewCunningham Radiotron ManualReviewed by Scott Frankland

To say that the triode is alive and wellseems almost a truism nowadays. Wit-ness this vintage 1934 tube manual,newly minted by Audio Amateur Press1.This edition of the Cunningham Ra-diotron Manual was originally pub-lished at the peak of the triode era. As

such, it lent technical support to RCA’sgrowing lineup of vacuum tubes, in-cluding the 2A3, first released in 1933.Once seen as a relic by “progressive”audiophiles seeking better sound viahigh-power beam tubes, the 2A3 is nowseen as a legitimate contender for bestpower tube of the 20th century2,3.

What first impressed me about thisbook is the striking way in which thecover announces its contents (Photo 1).The invitation is almost palpable; youneed only turn the page. So, what then,do you find inside?

THE CONTENTSThe first part of this 154-page manualcontains the familiar RCA primer ontubes, wherein tube lore of the most es-sential sort is laid out, including theusual caveats and operational hints.This section is often overlooked, but itretains its value as a boiled-down state-ment of what is most essential inputting tubes to work.

Following this is the central core ofthe book: the tube data. All of the tubes

listed were specifically designed for ei-ther power amplification or radio recep-tion. There are no television tubes, nohearing-aid tubes, and no computertubes. Anyone seeking definitive datafor power triodes such as the 2A3, 10,26, 31, 37, 45, 50, and 71-A need look nofurther (Photo 2).

Nearly every tube is accompanied bya graph of its plate characteristics, alongwith operating points and applicationnotes. In many cases the recommendedloadline is drawn across the plate char-acteristics, and the correspondingpower output is given. Small-signalaudio tubes are also represented, butthese usually combine their functionwith other tube elements designed for rfamplification, detection, or conversion.

For each tube type there is what ap-pears to be a very fine charcoal drawing(or duotone) of the tube itself. There are,in addition, numerous photos not usual-ly found in manuals—a dozen of whichdepict tubes on the assembly line (Photo3). These photos are welcome, as theyhelp put the book into context.*


PHOTO 1: Cunningham Radiotron Manual.

Those who consider pentodes inter-esting will find them well documentedhere. A number of vintage tubes arerepresented, such as the 2A5, 6A4, 33,38, 41, 42, 43, 47, and 59. In some cases,plate curves for triode-connected opera-tion are shown.**

There is also vintage-tube rectifierdata covering the 1-v, 5Z3, 12Z3, 25Z5,80, 81, 82, 83, and 84. The type 80 and 81include full families of graphical data,including a schematic of a typical appli-cation, in this case a complete choke-input power supply.

Three pages at the back of the bookare devoted to tube testing—the how-tosand the wherefores. Following this is acollection of circuits, most of which areradio-related, but some of which areamplifier-related. There are, for exam-ple, a push-pull triode circuit with inter-stage transformer, and a single-endedtriode circuit with input transformer.

Lastly, there are pages devoted totube dimensions, an index, and a read-ing list of vintage textbooks. In short,this is a thoroughly executed model ofwhat a good manual should be.

CUNNINGHAM AND DE FORESTI would be remiss if I didn’t remark onthe mysterious figure of Elmer T. Cun-ningham, one of the most prolific boot-leggers of the early tube era. From allreports, Cunningham was hugely suc-cessful on the West Coast, with a devot-ed following in the amateur market.

Then again, he was a thorn in the sideof de Forest, the inventor of the Audion(triode). De Forest wished to sell com-plete receiving kits along with his tubes,but Cunningham cheerfully undercuthim by offering raw tubes. Rather thanAudions, Cunningham labeled his tubes“AudioTrons” (Photo 4). The Audiotronwas made in the San Francisco BayArea and was the most widely sold of allthe independent brands4.

Cunningham’s ads boldly stated:“Every AudioTron is guaranteed to

have a life of 1,000 hours and to arrivein perfect condition.” One-thousand


ABOUT THE AUTHORBrought up as a musician, Scott Frankland studied phi-losophy of science at college, and in 1976 he turnedhis attention to electronics. Throughout the ’80s hewas a co-designer at MFA Systems, manufacturer ofthe Luminescence preamp. He was subsequently prin-cipal designer for Wavestream Kinetics, where his con-tinuing mathematical studies helped him secure apatent in 1996, covering multi-tube triode amplifiers operating in class AB2. He can be reached at [email protected].

PHOTO 2: Sample page.

PHOTO 3: Testing tubes on the assembly line.

hours was a long time in 1915, but Cun-ningham’s guarantee was evidently de-pendable. It is said that he (or his staff)discovered a new chemical reactionthat allowed Cunningham to manufac-ture long-life tubes ahead of his time5.

Nonetheless, in 1916 de Forest filedsuit against him for patent infringe-ment. Cunningham promptly postedbail and continued selling Audiotrons.Eventually, he and de Forest settledtheir differences, and apparently ajoint-venture of sorts ensued.

CUNNINGHAM AND RCAThe Radiotron label had been extantsince 1920, when RCA issued its firsttubes: the UV-200 and UV-201. Thatsame year, RCA brought suit againstCunningham. The upshot was that Cun-ningham agreed to be-come a distributor for RCA.It is clear that Cunning-ham’s long experience in thetube market afforded him se-rious leverage at the bargain-ing table.

As a distributor for RCA, Cunning-

ham wrote his own ticket. He boughtonly those tube types he chose to sell,and he persuaded RCA to label themwith his own name. In return, Cunning-ham sold an enormous number of tubesfor RCA.

Throughout the radio boom, RCAfunctioned as a joint distributing armfor GE and Westinghouse—those twoold rivals in the electrical business.RCA was a sales company, not a manu-facturer. In 1929, RCA purchased theVictor Talking Machine Company ofCamden, N.J.6. The resultingRCA Victor Company wasthe first step in DavidSarnoff’s grand unifi-cation schemet o

transform RCA into an integrated man-ufacturing, engineering, and sales enti-ty. Victor’s factories would provide thenecessary facilities for manufacturing.

As part of a complex deal with GEand Westinghouse, RCA spent millionsto adapt the Victor factories, and there-after began to manufacture radio setsusing Victor cabinets and phono-graphs. In April 1930, RCA obtainedcontrol of GE’s tube manufacturingplant in Harrison, N.J.7,8.

In May 1930, the US governmentfiled suit against the “Radio Trust”(GE, Westinghouse, AT&T, and RCA),and by 1932 forced the combine tobreak up. For Sarnoff, however, thisproved a blessing in disguise. RCA’sindependence was by this means as-sured, and Sarnoff’s unification plan

was fulfilled. The following year, 1933, saw the ar-

rival of the 2A3. It turned out to be a verygood year for audio, as this was thesame year in which Western Electric re-leased the 300A (a 300B except for the lo-cation of the bayonet pin9). In the mean-time, the radio boom having fizzled fol-lowing the market crash of 1929, Cun-ningham ads began to carry the tagline“A Subsidiary of Radio Corporation ofAmerica.” By 1931, Cunningham Inc.had been absorbed by RCA10, and Cun-ningham himself entered their employ.

But Cunningham soon got his re-venge. By 1933, he (Photo 5) had takencharge of the manufacturing division!For Cunningham, 1933 was a very goodyear indeed. From then until 1935, alltubes made by RCA were marked,“RCA Cunningham Radiotron.” After1935, the Cunningham name was grad-ually phased out, and RCA again tookcenter stage.

THE LAST WORDAudio Amateur Press has done a thor-oughly professional job in the construc-tion of this reprint. The paper, the re-productions, and the typography are allextra fine, as befits the subject matter.The Cunningham Radiotron Manualwas originally priced at 25 cents, butyou will have to pay a bit more for ittoday: $14.95. That, however, is lessthan half of what they’ve been tradingfor on Ebay lately—35 to 50 bucks acopy (circa February 2001). So this is a


PHOTO 4: Early unbased radio tube,

circa 1916–1919. Paper label is missing but could be

a Cunningham Audiotron or similar. There are no stampings on the

cylindrical plate (from Kilokat’s website).

Accuracy, Stability, RepeatabilityWill your microphones be accurate tomorrow?

Next Week? Next Year? After baking them in the car??? ACO Pacific Microphones will!

Manufacturered to meet IEC, ANSI and ASA standards.Stainless and Titanium Diaphragms, Quartz insulators

Aged at 150°°C.Try that with a “calibrated”consumer electret mic!

ACO Pacific, Inc.2604 Read Ave., Belmont, CA 94002

Tel: (650) 595-8588 FAX: (650) 591-2891 e-mail [email protected]

ACOustics Begins With ACO™

no-brainer. The Cunningham RadiotronManual is one of the classic databooksof the vintage triode era.

* You can see more such photos inDeketh’s classic text, Fundamentals ofRadio-Valve Technique.1,11

** Pentode fans may also wish to obtainValves for Audio Frequency Ampli-fiers,1,12 which contains extensive datafor the EL-34 and EL-84.


REFERENCES1. Audio Amateur Press, Peterborough, N.H., 2000(available from Old Colony Sound Lab at 1-888-924-9465 or from audioXpress.com).

2. “SE 2A3 Listening Evaluations,” Vacuum Tube Valley,Issue 12, Lakeport, Calif.

3. “Moondog 2A3,” Sound Practices, http://www.welbornelabs.com/sp_2a3rev.htm.

4. Gerald F.J. Tyne, Saga of the Vacuum Tube, 1994,Prompt Pubs., Tempe, Ariz.

5. http://www.ieee.org/organizations/history_center/oral_histories/transcripts/lederer4.html.

6. Gleason L. Archer, Big Business and Radio, Ameri-can Historical Co., Inc., N.Y., 1939.

7. http://www.sarnoff.com/sarnoff_story/history/.

8. http://www.rca.com/content/viewdetail/0,1407,EI99-CI263,00.html.

9. Bernard Magers, 75 Years of Western Electric TubeManufacturing, Second Ed., Antique Electronic Supply,Tempe, Ariz., 1994.

10. John W. Stokes, 70 Years of Radio Tubes andValves, Second Ed., Sonoran Publishing, Chandler,Ariz., 1997.

11. J. Deketh, Fundamentals of Radio-Valve Tech-nique, 1949, N.V. Philips, Eindhoven, reprinted 1999by Audio Amateur Press, Peterborough, N.H. Re-viewed, Glass Audio 1/00.

12. E. Rodenhuis, Valves for Audio Frequency Amplifiers, 1954, N.V. Philips, reprinted 1997 by AudioAmateur Press, Peterborough, N.H. Reviewed, GlassAudio 5/99.

PHOTO 5: Elmer T. Cunningham, President,RCA Manufacturing Co., Inc (RCA Victor).

Atma-Sphere Music Systems, Inc.160 So. Wheeler St., St. Paul, MN 55105

651-690-2246 FAX 651-699-1175www.atma-sphere.com

You’ll be hearing more from us!

Award-winning patented circuitry - World Class performance.

For 2 decades, Atma-Sphere has remained the unsurpassedleader in audio amplification. Our patented circuitry providesthe best of both single-ended and push-pull operation withouttransformers, without compromise.

The S-30amplifier:•All Triode•Class A•single gain stage•Circlotronic™ Output •Balanced Differential Design® •Handwired- NO circuit boards!•finest quality parts and materials,including custom manufacturedwire.•proven, reliable circuit•Standby feature•Classic styling

“I lust for this amp!”-Sam Tellig, STEREOPHILE

AUDIO TRANSFORMERSAMERICA’S PREMIER COIL WINDER

Engineering •• Rewinding •• Prototypes

McINTOSH - MARANTZ - HARMAN-KARDONWESTERN ELECTRIC - TRIAD - ACROSOUND

FISHER - CHICAGO - STANCOR - DYNACOLANGEVIN - PEERLESS - FENDER - MARSHALLELECTROSTATIC SPEAKER TRANSFORMERS

Williamson Amplifier Transformer SpecialistWE DESIGN AND BUILD TRANSFORMERS

FOR ANY POWER AMPLIFIER TUBE

PHONE: [414] 774-6625 FAX: [414] 774-4425AUDIO TRANSFORMERS

185 NORTH 85th STREETWAUWATOSA, WI 53226-4601

E-mail: [email protected] CATALOG CUSTOM WORK

Xpress Mail

PARALLEL FEEDWith regard to Barney Vincelette’scomments considering direct feed

versus parallel feed (“Letters,” GA 6/00),I agree with his opinion that normal SEtransformers are highly linear inductors.There has been increasing hype towardschoke (also a linear component) loadedtubes having a tiny OPT coupled with acapacitor. First, there was a tendency tohave huge linear irons and to eliminateall capacitors from the signal chain; nowit’s okay to add a cap to a critical place ofthe chain using minimal sized ironstacks. The idea is some 60 years old.

Without further criticism, there maybe a difference in sound. My purposehere is to offer one explanation. In thebeginning of my career I was involvedin designing transformers, mainly fer-rite cores for telephony carrier-waveequipment.

A common trick was to add a seriescapacitor to feed a transformer from aresistive source to a resistive load. Ifyou have a suitably sized capacitor, aseries resonance is present with thetransformer primary inductance form-ing a filter that extends the bandwidthabout half an octave downwards for afinal high-pass slope of 12dB/octave.

In the amplifiers’ case, you may havea capacitor of several microfarads andan inductance of tens of henries, fed bythe resistive triode impedance, loadedalso approximately resistively. Al-though probably not on purpose, theresonance mentioned may hit the areaof 5–30Hz and, with good luck, evengive some flattening to the response.

There is, however, another aspectthat may be more important. Down-wards, towards the resonance frequen-cy, the load impedance the tube seesalso falls markedly due to the reso-nance, and has sonic influence. Thebass to a lower transformer load ismore rigid than to a higher one usingthe same tube, at the expense of distor-

tion. More study is needed to look intothis practice.

Mauri PännäriFinland

MTM FEEDBACKAs a satisfied user of an MTM 18speaker system from Speaker City

(and after reading the dismal review inMarch 2001), I feel compelled to passalong some suggestions for optimizingthe speakers. I did not purchase a kitfrom Speaker City, as I had been usingmost of the drivers in a previous proj-ect. Since I was purchasing only thecabinets and crossovers, I was on myown in terms of optimizing the internalbracing, damping, and wiring.

My first recommendation is to pur-chase the 9500 tweeters rather than the9900s. They offer 95% of the perfor-mance of the Revelators at about halfthe price. Note that when purchasingthe crossovers that are designed to beused with the 9500s, both high- and low-pass sections are on a single board thatfits perfectly on the back wall of thecabinet. No need to reposition the bind-ing posts. About half of the money yousave here will be spent tweaking the in-sides of the cabinets and wiring up thecomponents.

The cabinets benefit from “bracing”as described by Mark Wheeler inSpeaker Builder 8/99 (“NavigatingSpeaker Design”). I used narrow stripsof solid oak, epoxied vertically alongthe center of the side and rear walls. Ialso glued Deflex panels (four per cabi-net, available from Madisound) alongthe side and back walls to minimizestanding wave problems and further re-duce panel resonances. Lastly, a mod-est (less than 1 pound per cabinet)amount of long fiber wool spread evenlythroughout the box seemed to eliminatethe last little bit of congestion (slapback through the woofer cones?). BTW,

the drivers take a long time to break in. I used Analysis Plus Oval Theater

(CL-3) speaker cable for all internalwiring. This cable is slender, very flexi-ble, and easy to work with. It is an oddshape, but crimps or solders easily. It isvery reasonably priced (less than $3 perfoot, and you need only 10 to 12 feet; Igot mine from Greg Lucas at MUSIC foryour EARS, 216-296-2263), and sonicallyit is wonderfully musical.

Neil [email protected]

I read with great interest Joe D’Appoli-to’s review of the Speaker City MTM-18kit. I noted your belief that the internalvolume of the speaker was too small. Iam interested in using these same driv-ers in a design of yours, the 717 forA&S. I computed the volume of the 717to be approximately 33 ltrs withoutbracing and other losses. This seemsclose to the MTM volume.

In addition, I have built an ARIA 5with the Raven, which seems to exhibitthe same problem at the bottom end asthe MTM-18 and has less volume thanyou seem to recommend for the MTM.I know that you can’t give individualsspeaker-building advice, but since youseem to give contradictory volume rec-ommendations, I wondered if you canclarify this.

Harold Goldman [email protected]

Joseph D’Appolito responds:

The A&S 717 was designed in the early’90s. Parameters for the Scan-Speak180mm woofer have changed substantiallysince that time. In particular, VAS has in-creased greatly, requiring a larger enclosure.

I am not aware of any low-end problemwith the ARIA 5 as I designed it, but, again,driver parameters have changed over the in-


tervening years. I have not made any updateto the enclosure design. It is quite possiblethat the volume originally specified is nolonger valid.

GOOD CHANGEWell, the change from Audio Elec-tronics to audioXpress is perhaps

giving me more tube electronics than Iwould desire, but, on the whole, thecombination of three publications intoone makes sense and I like it.

Regarding the Jan. ’01 issue, I hadpreviously read Anthony New’s articlein Electronics World, and I agree withsome of the response letters, which areworth mentioning. First, IM distortionmeasurements in audio have beenaround many, many years and were fre-quently, if not always, included as partof an amplifier specification.

Second, IMD and (T)HD are both mea-sures of linearity and go hand-in-hand. Areader might well conclude from Mr.New that if an amplifier sounds “bad,”yet has low HD throughout the audiorange, IMD measurements will revealthe problem. Wrong! There is no mecha-nism that I know of whereby an amplifi-er can have low HD and high IMD, orvice versa. I would agree, however, thatan inadequate HD measurement, say atonly 1000Hz, would not reveal problemsat lower and higher frequencies than astandard IMD test at 60 and 7000Hzmight, and in this sense an IMD test issuperior and faster.

Some years ago, perhaps in the ’60sor ’70s, the Consumer Reports peopletested power amps by an IMD methodthat I thought was great. As I recall, theyfed the amp with white noise filtered toremove frequencies below 500Hz or soand then measured the total amplifieroutput at all frequencies below 500Hz.This low-frequency output could only beas a result of IM distortion. Pretty neat!

Finally, Dr. Thagard’s article on aphono preamp design reminded methat I once worked out for my ownamusement a QuickBasic program for ageneric feedback phono preamp intowhich you could plug in various resis-tor/capacitor values and compare theresults with the standard RIAA values. Iwent back and polished up the programand made it more user-friendly, I hope.

Any reader interested can E-mail me

([email protected]) for my address andfor $1 (or a blank floppy and a couple ofstamps), I’ll mail back a disk contain-ing the BASIC program and also a com-piled stand-alone program for thosewho don’t know about or don’t care tobother with the Microsoft Qbasic pro-gram that is included with Windows 95and 98 (and, probably, Win 3.1) or inDOS versions 5.0 and higher.

Lawrence Wallcave Santa Rosa, Calif.

GOLDEN CALCULATIONWhile working on a new project, Iread “Determining Optimum Box

Dimensions” by Louis C. McClure (SB2/00). He presents a way of calculatingthe dimensions according to the goldenratio.

Being inspired by this article (thanks,Louis!), I found an easier and more accu-rate method of doing so, by using the lit-tle-known “solver” function in MicrosoftExcel. In this case, it solves the box vol-ume (defined by you) by changing thebox dimensions, where the ratios be-tween the dimensions are defined by thegolden ratio (0.618).

To do this you first determine cell A1as 1, A2 as “=0.618*A1” and cell A3 as“=0.618*A2”, where A1 to A3 defineeach dimension of the box. The box vol-ume is represented in cell A4 and de-fined as “=A1*A2*A3.”

Next, go to “tools” and then “solver.”The target cell is A4 (box volume),equal to the value of (fill in the desiredbox volume), by changing cell A1 (theonly independent dimension). Click“solve” and your optimum dimensionsare there! For example, if the box vol-ume is 100 ltr (100 cubic decimeters),the dimensions are 7.51, 4.64, and 2.87decimeters (75.1*46.4*28.7cm). Ofcourse, you can use this for any unit oflength, and any other ratio.

If the solver function does not appearunder the tools menu, you’ll need to in-stall it first by going to “tools” and then“add-ins.” Click on “solver add-in” andthen OK. Now the solver should appear.

The values are extremely accurate,since the solver’s accuracy is about amillionth of a percent by default. Mc-Clure’s method was found to be about2% accurate for each dimension (it all


Improve the imaging and clarity of your speakers byreducing cabinet resonance and baffle diffraction. Ab-sord standing waves in your listening room. Make yourcar or studio whisper quiet. Coustasheet, a .083" thick,1 lb. per sq. ft. loaded vinyl, will do all this for only$2.98 per sq. ft. (12 min.) plus $5 handling. FREE SHIPPING in USA or 30% discount on full rolls.

oustasheet®

50 _________________________

40 _________________________

30 _________________________

20 _________________________

10 _________________________

TYPICAL TRANSMISSION LOSS

100

250

500

1K 2K 4K 5K Hz

dB

GASOLINEALLEY,L.L.C.1700EastIron, Salina,KS674011.800.326.8372Web: www.gasolinealleyllc.com

Damping Material

depends on rounding errors). Onceyou’ve set it up it’s very easy and fast toget the optimal dimensions for a newbox volume.

Justus VerhagenOxford, England

TUBE WARM-UPIn Neil A. Haight’s letter entitled“Buyer Beware” (Jan. ’01 issue), he

refers to certain brands and types oftubes (Telefunken and Mullard 12AX7/12AU7/ 12AT7s, among others) having a

“defect,” which causes their heaters toflare brightly when initially powered up.This is not a defect in the tubes—rather,they are designed to warm up and reachoperating temperature quickly.

Almost all metal filament resistors(including incandescent lamp fila-ments and tube heaters) have a lowerresistance when cold than when hot—which causes them to draw a “warm-upsurge” when connected to a constantvoltage source. In these tubes that ef-fect has been maximized to provide aquick startup.

Offhand, I can’t recall where Ilearned about these, but I certainly re-member running across them—andthey do warm up more quickly thantheir “normal” counterparts. I don’tknow whether they normally have ashorter life expectancy than standardtubes, but I would not be surprised tofind that heater supplies that are capa-ble of supplying unusually largeamounts of current might provide anunexpectedly large warm-up surge andblow them out prematurely as well.

Keith LevkoffMassapequa Park, N.Y.

Neal A. Haight responds:

I would like to begin by thanking Mr. Levkofffor taking time to respond to my recent let-ter, regarding the behavior of the heaters inEuro 12AT7s/U7s/X7s.

After reading his letter, I decided to ex-periment to see whether I could verify whathe stated about Euro tubes being designedfor quicker warm-up. To do this, I loaded aTelefunken 12AX7 and a U.S.-made 12AX7into my stereo phono preamp. As is usuallythe case, each 12AX7 contains both preampstages for each channel.

Next, I played a monophonic recordalbum, i.e., placed the needle on a revolvingdisc, and then switched on the preamp withaudio amp already on, and set to normal lis-tening level. The result was that each chan-nel took the usual 11 seconds to operate. Iverified this visually with the aid of vu me-ters in a hooked-up tape recorder.

In conclusion, unless there is a lab testthat may prove otherwise, it appears that theEuros do not reach operating temperatureany faster than U.S. makes. It is also possi-ble that quicker warm-up may, however,have been the intention of the Euro manu-facturers, even if it didn’t materialize.

It won’t be long before Mullard and Tele-funken 12AX7s will go for no less than$100, due to the ever-dwindling supply.While I would never attempt to discourageanyone from buying them, I would, however,encourage anyone who does to heat themwith DC voltage. Not only will they lastlonger, they will also sound better, too.Those who don’t have big bucks to spend ontop-notch brands of 12AX7s can rest assuredthat affordable versions continue to improvein performance and reliability.


Issues #14 and 15 of Vacuum Tube Val-ley (PO Box 1499, Lakeport CA 95453, [email protected]) contain a two-partevaluation of 12AX7s, which may provehelpful to those trying to decide which brandof 12AX7s to purchase. It is important tonote that 12AX67s that sound good in guitaramps may not necessarily sound good in hi-fiamps, and vice versa. Until you decidewhich 12AX7 sounds good to your ears,avoid spending big bucks on bulk quantities.Also, make sure you know a tube retailer’sreturn policy—many retailers do not allow re-turns under any circumstances. Good luckand happy hunting.

POWER-SUPPLY DESIGNIf a little bit of filter capacitance isgood, then is a lot of filter capaci-

tance better? For example, Thagard’sexcellent article, “A Phono Pre-Pream-plifier for the CD Era” (audioXpress,Jan/Feb 2001), points out what wouldhappen if you try to build a power sup-ply with very small ripple (10mV) byusing only a capacitor input filter. Heestimates that even on his small 25mApower supply, you would get narrow 5Acurrent pulses flowing through thediodes, capacitors, and transformer.His solution was to use a two-stage fil-ter to minimize these current pulsesand reduce the ripple.

What he doesn’t mention is thatthese same narrow current pulses alsoflow through and radiate from thepower lines. While their fundamentalfrequency is only 60Hz, because of theirnarrow width, they will contain harmon-ics throughout the entire audio band.

If a 25mA power supply can causeampere size current pulses, considerwhat the much larger power supplies inyour PC, TV, power amplifiers, andother electronic equipment are doing.They are all conducting at the sametime—the peak of the line voltage wave.If you look at the voltage waveform ofthe power lines, you will probably no-tice that the tops of the sine wave havebeen flattened as these current pulsesload down the line.

In the standard 120/240V systemused in US homes, these current pulsestend to flow in one hot wire and out theother. Only the unbalanced currentflows in the neutral wire. But in an in-dustrial, or power distribution wye-con-

nected three-phase system, the currentpulses on each phase occur at differenttimes and all return through the neutralwire. The three sine-wave line currentsin a balanced three-phase system nor-mally add up to zero and no currentflows through the neutral wire, but inthis case they all add in the neutral line.

When these current pulses get to adelta-connected utility transformer,they cause a circulating current thatheats up the transformer. In officebuildings where there is a PC on everydesk, there have been cases where theinsulation on the neutral wire hasburned, due to these excessive currentpulses. This is because the heating de-pends on the square of the current—alarge narrow pulse will cause moreheating than the same charge in asmaller wider pulse.

Even if the contribution of yoursmall home audio system is not a prob-lem, it would be wise to design more en-vironmentally friendly equipment. Ithink a better solution would be to de-sign power supplies with 10 to 20% rip-ple and use a three-terminal regulator


to reduce this ripple to a reasonablelevel. This would significantly reducethe stresses on the transformer, diodes,filter capacitors, and the environment.

Walter SchillingerSanta Cruz, Calif.

Norman Thagard responds:

I had heard of the problems that Mr.Schillinger mentioned, but did not know thatcapacitor-input filters in line-powered equip-ment were the culprits. There is a tendencyto take a casual approach to power-supplydesign, but perhaps we should not, for thereasons that he gives.

In my summer semester special topicscourse, “Feedback Amplifier Principles,” Iteach this concept and assign a homeworkproblem similar to the computation of thepreamp capacitor-charging pulses that wasgiven in the article. At least my students un-derstand the ramifications of large capacitorsin capacitor-input filters.

It is too bad that transistor amps are rela-tively low-voltage, high-current components.Tube amplifier power supplies could eitherhave a choke-input or capacitor-input LC fil-ter that did not generate such huge currentspikes. This is possible when a power supplyprovides only a few hundred milliamps, butit is not an easy matter to find a choke of 1Hthat won’t saturate with several DC ampsthrough it.

Such phenomena as Mr. Schillinger de-scribes are in the purview of power engineer-ing, which is a specialty within electrical en-gineering. There is a recent renewal of inter-est in the field and in its practitioners andstudents for obvious reasons.

SERIES CROSSOVER NETWORKHere are some comments on XpressMail published in the May ’01 issue.

Angel Luis Rivera has written a longand cogent response concerning the ap-plication of series crossovers to loud-speaker design, and I wish to throw inmy two cents’ worth.

The one property of the first-order(6dB/octave) series network that bearsrepeating is that the sum of the volt-ages across the two drivers is alwaysequal to the amplifier voltage. The prac-tical consequence of this is that aberra-tions induced by one of the drivers orby an inductor will be compensated for

by the other driver, so in a sense, thistopology is “self-healing.” In practicaluse, this constitutes an advantage.

Another advantage, and one notmentioned by Angel, is that the sourceimpedance seen by a driver, lookingback into the amplifier, is low outsideof the driver’s working range in addi-tion to being low within the driver’sworking range. For instance, with 8Ωdrivers in a two-way system, both thewoofer and tweeter will see an imped-ance approaching 0Ω away from thecrossover frequency, and an impedanceof 8Ω right at the crossover frequency.This helps their sound by damping outdiaphragm resonances both within andoutside of their working range.

In the 1970s I designed a four-waysystem using the Electro-Voice T35Bhorn tweeter and accompanying Elec-tro-Voice X36 12dB/octave crossover.The lower three drivers were run fromthe X36’s low-pass through a compoundseries first-order network. I cascaded6dB/octave sections, each one a seriesconstant-voltage network, to achieve12dB/octave rolloffs for the 12″ woofer’slow-pass and a 2″ dome midrange’shigh-pass, and 6dB/octave for high-passand low-pass on a 5¼″ filler driver work-ing between 500 and 800Hz.

This four-way system is linear phasefrom about 10kHz all the way down tothe woofer’s fundamental resonance (Iused a Bessel fourth-order reflex align-ment). This setup has extremely com-fortable and dynamic sound. Since a6dB/octave slope by itself is not steepenough for most purposes, you mustelaborate on it and adapt it to a collec-tion of drivers that make sense in a par-ticular application.

All four drivers had Zobel networksadded across them as part of thecrossover network, and several resistorswere added strategically so that L-padswould work properly with the domemidrange and horn tweeter, without al-tering the frequency response duringadjustment. The horn tweeter was time-adjusted with the dome midrange.

The trick to making it all work, as Imentioned, was to choose the crossoverslopes and frequency ranges to matchthe driver complement, and to build itup with cascaded series first-order sec-tions. The crossover concept and the


TubesPartsBooksSuppliesPhone, Fax, or e-mail today for our free 80 page 2001 catalog!

6221

S. M

ap

le A

ve. T

em

pe

, AZ

8528

3 U

SA

Phone: 480.820.5411

Fax: 480.820.4643or 1.800.706.6789

ww

w.

tu

be

sa

nd

mo

re

.c

om

driver complement needed to be speci-fied as an integrated system.

Victor Staggs [email protected]

CURRENT PULLDOWNI have some questions concerningop-amp applications and a little bit

of input on various topics, as well. It’sbeen a few years since current pull-down on op amps has been discussed,and I wonder whether anything haschanged, or if indeed it is useful onnewer op amps such as the AD817 ormaybe the AD825. The latest I recall in-volved 2mA; I use the LM334Z with 33Ω

bias. I have always used a heatsink onthe chip when applying pulldown. Isthis really necessary?

I’ve been having trouble findingAD827 dual op amps. Are there otherdual op amps of similar or better perfor-mance out there?

Apparently a number of people pre-ferred the AD825 to the AD817 in Mr.Jung’s super regulators. Would theAD825 be a good bet for a line stagealso?

I’ve been looking for an upgrade formy LT1115-based phono preamps (TAA2/90). I’m also looking for a use for allof the AD797s I pulled out of my superregulators. Would the AD797s work

well in the phono preamps? If not, whatwould? Are there any other ideas forthe AD797s?

A recent issue of Stereophile men-tioned that Holco resistors are nolonger non-magnetic. It’s easy to tellthe newer ones—the steel leads aremuch stiffer (you could also use a mag-net). They sound pretty much like “Re-sistas” to me. Welborne Labs (www.welbornelabs.com) is shipping Vishay-Dale RN60Ds in place of Holco H4s. Ilike them; Mouser (www.mouser.com)also sells them.

Here are some tips for anybody whomight be trying to quickly put togethera good, inexpensive amp. MCM sellspower-amp modules—discrete transis-tor Class AB boards complete with(rather small) heatsinks—by a companycalled Cebek. They also sell power sup-plies for them (a bit wimpy).

I bought a pair of 25-watters (variouspower ratings available), did a resis-tor/cap upgrade including some addedon-the-board rail capacitance, and usedan oversized choke-input power supply.It sounds better than my “baseline”Adcom GFA 5200—even to my wife!MCM (www.mcmelectronics. com)“AXON” and Parts Express “Dayton”(www.partsexpress.com) poly capswork well for projects like this; take alook at them and you might have a sus-picion about who actually makes them.I also like Nichion “Muse” electrolyt-ics—try Michael Percy for them.


FIGURE 1: Sample pulldown current.

Finally, before you spend majormoney on line conditioners, look at theSola CVS 2000W constant voltage trans-formers that Fair Radio (www.fairradio.com) has for $250 each, 2 for$450. The Navy uses these extensivelyin small-signal applications where min-imal line noise is a must and they reallywork.

M.B. WilcoxAltamont, Tenn.

Charles Hansen responds:

I have used pulldown biasing to cure“scoop-out” and crossover distortion prob-lems in the LM324 quad single-supply opamp (Fig. 1). In terms of audio use, GaryGalo gave us three excellent examples ofcurrent pulldown using a resistor, a con-stant-current FET, and the LM334 in the“New Buffers” designs by Gary, RichMarkell, and Walt Jung (TAA 2/90 p. 48).The pulldown current is indeed 2mA and bi-ases the op amp into Class A operation. Ad-ditional discussions on the topic can befound in TAA 1/91, pp. 48-49; 1/94, p. 37;and 3/97, p. 39. The authors of the 2/90article reported cleaner, smoother, and moredynamic sound using pulldown biasing.

The 2mA pulldown current adds addition-al power dissipation to the IC. Whether thisis a problem or not can be determined by aworst-case power dissipation analysis. I willuse the AD825 data sheet for the analysisinformation.

Assume a 500Ω load and ±15V DC sup-ply rails. You find the worst-case supply cur-rent (over the full temperature range, sincethe chip will run hot) is 7.5mA or 225mW.The 2mA pulldown bias adds 30mW to theoutput stage. The AD825 can swing a maxi-mum output of ±12.9V DC with the ±15VDC supply. The worst-case chip dissipationdue to the load current occurs at an outputvoltage of about 6.75V RMS, or 83mW. Forthis example, the worst-case chip dissipationis the sum of these three components, or338mW.

According to Figure 2 of the data sheet,the AD825 8-lead plastic SOIC package candissipate 730mW at 25°C ambient and ajunction temperature Tjmax of 150°C. To beconservative, derate Tjmax to 125°C. Thisputs the useful dissipation at room ambientat about 600mW.

Using the package θJA of 160°C/W, the

temperature rise for 338mW will be 54°C,for a case temperature of 79°C at room am-bient (25°C). While 79°C is not an excessiveamount of power dissipation, my rule ofthumb is that any semiconductor with a casetemperature that I cannot comfortably touch(50°C) at room ambient is running too hotand requires a heatsink. Removing the 2mApulldown bias won’t help very much—thecase temp will drop only 5°C.

So, whether or not you need a heatsink inyour application will depend more on theload current than the 30mW added due topulldown bias. And remember, the majorityof that temperature rise will be due to thesupply current alone.

The AD827 dual op amp is available fromNewark Electronics (www.newark.com or 1-800-4NEWARK) in four different packagetypes. Other duals worthy of considerationfor high-quality audio applications includethe Analog Devices AD826 and the Burr-Brown OPA2134 and OPA2604.

I have no personal experience with theAD797. It has a bipolar transistor inputstage, so at first glance it looks like it maywork in a phono preamp. Readers who havetried the AD797 in this application are en-couraged to share their experiences

LINE ARRAYSAs a long-time subscriber to SpeakerBuilder, I am interested in building

speakers that are line source instead ofpoint source. I would like to knowwhere I might find information on de-sign. I don’t recall seeing anything inyour magazine on this configuration. Isthere a reason for this? I would appreci-ate any help you can give.

Keith Hamilton [email protected]

Bill Waslo responds:

I am assuming that your question is aboutline source (or quasi line source) bass arraysrather than midrange or tweeter drivers. Butjust in case, I’ll first mention that there havebeen a number of articles about using linesource ribbon drivers for the higher frequen-cies, including one in the Jan. ’01 issue ofaudioXpress. You can find references tomore articles on this driver type by going toaudioXpress’s website, www.audioXpress.com, and doing a search on the word “rib-bon.” One notable article from Speaker


aauuddiiooXXpprreessss, PO Box 876 Dept. X1,Peterborough, NH 03458-0876 USA,603-924-9464, Fax: 603-924-9467,

E-mail: [email protected]

Comprehensive inscope, aauuddiiooXXpprreessss

brings your audio system needs together

in one monthly publication. You’ll find more articles, projects, reviews

and iformation than ever before.

audioXpresscontinuing

Audio Electronics,Glass Audio, andSpeaker Builder

PREMIERE ISSUE

January 2001

US $7.00 Canada $10.00

www.audioXpress.com

IT JUST SOUNDS BETTER

A 30W Monoblock

Amplifier

Do Harmonic Distortion

Measurements Mean

ANYTHING?

Constructing A

Super Ribbon

Speaker

Continuing Audio Electronics, Glass Audio & Speaker Builder

Tilting SystemResponse

With Tubes

AN OVERFLOWING

EQUIPMENTREVIEWS SECTION

FitzmauriceBuilds A

Double Reverse

Horn

Class DAmps For

Subwoofers

Build This Astronaut’s Elegant Preamp Design

$2995CANADA ADD $12 PER YR.,

OVERSEAS RATE $59.95 PER YR.

ONLY

for 12 monthly issues!

For Fast Service: 1-888-924-9465

Order On-Line: www.au-dioXpress.comit’s a

ll he

re, i

n on

e mag

azin

e!

Builder 2/92 described a 20′ ribbon tweetermade by SpeakerLab.

I’ve not seen as much published on thesubject of line array woofer systems. There isthe July ’63 AES Journal article, “ConstantDirectional Characteristics from a LineSource Array,” by Klepper and Steele (re-published in the AES’s “Loudspeaker An-thology” Vol. 1). Phillip Witham’s LinearArray articles (starting in Speaker Builder5/94) describe some of the directional ef-fects of driver arrays as related to his sound-field replication system.

I did some work with woofer line arraysduring my focused array loudspeaker project(starting Speaker Builder 5/98), though onlyas a complement to the high-frequency sec-tion, which was the subject of the articles. Ican say that, sonically, these stacks of port-ed boxes gave the best bass I’ve ever heardin any of my systems: very powerful and verysmooth, with the major disadvantage of total-ly dominating the room with their massive-ness. I’ve had a more reasonably sizedwoofer array cabinet design on paper forsome years now, waiting for me to find theconstruction time.

Line woofer arrays are a good approach to-ward generating clean, high-level bass, whileproviding some relief from room boundary ef-fects. If you take on such a project and areinterested in documenting the process andits results, I would certainly look forward toreading about it.

JOB SEARCHI read your article in the March 2001audioXpress (“Confessions of a For-

mer Driver Designer,” p. 70) and foundit quite intriguing. I am a sophomore atthe University of Maine studying elec-trical engineering. I, too, have beenbuilding stereo speakers from the ageof twelve.

My father got into it and I kind of fol-lowed in his footsteps, reading SpeakerBuilder magazine and others. I lovemusic (I am a guitarist myself) and havealways thought the perfect job would beto build and design speakers. I never re-ally considered actual driver design asan option. Where would an EE find hisplace in stereo speaker design?

So did you really enjoy the job, find itchallenging, and so on? I am workingfor Massachusetts Electric this sum-mer, just for a co-op. I would really liketo get a toehold into the audio industry

for next summer’s co-op. If you couldgive me any tips or pointers it would begreatly appreciated.

Nathan SherwoodUniversity of Maine

Perry Sink responds:

Well, try places like Harman, Onkyo, largerconsumer electronics or professional compa-nies, and so on. Yes, these are real jobs andreal careers. They tend to pay slightly lessthan other engineering disciplines, but youmay find the work more enjoyable.

I liked the challenge of making somethingso inexpensive sound respectably good. AndI worked on a few “skunkworks” projectsthat were a lot of fun, too. Start anywhereyou can get a toehold and just begin talkingto people that are inside the industry. Likeany specialty, it’s a fairly small industry andeveryone knows each other.

Once you know a few key people, it won’tbe hard to find something. First, look at thejob advertisements in the AES Journal orVoice Coil. Be a little ballsy and get in touchwith the engineering people (bypass the HRdepartments) and ask them whether they’relooking for an engineering co-op. In manycases they could really use some extra help.OK, you might not get a prestigious projectat first—you may sit around running re-sponse curves half the week—but you’ll havea good start.

I worked for Oxford Speaker Companyin Chicago around the same time thatyou were at Jensen (yes, the competi-tion). I was on the Ford account, andreading your article just gave me a huge “Matrix”-type flashback ofmemories.

Working at Oxford was one of thehighlights of my life. I couldn’t agreewith you more that you learn so muchabout engineering when you have nobudget to get high performance from a“mere paper” cone. For example, itmust withstand a drive from Alaska toDeath Valley in 15 minutes, 50 times aday, and withstand side impacts, explo-sions, high-pressure hoses, and—oh, bythe way, if it is free it is too expensive,so make it cheaper!

I left the high-pressure, fast pace ofOxford after three years and venturedto the luxury and Club Med feeling of


MAHOGANY SOUNDThe Transmission Line Specialist

P.O. Box 9044Mobile, AL 36691-0044

334-633-2054

Acousta-StufThe Very Best Damping Material For

Speaker Systems. It Produces DeeperBass, Cleaner Mids, And Greater

Dynamic Range. Acousta-Stuf Costs$9.50 Per Pound UPS Paid.

Q&ETLDQuick & Easy Transmission Line

Speaker Design Booklet & LOTUS1-2-3 Software. Learn How To

Design Optimizes 1/4 WavelengthTL Speakers. Q&ETLD Costs

$8.95 Plus $2.05 P&H.

Acousta-TubesRound Paper Tubes For BuildingCylindrical Speaker Enclosures.

Please Note Our New Address& Area Code

Call Or Write For A Free Catalog

northern California and Lucasfilm. Ac-cording to my Oxford co-workers, I had“made it out,” instead of the usual in-cestuous “let’s get a job at Onkyo, Pana-sonic, Foster, Harman.”

People here in the high-end markethave absolutely no concept of OEM,and I think that they actually disbelieveme when I tell them the stories that youhave actually written about.

Warren MansfieldProduct Development Engineer Lucasfilm, Ltd.

UNDERSTANDING VASWas the article by Ray Alden on theBella Voce (SB 4/00) correct in stat-

ing that the VAS (liters) is doubled? Areall my books wrong when they say thatthe VAS is halved?

Chris [email protected]

Ray Alden responds:

In response to your question regardingwhether VAS should be doubled or halvedwith regard to the Bella Voce speaker sys-tem, I am sure that doubling VAS is correctfor my situation. Naturally, this seems toraise a red flag if you are correct that allyour books say it should be halved. Howev-er, based on where in the book you are get-ting this assertion, there might be a simplesolution.

On page 30, section 1.91, of the Loud-speaker Design Cookbook (Vance Dickason,version 4), it states that for a standard con-figuration (two identical woofers in the sameenclosure), “VAS (and the associated box vol-ume VB) will be twice that of a single driver.”On page 32 in section 1.93, he states thatfor a compound configuration, “VAS (and theassociated box volume VB) will be half thatof a single driver.” Since the Bella Voce usestwo woofers in standard configuration, VAS isdoubled for use in all calculations. Possiblyyou are looking in your books at the com-pound configuration?

To get a better intuitive idea as to whyVAS behaves differently, I suggest you readpage 16 of Testing Loudspeakers, by JosephD’Appolito. His section 2.5.2 on “airsprings, mechanical compliance and VAS”should help you visualize these two situa-tions. Also, pages 17, 27, and 61 of my

book, Advanced Speaker Systems, will helpyou imagine VAS in a more intuitive mannerand describe why there is a different behav-ior when using the standard as opposed tocompound configuration.

NO FOOLIN’I’ve enjoyed reading some of yourpublications over the past few years

and have found little to quibble aboutin the articles. They have been sober-minded even when they were obviouscommercial messages. However, I musttake exception to the article entitled “A50W/Channel Composite Amplifier”(April ’01 aX, p. 6) and another one inthis same issue.

If this was an attempt at April 1stfoolishness, it must be rated as quitesuccessful! However, it is usually con-sidered good practice to inform thereaders of this at the end of the article.

An article addressing “Bass-BoostingNetwork” (p. 24) appears to be a gooddiscussion of crossover networks, com-plete with test data. The author is cred-ited with a patent application for cir-cuitry to improve incandescent lightingefficiency. Don Lancaster has a reviewof one such approach on his website,www.tinaja.com. The technique in-volves operating 30V bulbs from stan-dard 120V line voltage, using a dimmerfunction to reduce the effective voltage.He indicates that the measurementtechniques could not produce accurateresults and generally debunks the idea.

If this is the same person or tech-nique, then this article is also suspect.Not a nice thing to provide to your lessexperienced readers!

The referenced article has all themakings of a commercial announce-ment—verbiage that applies new wordsto standard techniques, enough techni-cal discussion to lend an air of validityto the presentation, and so on. The au-thor’s use of terms related to feedbackand servo techniques certainly indi-cates his familiarity and expertise inthat field. This, however, is not an indi-cation of the efficacy of this design.

The article text seems to be a mix ofassembly manual, advertising material,and theoretical tutorial. Well mixed up!Use of the term “composite” seems tobe overplayed, since the term is sobroad. Almost all amplifiers consist of


Sound Clearing House

130-page Online CatalogOther High-Quality Kits Available

www.speakerpage.com1-888-3-SPEAKER

Kit K05C22• $110 basic kit• Vifa parts• Assembled 8-element

crossover

CenterChannel

Electronic CrossoversTube

Solid StatePowered

Passive Preamps46 Step Attenuator

Free Catalog:Marchand Electronics Inc.

PO Box 473Webster, NY 14580

Phone (716) 872 0980FAX (716) 872 1960

[email protected]

Power Amps100% MOSFET

All available as kit also

XM126 tube xover, available 2,3 and 4 way

more than one gain stage, including theop amps used here. To consider apower amplifier as an op amp seems tostretch terminology, too. The first para-graph under “Other Features” seems toequate “compound” with composite,without any introduction. This couldwell be confusing to a novice.

The schematics on page 6, Fig. 1–3,have severe errors—there is no Vin terminal provided for two of them, andthe third has an unlabeled open resis-tor connection that might be the in-tended input. Again, very confusing to beginners.

In the main circuit diagram, a curi-ous mix of European and IEEE partvalue labeling is used, sometimes forthe same component type! And the ad-ditional definition of mF* as millifaradsin very tiny print is difficult to under-stand.* Consistency would be benefi-cial, particularly to newbies.

There are two jumpers whose func-tions don’t seem to be explained in thetext. It makes reference to “PC100”without any description thereof. Someterms in the text appear to be of Euro-pean origin; e.g., “stuffing guide,”which is more commonly referred to asthe component placement diagram.

And, no doubt, Mr. Zobel and theIEEE would certainly appreciate capi-talization of the network that is namedafter him. By the way, it would be nice ifsuch complex schematics were allocat-ed more space.

In paragraph two of this section, thestatement is made to the effect that theslew rates of the amplifiers multiplyeach other. If I understand this com-ment, it is saying that frequencies thatthe first stage couldn’t amplify will beamplified in a faster stage followingthe first. Or, to put it differently, if thefirst stage has a slew rate of 10V/mi-croseconds, a signal of 20V/m-secondwould not be passed to the secondstage, and its 100V/microseconds slewrate would allow the signal to be am-plified! NOT! Slew rate and frequencyresponse are effectively inversely pro-portional to each other, if I remembercorrectly.

In paragraph 3 of that section, thisstatement is made: “Two amplifiers in aloop virtually guarantee oscillation...” Ifmemory serves me, that is only true if

there is positive feedback. Negativefeedback negates this.

Despite the avowed intention of zerodistortion, the design seems to violateat least some of the more commonlyheld beliefs in the field. It said to be op-erating in Class B mode, which is a no-no to most purists. As is negative feed-back. The statement is made that suchfeedback corrects distortion. If memoryserves me, there is little if any connec-tion between negative feedback andlow distortion, other than perhapskeeping the active devices in their nom-inal operating regions. It won’t elimi-nate crossover distortion in a push-pullstage, for instance.

There is a reference to mounting fuseholders to a metal surface with siliconesealant. The insulating properties of thesealant are important here, becausemany fuse holders have exposed con-nections on the underside. Note thatRTV sealants contain acetic acid, whichis why they smell like vinegar, and arenot, therefore, a viable choice here. Theacid will eat away at the metal partsand cause eventual failure.

All in all, these discrepancies couldwell lead to many problems in imple-menting or understanding this design. Iam disappointed that it appeared inyour otherwise laudatory publication.

Phillip MilksIndianapolis, Ind.

*Reader Milks’ comment is typical ofthe very provincial attitudes of US engi-neers as to nomenclature for capaci-tors. While millifarads are not commonfor capacitor values they should be. Wehave no problems with such referencesin chokes. And nano is a standard in allmathematical values. Americans arejust too backward (or worse) to use theterm. The nanofarad would effectivelybanish all those 0.001µFs with 1nF and0.1µF with 100nF—E.T.D.

Kenneth P. Miller responds:

First, given the overall tone of your letter, Iprobably will not find the right words tochange your opinion of my amplifier. Thus Iwill focus on other objectives.


For one thing, your letter does offer somelegitimate criticism. This needs to be ad-dressed and I will do that. Some of yourcomments are based on a faulty understand-ing of elementary electronic terms, defini-tions, and concepts. Here, too, I can help.Finally, some of your judgments are basedon misconceptions picked up during yourtravels through the audio field. I’m not at allsure I can fix those problems, but I’ll give itmy best shot.

You have one particular misconceptionthat will be very, very difficult, indeed, todismiss. Most of this difficulty is due to thefact that you are not traveling alone. I willsave the tough task for last. I would prefer toaddress each issue in the same order inwhich it was raised in your letter, but I canfind no way to do this coherently, so I will, inmy own way, eventually answer most of yourcomments.

Despite heroic efforts from the publisher,the author, and the printer, there are mis-takes in my article. We both know that. Yourassertion that Zobel is a proper noun is cor-rect. I did not catch this mistake when Iproofread the copy, nor did the editor. Evenworse, I misspelled the word elsewhere in mypresentation.

The unlabeled open resistor (Fig. 2, p. 6)should go to ground. I think this a printingerror as it did not appear on my proofreadcopy. In Figs. 1, 2, and 3, the input signal isapplied to the + terminal of A1. In Fig. 5,Vin is applied at the + terminal of the BUFFop amp.

I agree that the mix of labels is confusing.I submit all capacitor values in either pico-farads or microfarads. The editor prefersother units.

There are mistakes in the amp schematic,Fig. 6, page 7. The collector of Q2 should beconnected to the base of Q4. And the inputside of C3 should come from the junction ofR2 and C2 rather than from the output termi-nal of A1. The jumpers are also missing onthis schematic. These are nothing more thanpieces of wire used to connect eyelet 2 to eye-let 3, and 4 to 5. The link between 2 and 3connects the output of A2 to the input of A3,and the other link connects the base of Q2 toground. This is the normal condition of thelinks during composite operation. During theconstruction process or for troubleshootingpurposes, the links can be removed and repo-sitioned such that A2 and A3 can be operatedas conventional, independent amplifiers. Thisdetail is of most interest to someone building

the amp and covered thoroughly in my sepa-rate stuffing guide.

I have no plans to commercially exploit myamplifier. I sell my stuffing guide and boardsat cost to those who would like to try my amp.Although it is not illegal to turn my projectinto a cash cow, it is probably a violation ofboth the spirit and intent of the magazine.The real motivation and satisfaction is in shar-ing a better idea with other audiophiles.

The fuse holder I specify has no exposedconnections on the underside, and thus theinsulating properties of the sealant are irrele-vant. Since I’m not a chemist, I can’t refuteyour claim that the sealant will eventually eataway at the metal. But I did check my ownunit, which has been in continuous service forthe past 16 months and found not a trace ofdeterioration. I will contact the manufacturerand if he confirms your claim, I will advisebuilders to use a different mounting system.

The term “composite” is a definition. It isonly broad enough to include all topologiesin the family.

Two op amps inside a purely resistive feed-back will definitely oscillate. The high fre-quencies will be shifted 180° relative to thelower ones, exactly the right condition for pos-itive feedback. The output will oscillate vigor-ously, usually at the lowest frequency capableof producing the full 180° phase shift.

I can understand your concern over Class Boperation. Optimum Class B is a term refer-ring to recent discoveries and is probably notcommon knowledge just yet. It is well knownthat push-pull devices deliver minimum dis-tortion when biased for full Class A operation.If, for the sake of efficiency, the bias is re-duced, pushing the output into Class B or ABoperation, then distortion will increase. Asbias is increased toward Class A operation,you can expect the distortion to reduce.

This is all true with one exception: thebipolar transistor. I use a bipolar output andrecommend a bias current of 11.5mA. Atthis current, my amp delivers minimum dis-tortion out, except for full Class A operation.If the bias setting is changed from 11.5mAto 5mA, for example, distortion will increasedue to crossover artifacts in the output. Ifthe bias is increased to, say, 20mA, distor-tion will increase once again, but not due tocrossover artifacts, but rather to a phenome-non known as transconductance doubling.The distortion will stay at this higher leveluntil it is adjusted to near Class A operationwhere it drops to the lowest level of all.

If you get the drift of what I’m saying, it


CleaningKits from OLD COLONY

SOUND LABORATORY

DeoxIt/PreservIt Electronic Maintenance and Contact Con-ditioner SystemCAIG Laboratories, Inc.Reduces noise and distortion in your systemcaused by poor electrical contacts in thenumerous connectors. Although DeoxIT isnonconductive, it enhances the electricalcontact, thus improving the energy transferthrough the connector. If the film’s integrityis broken, it “heals” itself, automatically re-distributing the coating over the entire sur-face to reseal the breaks. DeoxIT is formu-lated specifically to eliminate contaminationwherever contacts meet without solder-PCboards, connectors, sockets (includinggold), in audio and electronic equipment,communications gear, and computers. Shipping wt: 1 lb. KM-6 ..................$29.95

ProGold Conditioner, Preservative and Lubricantfor Plated Metal Electrical ConnectionsCAIG Laboratories, Inc.Specifically formulated to improve conduc-tivity, lubricate and protect gold, base met-als and other precious metal surfaces. Dueto its unique properties, it deoxidizes andcleans surface contamination, penetratesplated surfaces and molecularly bonds tobase metals. ProGold increases the perfor-mance and reliability of all electrical equip-ment, improving conductivity for greater de-finition and dynamics, reducing noise andinterference and lowering distortion. Shipping wt: 1 lb. KM-6PGL7 ........$19.95

Williamson LP Record Care KitReg Williamson Knock out static charges and deep clean yourrecords with this classic two-step method stillin demand by high-end audiophiles. The de-static compound supplies molecules whichattract and retain static charges—one treat-ment usually lasts for the life of the record.The deep cleaner dissolves and flushes thedirt out of the grooves—and then holds it in afilm which you simply peel away.Shipping wt: 1 lb. KM-9 ..................$9.95Shipping wt: 1 lb. KM-9R ................$4.95

3-oz. refill of cleanerShipping wt: 1 lb. KM-9S ................$4.95½-oz. refill of liquid de-static treatment

For Fast Service Call 1-888-924-9465

Order On-Line Atwww.audioXpress.com

Old Colony Sound Laboratory, PO Box 876,Dept. X1, Peterborough, NH 03458-0876 USA

Phone: 603-924-9464 Fax: 603-924-9467E-mail: [email protected]

should be obvious that, with bipolar transis-tors, there are only two viable bias options—optimum Class B or full Class A. Why notuse a different output device, such as apower MOSFET, since it is immune totransconductance doubling? Because anybipolar pair operating optimum Class B willproduce lower distortion than any pair ofMOSFETs biased at Class AB. The same istrue at Class A. There can be legitimate rea-sons for using MOSFETs in the output stage,but there is always a distortion penalty.

Your statement that there is no connectionbetween negative feedback and low levels ofdistortion is simply untrue. I suspect that yourclaim is based on some “zero-feedback” am-plifier having respectable distortion figures.

The catch is this. The manufacturer heremeans zero global feedback. His amp useslocal negative feedback and gobs of it. Thereis no way to have respectable distortion fig-ures without the heavy application of nega-tive feedback. Negative feedback is negativefeedback, whether local or global.

There is a misconception about feedback,which goes something like this: Negativefeedback in an amplifier is generally a badthing. The less feedback used, the better—zero feedback is best. This all started someyears ago and, of course, has been heard bydesigners and others. Their reaction to it?Most audio designers do not buy it, andacademia totally rejects it. That’s why uni-versities are still teaching conventional nega-tive feedback theory.

In a nutshell, it goes something like this.Negative feedback is a powerful tool, so use it.The one and only penalty for too much nega-tive feedback is instability. Instability is hardlya subtle defect and easy to cure—back off thefeedback level a bit until stability is sufficient.

Let me say at once that I do not questionthe sincerity of those who believe that nega-tive feedback is a bad thing—they are quitesincere about their belief. But the truth or re-ality of any given proposition cannot be deter-mined from the sincerity of convictions. Mostmistakes are honest. The preponderance ofevidence is what matters and it supports theuse of negative feedback. As far as proof goes,it can’t be had—but only because the puristscling to a disprovable hypothesis.

TONE CONTROLThanks for publishing the projecton “An Unusual Tone Control”

(April 2001, p. 50). For years I’ve beenlooking for a loudness control schemat-

ic that did not involve exotic multi-poleswitches or tapped potentiometers.There is, however, a detail missing fromthe parts list and text. What type of po-tentiometers are P1, P2, P3, and P4?Logarithmic or linear?

Simon BaldersonS. [email protected]

Tech staff here thinks they are all log taper.Some have suggested that the unit wouldbenefit from an added cathode follower out-put stage.—E.T.D.

BEST EDITIONI enjoyed Larry Lisle’s “Books forthe Tube Audio Beginner” (April ’01,

p. 76). I have most of the books he lists,including my introduction to tube am-plifiers (the two-stage 6BF6 job!), circa1965, at the tender age of 8 years: AlfredMorgan’s The Boys’ First Book of Radioand Electronics. I would like to know if,possibly, he can recommend one partic-ular year in the 1950s that the ARRLHandbook may have stood out as the“best” edition for tube audiophiles. Withten editions (1950-1959) at about $30 apop (from one used book source), itmight get expensive for me, as well asredundant, to build this segment of mylibrary!

John [email protected]

Larry Lisle responds:

Thank you for your interest. The ARRL Hand-book changed very little from year to year.Editions from the late ’40s and early ’50shad a greater variety of building techniques,using wooden racks as well as all-metal con-struction. Later editions will have more re-cent tube data. These are about the only dif-ferences for our purposes, and I’m sure youwill find any Handbook useful.

VINTAGE AMPSGoing through the pages of a recentedition of audioXpress (March 2001),

I became very interested in the articleabout remaking the Dynaco Mark III byKara Chaffee (p. 18). The reason I am sointerested is that I also own a vintageamp, which is a Harman Kardon Cita-tion V in original working condition. I


HAGERMAN T E C H N O L O G Y

Vacuum Tube Curve Tracer

See what you’ve been missing. Invest in a VacuTrace . Assembled: $999, Kit: $899.

www.hagtech.com Hagerman Technology LLC PO Box 26437 Honolulu, HI 96825 USA Phone: (808)-383-2704 Fax: (808)-394-6076

would like to remake it like Kara Chaf-fee did with the Dynaco Mark III.

Would it be possible to see this proj-ect in the pages of audioXpress soon? Iwould like to know whether this ampli-fier has ever been rebuilt to today’sstandards, or if anybody out there is in-terested in this project.

Fernand BinetteVictoriaville, Quebec

SHED SOME LIGHTAs a reader of your magazine andtube audio enthusiast, I was pleased

to see one of JKL’s products mentionedin “Switching Power Supplies for TubePreamps” by Eric Barbour (April 2001,p. 16). I’d like to add a few notes on theJKL 12200 lamp inverter. The 12200 isthe hobbyist version of the JKL BXA-12529, which is designed with pins tomount onto a circuit board. The BXA-12529 and other JKL inverters, includ-ing EL inverters, are available throughDigi-Key (WWW.DIGIKEY.COM). Asseen from the accompanying schematic(Fig. 2), the DC-AC inverter is a Royer-type oscillator circuit; the frequency ofthe oscillator is about 35kHz.

A word of caution: The open-circuitoutput voltage of the inverter can ex-ceed 1000V RMS, as it is primarily in-tended for ignition and operation of dis-charge lamps. Capacitor C1 is used togive the output waveform a more sinu-soidal shape; capacitors C2 and C3 splitthe output from the transformer for twolamps and provide a means of ballast forthe lamps by limiting lamp current toabout 5mA RMS each. When using theBXA-12529 (12200) as a high-voltage DC-DC converter, you should use diodesand capacitors that can handle the ag-gressively high voltage and frequency.

I look forward to more insightful arti-cles in your publication.

John KahlJKL Components Corp.

SOUND RECOMMENDATIONSIn his review of Cool Edit Softwarein the April issue of audioXpress

Perry Sink mentions that you must becareful what sound card you use forrecording, and that some cards intro-duce significant noise pickup or havemarginal signal-to-noise ratios. From hisexperience, can he tell us any soundcards that should be avoided, and alsowhich sound cards he recommends?

Richard MainsYpsilanti, Mich.

Perry Sink responds:

I have done no personal evaluation of variouscards myself, other than to note that the typ-ical card that comes in a PC is most likelybuilt to minimal standards and is noise-prone. If you get excellent sound from such acard, you’re a bit on the lucky side.

However, I know that many high-perfor-mance cards are available at price points upto $1000+, often designed with serious 24-bit home recording in mind. In the middleare a number of excellent products for under$100.

The following are some resources on theweb that may take you a step further:

http://www.epanorama.net/pc/sound.html http://www.music-software-reviews.com/sound_cards_PC.html

Also, check sources such as ZDnet, PCMagazine, and various musician magazines.Note: Turtle Beach cards have been very


FIGURE 2: JKL’s BXA-12529 inverter.

ADVERTISE IN THE NEW

audioXpressmagazine

It’s a cost-effective way to reach 15,000 audio enthusiasts

with your products!

ONE MONTHLY PUBLICATION will

keep your message in front of active audio consumers!

Continuing Audio Electron-ics, Glass Audio and Speak-

er Builder

Contact Louise Sullivan at (603) 899-5068 or e-mail at [email protected].

Don’t Be Left Out!

Call Today For InformationOn The Next

Available Issue!!

YOUR ONE-STOP OPPORTUNITY TO REACH

OUR READERS!

positively reviewed in PC Magazine by CraigL. Stark.—E.T.D.

I SEE ICSRegarding Charles Hansen’s column“Chips Gone from the Block” (April

2001, p. 89), I’m sure you’ll get a ton of e-mail pointing out that datasheets forboth the BA1404 and the TC9147 arefreely available for download fromhttp://www.freetradezone.com/ (Actual-ly, they had three different versions ofthe Rohm datasheet, and just one ver-sion of the Toshiba datasheet.)

Scott NewellFort Smith, Ark.

Charles Hansen responds:

Thanks for the information. This site is cer-tainly worth a Favorites/Bookmark!

HELP WANTEDDoes anyone have the instructions anddrawings for the Le Petit Onken? Also,do you have enclosure instructions anddrawings for any Norelco or Philips full-

range 12″ drivers? These enclosureswere made back in the late 50s. I havesome vintage 12″ drivers and wish tobuild the Le Petit Onken for them, and Iwas hoping that some reader wouldhave the solution to my problem.

Evangelos [email protected]

I’m trying to find out the value of aJensen model 610, 15″ triaxial withbridge tweeter circa 1953. I’ve been toldthat it is the holy grail of vintage speak-ers. Can anyone enlighten me?

Paul [email protected]

Many years ago (approximately 20), Isubscribed to the Audio Amateur maga-zine (and Speaker Builder) when I was aProfessor of Chemistry at Syracuse Uni-versity. A few years ago I retired fromteaching and settled in Keene, N.H.

Searching the web, I noticed that thesuccessor to these magazines is head-quartered in Peterborough, N.H. I was

disappointed that there were no audioamateur clubs listed on your website(not even in Peterborough!) except forone in Connecticut (no address given).Are there no groups of people in theKeene-Peterborough area who are inter-ested in getting together to listen togood, well-reproduced music and talkabout audio (the old Hi-Fi)?

Clarence [email protected], N.H.

I have a pair of Clements 206di speak-ers, which I enjoy, but was wondering ifsome modifications or upgrades wouldbe possible. To begin with, the bindingposts could definitely be upgraded, andI thought maybe a similar crossoverwith higher-performance parts could re-place the existing one. I am not an ex-pert in speaker building, but I have hadmy DAC modified by an expert namedDan Wright in Portland, with very goodresults. Could anyone provide informa-tion and/or service on modifying or up-grading speakers?

Jason [email protected]

I’m looking for information and noticesabout the triphonic JR loudspeakers(model lpa). I think it’s 15/20 years old.

J.J. [email protected]

I am building a box for a 1998 DodgeRam 1500 quad cab for 2 rockfordpunch 10″. Is there box designer forthis?

Robert Dicks [email protected]

I have an Altec 5″ tweeter that is notworking, and I wish to find out how torebuild it.

Greg [email protected]

Readers with information on these top-ics are encouraged to respond directlyto the letter writers at the addressesprovided. −Eds.


Classic CircuitryPilot 264

This is a seemingly generic-lookingbasic stereo power amp from 1962, butlook at the output section. The 4Ω tap isgrounded! Audio Research did not orig-inate that concept. Further, feedbackcomes both from the ungrounded “com-

mon” tap and from the 16Ω tap. Mostunusual.

Lance CochraneSan Francisco, Calif.

Pilot 264.


the Circlotron. It is also the key featureof the 6AS7/6080 I described in the ear-lier GA article.

Now, if someone would contributethe transformers, we would have a project.

Obviously bitten by the Circlotronbug, John Stewart, in a subsequentissue, provides further details about this

exotic amp...and enough information foryou to build your own.−Eds.Circlotron Amp

from page 29

Ask AX

ENCLOSURE VOLUMEIn figuring box volume, a higher VASspeaker would require a larger boxthan a speaker with a lower VAS. By def-inition, VAS is the volume of air hav-ing the same acoustic compliance asthe driver’s suspension. To me thatsounds as though a low VAS driverwould have a stiff suspension and ahigh VAS driver would have a loosesuspension. But doesn’t a loose speak-er require a small box and a tightspeaker need a large box?

Wilbur [email protected]

Dick Pierce responds:

First, you are correct about the meaning ofVAS: it is the volume of air having the sameeffective acoustic compliance as the driver,and there is a direct relation between themechanical compliance CMS of the driverand its VAS. The following equation showsthat relationship:

where ρO is the density of air (about1.18kg/m3), c is the speed of sound(342m/s), CMS is the mechanical complianceof the driver suspension (in m/N), and SD isthe effective emissive area of the diaphragm.This equation confirms your notion that ahigh VAS implies a high CMS and vice versa.And the correlation is direct, assuming thecone diameters are the same.

However, your guess—as intuitive as itmight seem—that a driver with a large VASneeds a small enclosure and vice versa is notcorrect. One of the problems with makingsuch a sweeping generalization is that it ig-nores a large number of important factors.

In a sealed box system, for example, thecombination of the driver’s VAS and the en-closure volume VB together determine thetotal mechanical stiffness of the system.That stiffness, combined with the effectivemoving mass of the cone, determines thesystem resonance. So, you could correctlyconclude from this that the proper enclosure

volume VB for a given driver is that needed toraise the system stiffness to achieve the tar-get system resonant frequency. This conclu-sion makes no suggestion that a large VAS re-quires a small VB, or vice versa.

There is another relationship that’s impor-tant here. Raising the resonant frequencyalso raises the system Q at resonance. The Qincreases directly as the system resonant fre-quency increases.

These two parameters—system resonanceand system Q—change together in a pre-dictable fashion. In a closed box system, forexample, there is a parameter, cx, which isthe so-called “tuning ratio.” It is the ratiobetween the driver equivalent volume VASand the enclosure volume VB:

The change in resonance and the changein system Q can be calculated once youknow the tuning ratio. For example, you cancalculate the resonant frequency of a driverin the enclosure FC from the tuning ratio andfree-air resonance of the driver FS:

Similarly, you can calculate the system’selectrical Q, QEC, from the driver’s free-airelectrical Q, QES, as follows:

From this, I hope it’s clear that it’s theratio of the driver compliance to box com-pliance that’s important, regardless ofwhether you have a high-compliance or low-compliance driver. For example, if you hadtwo drivers, both with the same free-air res-onant frequency and the same free-air Q,but with different equivalent compliancevolumes, to achieve the same target reso-nant frequency and system Q, the ratio ofthe VAS to VB for both speakers would needto be the same. The result is that thespeaker with the larger VAS would require alarge enclosure volume VB.

Now, seldom do you encounter such a sit-uation in actual practice (though there is noinherent physical reason why it couldn’t hap-

pen), but the example serves to illustrate thepoint that the necessary enclosure volume isnot inversely related to VAS. Rather, the idealcabinet volume is determined by the desiredsystem response and the total of the driver’selectromechanical parameters. There is nohard-and-fast rule that suggests that high-compliance speakers require small boxesand vice versa.

Part of your confusion may derive fromthe original claims made by Acoustic Re-search back in the early 60s. They used avery high-compliance driver in a very smallenclosure, at a time when people generallyused low-compliance drivers in large en-closures. The latter is easy to explain; peo-ple simply did not understand how loud-speakers and enclosures worked as a sys-tem. The breakthrough that Villchur andAllison achieved was to consider the driverand enclosure as a system, though itwould take Thiele and Small to generalizethe principle.

The AR idea was to marry a high-compli-ance, high-mass driver with a low-compli-ance enclosure. The primary reason fordoing this was that it was the enclosure thatlargely determined the total system stiff-ness. In AR and similar “acoustic suspen-sion” systems, the tuning ratio, cx, wasoften on the order of five and higher. Theconsequence of this is twofold.

First, it meant that the system could tol-erate large variations in the driver compli-ance, and those variations would have aminimal effect on the system overall. Forexample, if the enclosure compliance was¹⁄₅ that of the driver compliance (a tuningratio of 5), any variations in driver compli-ance on the total system compliance will bereduced by a factor of 6. A ±25% variationin driver compliance will result in only a±4% or so change in system compliance,and then only about a ±1% change in sys-tem resonant frequency.

Second, it is the enclosed air that pro-vides most of the restoring force for thecone, not the driver’s suspension. In thatway, the more linear operation of the en-trapped air overwhelmed the nonlinear me-chanical suspension. However, as a trade-off, the acoustic suspension system requires

1QQ ESEC +α=

1FF SC +α=

B

AS

V

V=α

2DMS

20AS SCcV ρ=


a driver with a high moving mass to achievea low system resonance, and that combinedwith the long-overhung voice coil resulted ina system with very low efficiency.

This serves further to illustrate the point Iam trying to make: the enclosure size is de-termined by the parameters of the driver(compliance, Q, resonant frequency, and soon) and the target system performance (sys-tem resonance, system Q). In some designs,that means a high-compliance driver in asmall box, in other systems, it may meansomething else.

ZOBEL POWER DISSIPATIONAs an avid subscriber to your journal, Iwas wondering whether I could botheryou for some technical advice concern-ing Zobel networks in two-waycrossovers. It seems the problem is inthe power-handling characteristics ofthe board we’ve designed. Because theZobel network equalizes the impedanceof the woofer in the upper part of thewoofer’s range, it necessarily must dis-sipate some power. My calculationsshow that, over a narrowband of fre-quencies, the power required is 40-50Wwhen an amplifier is putting 100W intothe speaker at 8Ω.

In order to claim the real power-han-dling characteristics of a speaker, is itnecessary to “beef up” the Zobel net-work to account for the power dissipa-tion through the network?

Bob McChesneyBobM@Port of Everett.com

G. R. Koonce responds:

The power requirement for resistors used indriver Zobels has always been a bit of a mys-tery. The “standard” Zobel consists of a se-ries-connected resistor/capacitor networkconnected across the driver terminals wherethe resistor value in ohms nearly matchesthe driver resistance. It is clear that whensuch a network is driven by a voltage source,the current through the network increases asfrequency rises. The maximum Zobel resistordissipation will thus tend to be at the topend of the crossover range for eachdriver/Zobel combination.

If the woofer crossover frequency is highenough—likely with a two-way system—theZobel resistor may dissipate more power thanthe woofer at the top of the band because ofthe woofer’s rising input impedance. That is

exactly what the Zobel is supposed to do!The problem is how to size the Zobel resistorfor your intended application.

Just like semiconductors, resistors needderating. For high ambient temperature ap-plications such as in military gear, a resistoris used only to half its nominal rated power.Typical commercial application would re-quire less derating. In a high-power speakersystem where the Zobel is mounted insidethe box with all the driver dissipation, youshould consider the possible rise in ambienttemperature within the box. The remainderof this letter refers to the derated value ofthe Zobel resistor.

Narrowband ApplicationsIn a case where the input signal may be a si-nusoid or other narrowband signal, you cancompute the actual worst-case Zobel resistordissipation, as when you plan to test with si-nusoids at high power or pro-sound applica-tions such as bass guitar or organ music.Here you can take the maximum input volt-age the system will see, and at the maximumfrequency the Zobel will compute the currentto the Zobel resistor and thus the dissipa-tion. The computation should include any“peaking” in the crossover network that cansubject the driver/Zobel combination to avoltage somewhat higher than the maximumsystem input voltage. In these cases you mayfind the Zobel resistor must be rated to thesame power rating as that of the driver.

Music ApplicationsWhen you play music, the Zobel resistor hasreduced power dissipation as it receives lesspower at the low end of the driver frequencyrange and music tends to spread the energyout over frequency. Many years ago, mycomputations of the power a Zobel resistorcould dissipate matched those results re-ported by Mr. McChesney; i.e., the Zobel re-sistor could see about one half of the powerdelivered to the driver/Zobel combination.Thus for a 100W woofer pushed to its limit,you might need a Zobel resistor rated at50W with music.

Practical Results with MusicI normally develop speaker systems with100W amplifiers. With the high peak to av-erage power requirements of CDs and LPs,this means about 10W-20W of averagepower available. The limiting used on FMwould allow a somewhat higher averagepower from a 100W amplifier, but 10-20W

produces loud music and I normally wouldnot exceed this level.

I started developing systems with 10W re-sistors in the Zobels, and after playing quitea while at 10-20W average, the resistorswere never even warm. Thus they clearlywere not dissipating anything like the 5W to10W the worst-case calculation would indi-cate. Why was this?

First, while the music may reach 20W av-erage for some period of time, it generallydoes not stay at this level. Thus some timeaveraging of the dissipation occurs. Powerresistors have the ability to take very highpower dissipations for short periods of time;it is the average dissipation over time thatsets the resistor heating.

Next, testing showed that with a three-waysystem with the lower crossover frequency inthe 800Hz range, about one half of the aver-age system input power goes to the wooferand the other half to the midrange andtweeter. The tweeter normally receives about10% to 15% of the power. For a two-waysystem with its relatively high crossover fre-quency, it’s best to assume the woofer seesall the system input power.

My current rule of thumb for home sys-tems is that I use at least a 10W resistor onthe woofer Zobel and at least a 5W resistoron the midrange Zobel. The lower power rat-ing on the midrange is based on the fact themidrange is normally more sensitive than thewoofer and thus has some attenuation aheadof the Zobel. I normally don’t worry about thetweeter Zobel resistor power at all; I use 5Wunits for convenience. In over 30 years Ihave never had a system returned with afailed Zobel resistor.

Why not just use high-power resistors,which are relatively cheap? The problem isfinding room for these physically large resis-tors, which will run cooler than smaller onesbut probably actually put out more heat.This is because the resistor types normallyused (wire wound) have a resistance thatrises as they heat up, making them some-what self protecting. This resistance changewill, however, slightly modify the Zobel fromits design values.

As noted, there are some cases that re-quire special attention to the dissipation inthe Zobel resistors as follows:

1. Very high average system input power.This would be in music or pro-soundapplication where the driver is pushed

(continued to page 95)


New Chips on the BlockNational Semiconductor LM 4651/4652 Class-D AmpBy Charles Hansen

The IC combination of the NationalSemiconductor Overture LM4651driver and the LM4652 MOSFET poweramplifier provides a high-efficiency,two-channel, Class-D subwoofer amplifi-er. Applications include powered sub-woofers for home theater and PCs, carbooster amplifiers, and self-poweredspeakers.

The LM4651 is a fully integrated con-ventional pulse-width modulator(PWM) driver IC. The IC contains shortcircuit, under voltage, over modulation,and thermal shutdown protection cir-cuitry. It contains a standby function,which shuts down the PWM and mini-mizes supply current.

The LM4652 is a fully integrated H-bridge power MOSFET IC in a TO-220power package. Together, these two ICsform a simple, compact, high-poweraudio amplifier solution complete withprotection normally seen only in ClassAB amplifiers. Fewer external compo-nents and minimal traces between theICs keep the PCB area small and aids

in EMI control. The near rail-to-railswitching amplifier substantially in-creases the efficiency and reduces theheatsink size compared to a Class ABIC of the same power level.

The LM4651N is housed in a 28-lead

DIP package. The LM4652 is availablein 15-lead TO-220, either isolated(LM4652TF) or non-isolated (LM4652T).

National Semiconductor, Santa Clara,CA, 1-800-272-9959, www.national.com.


FEATURESConventional pulse-width modulation.Externally controllable switching frequency range:75kHz to 200kHz.Integrated error amp and feedback amp.Turn-on soft start and undervoltage lockout.Over modulation protection (soft clipping).Short circuit current limiting and thermal shutdown protection.Self-checking protection diagnostic.

KEY SPECIFICATIONSUser supply voltage: ±22V Power supply range: ±10 to ±50V Maximum power at 4Ω, THD ≤1%: 135WMaximum power at 8Ω, THD ≤1%: 75W Maximum power at 4Ω, THD ≤10%: 170WMaximum power at 8Ω, THD ≤10%: 90WSignal-to-noise ratio: 80dB (typ) THD+N 0.3% (typ), 10W, 4Ω, 10−500HzMaximum efficiency at 125W: 85% (typ)Standby attenuation: >100dB (minimum)


New Chips on the BlockLinear SystemsBy Charles Hansen

Linear Systems is a full service US-based manufacturer of specialty linearsemiconductors, providing superior

performance for applications requiringlow offset, low leakage, low noise, ortight thermal tracking. The product lineincludes single and monolithic dualand matched NPN bipolar transistors,N-channel JFETs, N-channel MOSFETs,PNP bipolar transistors, P-channelJFETs, and P-channel MOSFETs. Theseproducts are specifically designed toservice the new design and second-source requirements of the audio, hy-brid, medical, military, space-cryogenic,test, and measurement markets. Elec-trical selections to meet specific appli-

cations or design requirements areavailable for all discretes.

Linear Systems also offers pin-for-pinreplacements for current and discontin-ued discretes made by Intersil, Motoro-la, National, and Siliconix. Productscome in the four standard packages,TO-71, TO-78, P-DIP, and SOIC. Die in-formation is also available.

Linear Systems, Inc., 4042 ClipperCourt, Fremont, CA 94538, (510) 490-9160, FAX: (510) 353-0261, www.linearsystems.com.

For More Audio Information,Check Us Out.

or E-mail us [email protected]

Audio Amateur, Inc.,PO Box 876, Peterborough,

NH 03458-0876 USAPhone: 603-924-9464 w Fax: 603-924-9467

Visit Our Website!

audioXpress.com

BOOKS, BACK ISSUES,

MAGAZINES,

SOFTWARE DEMOS,

BARGAINS &

CLOSEOUTS

It is not enough to put oil Forget about resistive devices It is not enough into a Capacitor to make for the volume setting. to use silver wire

it musical.... We offer a silver wired for a good SE-OPT...transformer approach

Silver Rock Silver RockTransformer- Output transformerPotentiometer

There is only one optimum solution for a given problem.

Audio Consulting / 14B chemin des Vignes / 1291 Commugny / SwitzerlandFax: 00-41-22-960-12-59 / e-mail: [email protected]

http://www.audio-consulting.ch

Along with negative feedback, theseemingly innocuous and trivial cou-pling capacitor is arguably one of thegreatest single factors you can easilycontrol in determining frequency re-sponse of an amplifier. Other factors,such as the Miller effect, the upper fre-quency response limit, and the outputtransformer response, are somewhatset and not easily altered.

In correspondence with Mr. Dell, re-cently, he questioned the size of thecoupling caps I was going to use in myEico HF-60 amplifier. This raised aquestion in me, how do you go about se-lecting the value of coupling capacitorsin an amplifier? So I hit the books tosee what the experts said, and thenstarted to experiment with my amplifier.

FORMULAIC SOLUTIONLet me first relate my understanding ofhow to derive the values for the cou-pling caps. The Eico HF-60 manualstates that the lower frequency limit is20Hz. The coupling cap of 25nF and theoutput tube grid resistor of 330kΩ forma simple RC high-pass filter. By plug-ging these two values into the formulafor an RC filter circuit, you can find the−3dB rolloff point of the high-pass filter:

freq = 1/(2πRC)

which equals 19.3Hz, very close to the20Hz lower limit.

Electronic Designer’s Handbook(Robert W. Landee, et al., McGraw-Hill,1957, p. 3–19) confirms this as long as Ris much greater than the parallel com-bination of the driving tube’s plate loadresistor and its dynamic plate resis-tance. For a more accurate value for C,add this parallel combination resis-tance to the grid resistance in the previ-ous formula. But that only slightly in-creases C by a few nanofarads, which isnot much difference.

F. Langford-Smith’s Radiotron De-signer’s Handbook (p. 483) also saysthe preceding formula is valid, but

adds, “In high fidelity amplifiers a fairlylarge value of C is generally adopted,thus not only improving the low fre-quency response but also reducingphase shift and possibly also improvingthe response to transients. However, ex-cessively large values of C are undesir-able.” This last sentence is particularlytrue because with negative feedback ex-cessively large coupling caps couldlead to distortion and oscillation. But Iwonder how large he meant by “a fairlylarge value of C”?

Radio Handbook (William I. Orr, Edi-tors and Engineers, 1970, p. 147) givesan empirical formula for coupling caps.“For high-fidelity work the product ofthe grid resistor in ohms times the cou-pling capacitor in microfarads shouldequal 25,000 (i.e.; 500,000 ohms ×0.05µfd = 25,000).” In my case CµFwould equal 25,000/330k, which is.076µF or 76nF. Plugging this value forC back into the formula for an RC filtercircuit (just to see where the new lowlimit would be) comes to 6Hz, which isquite a bit lower than the original.

REACTANCEAnother way to look at the capacitanceof the coupling caps is from the stand-point of capacitive reactance. The reac-tance of a 25nF cap at 20Hz is Xc =1/(2πfC), or 318kΩ. The reactance of a76nF at the same frequency is 105kΩ.Quite a large difference! From this view,using the larger cap would make moresense, but again, this is predicated onthe condition that it does not cause in-stability, oscillation, or distortion. Table1 shows the reactances of the two capsversus frequency from 1Hz to 20kHz.You can see how capacitive reactanceaffects the whole audio frequency band,not only the lower limit, but the upperend as well. The 78nF cap is about threetimes larger than the 25nF, and aboutthree times lower in reactance and inthe low-frequency limit, too.

I realize that these equations and for-mulas—empirical or otherwise—are just

intended as a starting point. You mustexperiment by increasing the capaci-tance (lower the low-frequency limit)until the amplifier shows excessive dis-tortion or oscillates (if it does), thenback it off for a safety margin. This isthe value to use. You do this, of course,using an oscilloscope connected to theproperly loaded output, with some kindof generator injecting an audio test sig-nal to the input.

In my particular case, after tryingvarious values, I ended up using 47nFcoupling caps. Using larger valuescaused high-frequency oscillations toride piggyback on my output, testsquare waves, but the 47nF behavedperfectly through the whole audiorange. The new low-frequency limit isnow 10Hz, and my amp sounds sweeter,with a more robust bass. So with a littleexperimentation I was able to success-fully double the size of the original cou-pling capacitors and improve the re-sponse, while introducing no degradingside effects.

Michael KornackerPalatine, IL


TABLE 1REACTANCES OF A 25NF

COMPARED TO A 76NF CAP VERSUS FREQUENCY ALONG

WITH THE 47NF.

FREQUENCY Xc FOR Xc FOR Xc FOR IN Hz 25nF (ΩΩ) 47nF (ΩΩ) 76nF (ΩΩ)1 6.4M 3.3M 2.1M10 636k 338k 210k50 127k 68k 42k100 64k 34k 21k200 32k 17k 10k400 16k 8.4k 5.2k600 11k 5.6k 3.5k800 8.0k 4.2k 2.6k1k 6.4k 3.4k 2.1k2k 3.2k 1.7k 1.0k4k 1.6k 850 5206k 1.1k 565 3508k 800 423 26010k 640 340 21012k 530 282 17514k 450 242 15016k 400 212 13018k 350 188 11620k 318 169 105

Glass ShardChoosing Coupling Capacitor Values


Classifieds

VENDORSAudioclassics.com buys, sells, trades, and repairs high end audio. 607-766-3501, 3501Old Vestal Rd., Vestal, NY 13850

Summer sale, all kits 10% off www.borbelyaudio.com.

Principles of Power, tube audio books, kits,FAQ www.londonpower.com.

The LP track carries both stereo channels vs.alternating pits on the CD. The LP remains su-

perior with “tight coupling for audio,” $15; usingan “OTL amplifier circuit," $35; driving efficient“Bass Horn” plans, $50. Visa, MC. STEENAUDIO LAB, PO Box 2185, Vancouver, BCV6B 3V7, CANADA.

Just invented! The amazing new Bullet Plug®connector – the best sounding RCA plug evermade. Improves the performance of any ana-

logue, digital, or video Interconnect. Now US$36, set of 4. Information and Dealer listing atwww.bulletplug.com.

The BassZilla has Landed! Finally, a high-efficiency and affordable loudspeaker kit by Dick Olsher. Kit plans at: http://www.blackdahlia.com.

AD INDEXADVERTISER PAGEACO Pacific Inc....................................................70Adire Audio ..........................................................83 All Electronics Corp. ............................................15Antique Electronic Supply....................................76Antique Radio Classified......................................75Antique Radio New ..............................................67Aquablue ..............................................................29Atma-Sphere Music Systems ..............................71Audio Amateur Corp.

audioXpress subscription..................................78Circuit Actions CD ............................................44Classifieds .................................................. 93,94Cleaning Kits ....................................................82Loudspeaker Design Cookbook 6th Edition ......32PC Boards ........................................................52Velleman Kits ....................................................24Web Shopping ..................................................63

Audio Consulting..................................................91Audio Electronics Supply

(Cary Audio Design) ......................................CV3Audio Transformers ............................................71Beige Bag Software ............................................45Buggtussel ..........................................................28CE Distribution ......................................................6Classic Audio Reproduction ................................23Curcio Audio Engineering ....................................69Danish Audio ConnecT........................................35Divergent Technologies ..................................19,37Dynasonic Ltd ......................................................57EIFL......................................................................27

ADVERTISER PAGEGasoline Alley ......................................................73Hagerman Technology ........................................83Hammond Manufacturing ......................................7Harris Technologies ............................................33Hi-Fi Do Inc. ........................................................21Hovland Company ..............................................75International Audio Group....................................53International Components....................................74Imagaural Acoustics ............................................77Kimber Kable/WBT-USA ..................................CV2Lambda Acoustics................................................31Langrex Supplies ................................................43 Linear Integrated Systems ..................................73Madisound Speakers ..........................................11Mahogany Sound ................................................79Mapletree Audio...................................................76Marchand Electronics, Inc. ................................80McFeely’s ............................................................79Meniscus Audio. ..................................................81Monarchy Audio ..................................................13Morel (IL) ................................................................3Parts Express Int'l., Inc. ..................................CV4Pioneer Speakers

Employment ......................................................94Plitron Manufacturing, Inc. ................................1,23Solen, Inc. ........................................................ 9,17Sonic Craft ..........................................................41Sound Clearing House ........................................80Speaker City USA................................................86Swans Speaker....................................................85

ADVERTISER PAGEThe Parts Connection ....................................46-49Thetubestore.com................................................68Vacuum Tube Valley............................................51WBT-USA/ Kimber Kable ................................CV2World Audio Publishing........................................25Zalytron Industries ..............................................55

CLASSIFIEDSApex Jr ................................................................93Audio Classics......................................................93Billington Exports ................................................93Black Dahlia Music ..............................................93Borbely Audio.......................................................93Carl Mosely ..........................................................94Eichmann Technology ........................................93ESL Information ..................................................93Forte Acoustics ....................................................94London Power......................................................93Michael Percy Audio Products ............................94Sonny Goldson ....................................................94Steen Audio Lab ..................................................93The Cable Company............................................94

AUDIO NEWS/NEW PRODUCTSAMS, Inc.................................................................6Aved/Generation Electronics ................................6Electro-Harmonix ..................................................6SMART Devices, Inc. ............................................6

ElectrostaticLoudspeakerComponents &

InformationParts for the DIY ESL

EnthusiastFrom: Barry Waldron's

ESL Information eXchange1847 Country Club Drive

Placerville, CA 95667(530) 622-1539

Website: www.jps.net/eslinfo.

Your source for Affordable ElectronicsSub-Amps, Caps, Teflon Wire & More

WWW.APEXJR.COM3045 Orange Ave. 818-248-0416 VLa Crescenta, CA 91214 818 248-0490 Fax

All types of audio tubes. 300B 6DJ8ECC81 ECC83 KT88 Mullard GECSylvania. Discount for large quantity.Minimum order £100 Export (£50UK)Billington UK. Tel (0)1403 784961.Fax (0)1403 783519.Email/website www.bel-tubes.co.uk


Pioneer Speakers, Inc. has immediate openings at its Design Center in San Diego, CA:

Senior Transducer EngineersResponsibilities include designing transducers and managing projects from prototyping through production. Requires experience with advanced,computerized speaker analysis and modeling software, background in project management, strongknowledge of materials, and excellent communicationskills. BSEE, BSME or equivalent and minimum5 years experience in the design, development, prototyping, testing and manufacturing of loudspeakertransducers.

Assistant Transducer EngineersResponsibilities include assisting in the design anddevelopment of transducers and components, buildingloudspeaker samples, performing acoustical tests andanalyzing test data, producing component drawings,BOM's and other documents, and the management ofdesign projects from prototyping through production.Requires experience with CAD, computerized analysis and simulation. BSEE, BSME or equivalent,minimum 3 years experience in audio, and excellentcommunication skills.

Pioneer Speakers, Inc. offers competitive salaries,complete relocation assistance, and a benefits planthat includes health and life insurance and 401(k).Resume and salary requirements to:

Pioneer Speakers, Inc.Attn: Human Resources132 Mile of Cars WayNational City, CA 91950Fax: (619) 477-1702 Email: [email protected] EOE

WANTEDTUBE HIFI, HORN SPEAKERS, and HIGHEND. Altec, Marantz, JBL, Mcintosh, Electro-voice, Fisher, Scott, Dynaco, Leak, Eico, AudioResearch, CJ, Quad, Bedini, etc. Sonny,Phone: 850-314-0321 Fax: 850-314-0284 [email protected].

Older Altec, JBL, Electrovoice, Klipsch, speak-ers; Marantz, Mcintosh, Fisher, etc. electronics.Carl, 770-992-0653 or fax 770-992-3802, [email protected].

FOR SALEMatched pairs of Raytheon GERMANIUMPNP audio transistors for audio preamps orguitar effects. Cross to ECG 102A or 2N44A.NOS, $6.50 per matched pair, plus shipping.Quad matched sets and matched sets of 6 ormore also available. Richard, 310-456-9277,e-mail [email protected].

New: Five Dynaudio D52 af, $75 each; threeScanspeak 18W/8546 7″ Kevlar, $90 each;one poly, $40; six Vifa M26WR 10″ woofers,$55 each; one Eton 4″ 4-206/25-HEXmidrange, $30. All unused parts, tested forpurity. Includes shipping. Daryl, 530-265-2575, e-mail [email protected].

WANTEDSchematic for Adcom preamp/tuner ModelGTP-500 or GSA 545 amplifier. Will paypostage and handling. Brett, e-mail: [email protected].

I have 12 pairs of Hitachi J56 and K176MOSCODE and want to build a MOSCODE600, but I don’t have any of its design info.Can anyone help me? I can spend somemoney for its PCB, schematic, and operationguides. C.K. Wong, [email protected].

I have an old 1986 vintage Akai CD player,model CD-A30, but no manual or schematicfor it. It has a multipin outlet on the back la-beled “Digital Subcode.” I am trying to figureout what outputs are available through thisoutlet. Any information GREATLY appreciat-ed. Please contact Don, [email protected].

Yard Sale

usedcable.comWe buy used cables.We sell used cables.

Good advice.

FORTE ACOUSTICS, INC.

Complete Design/Testing Services for theSpeaker Builder−Custom Loudspeakers

www.forteacoustics.com(516) 249-9632


HP339A audio distortion analyzer. Must be ingood working order. Andy Rawling, FAX +1630-214-3687, e-mail [email protected].

Two Focal 7N402DBE drivers, preferablyNOS. Ruud, Phone/FAX 415-383-4958.

5. Fleetwood Mac, The Dance. WarnerRecords 38486-2. This group’s person-al conflicts and music defined a gener-ation. Twenty-two pure gold songs.First-rate cinematography. Double-sided disc with 5.1 Dolby Digital onone side, stereo LPCM on the other.Solid live quintet sound, with a sub-stantial amount of backup musicians.Plenty of instrument details, but thesoundstage is not quite there. The bassis a little overwhelming on the stereotrack, but it is nothing that cannot becompensated with the bass control.Lyrics may be subtitled.

6. Peter Frampton, Live in Detroit.NuagesMusic D88161GDVD. Nobodyrocks the house like Peter Frampton.This is a solid performance of classicFrampton songs. Extra-crisp wide-screen video thanks to HDTV master-ing. Plenty of bonus features. Availablesurround tracks in both 5.1 Dolby Digi-tal and DTS. Choose the latter. Thefour-piece rock band does not havemuch soundstage, but the recordinghas engaging audience ambience. Thestereo track is heavy on bass, so turndown that control.

Fernando GarciaBrownsville, Tex.

Let’s hear from you. Simply describeyour seven favorite pieces (not to ex-ceed 1,000 words); include the namesof the music, composer, manufacturer,and manufacturer’s number; and sendto “Test Tracks,” Audio Amateur, Inc.,Box 876, Peterborough, NH 03458. Wewill pay a modest stipend to readerswhose submissions are chosen forpublication.

Test Tracksfrom page 96

visit us onlinewww.audioXpress.com

Editorialfrom page 2

today. But not here. It cost NASA a Marssatellite the other day.

If you argue that decimalled faradsare like Gary Galo’s (one of our regularcontributors) favorite, the current flowissue, in that the millions of textbookshave all adopted the “mistake,” and itwould be just too expensive and diffi-cult to start changing all the literatureat this stage of our existence, then I donot believe the issues are comparable.We have made the change from the oldmicro-micro farad of the thirties andforties. We could easily begin changesin the books, catalogs, and spec sheetsstarting in the new century. Most of theEuropean electronics press already hasthis style in place.

One historical note might be in orderhere. We honor Faraday and other dis-coverers of electronic properties by des-ignating such quantities with a numberand the capital letter of their surnames.Faraday’s µF, Henry’s µH and mH, Ni-colo Volta’s 1V or µV. Poor Georg Ohmhas the misfortune to have a namewhose capital is too similar to the zero,so the elders resorted to the good oldGreeks again, picking the omega, Ω, todesignate the ohm. When we spellthese quantities, however, the properuse is lowercase, which indicates thatwe’re talking about a quantity, not thedistinguished person.

It is time to stop decimating thefarad. The standard math notation forthese little quantities is much moreeconomical, rational, logical, and alsoneater. These magazines will continueto convert those ugly decimals to theirproper designators. I hope you willspend some time thinking through thechange and adopt it for yourself.—E.T.D.

This column originally appeared inGA 6/99, SB 7/99, and AE 6/99.

to its rated electrical power: the caseMr. McChesney raises. There is the im-plicit assumption here that your box de-sign is not displacement-limited at anyfrequency so that the electrical powerlimit applies.

Another case would be the home userwho uses 1kW amplifiers and plays at veryhigh power levels. If I was sure the appli-cation was playing full bandwidth “typi-cal” music, I would size the Zobel resistorat about one quarter the driver rating. Ifthe music to be played maintained a con-stant high level, then I would raise theZobel resistor rating to one half the driverpower rating.

2. Narrowband input frequency range. Thisis where the system may see a high powerinput at a single frequency or narrowbandof frequencies. Here, for safety, I woulduse Zobel resistors with a power ratingmatching the average power you plan toput into the system. To consider physical-ly smaller resistors, I would measure orcalculate the maximum Zobel resistor dis-sipation that could occur.

3. Multiple driver systems. Remember, if youhave a group of woofers using a singleZobel, the resistor (and capacitor) mustbe sized for the power to all thesewoofers. Multiple Zobels might make bet-ter sense in terms of physically smaller re-sistors and less current through the Zobelcapacitors.

SummarySo, in general, for home speakers used at ra-tional power levels, I get by with 10W Zobelresistors for the woofer and 5W for themidrange driver. When running a woofer toits full power rating, the Zobel resistor dissi-pation might approach one quarter of the fullpower rating of the woofer, depending on theapplication. Not knowing the application, I’mafraid I can’t be more specific than this.

As always, experience is the best guide. Ifyou are not having problems with the Zobelresistors as you are currently sizing them,then why change? I address the subject ofpower dissipation in the resistors used incrossovers more generally in SB 8/96 (“SBMailbox,” p. 51).

Ask AXfrom page 89

Test Tracks

Reader-submitted favorite selections totest audio systems.

Those who have experienced musicvideos on DVD know that it will revolu-tionize the way music is enjoyed, espe-cially for live performances. Not onlydoes it have the benefits of the plain oldCDs, but its added visual dimensionprovides a stunning extension to thelistening pleasure.

Additionally, a well-produced DVDhas a host of features that conventionalCDs can only dream of. True, somenewer CDs feature video clips—withpuny resolution—and computer interac-tive software, but the medium is con-strained by its limited capacity. DVDsoffer that plus high-quality video clips,still pictures, bios, lyrics, interactivefeatures, and much more. Some comewith selectable camera angles.

And the sound...well, the sound willliterally blow you away. Many peoplewill hotly argue this statement. I be-lieve that it may be related to a bias—or,shall I say prejudice—against multi-channel sound. I must confess I was inthat same group. After all, surroundsound employs lossy compression, andwe audiophiles are always complainingthat even linear PCM steals away pre-cious music information.

Well, don’t despair. DVDs alwayscarry an LPCM stereo track with atleast 48k/16-bit sampling; you can al-ways listen in that mode if you wish.But the good news is that audio engi-neers are finally catching up with sur-round sound’s capabilities and limita-tions. Early surround recordings, espe-cially those found in movies, were artifi-cial, aggressive, and in-your-face—morea technology showcase than a musicalexperience. But an increasing numberof music videos, some of which I havepicked here, have a three-dimensionalquality and open soundstage in the sur-round mode that cannot be reproducedby only two speakers, no matter howhigh their quality.

About the only complaint I have isthat it can quickly escalate into an ex-pensive proposition. Think of it as anaudio hobby on steroids. You have theadded expense of not only the audiocomponents (multiple speakers and am-plifiers and surround decoder), but thatof the television monitor itself. The highvideo resolution will leave you starvingfor a large-screen monitor with compo-nent or S-video inputs, interlaced to pro-gressive scanning conversion, or maybea widescreen format. Hey, an upgradeto a HDTV set would be nice!

Second mortgage, anyone?

1. Eagles, Hell Freezes Over. GeffenRecords ID5529EADVD. The standardby which music DVDs will be mea-sured for years to come. The produc-tion is impeccable. Video quality isfirst rate. The sound quality is out-standing and detailed.

This DVD features both LPCM andDTS-encoded tracks, and I like listen-ing to it in one mode or the other de-pending on my mood. The stereo LPCMtrack is incredibly tight and detailed,with powerful percussion and preciseimaging. But it is the 5.1 channel, DTS-encoded track that really shines. Mypersonal judgement, which I have seenexpressed many times over, is that DTSis superior to Dolby Digital for music.The DTS version of the songs, althoughnot as detailed and powerful as thestereo track, nevertheless offer the bestsoundstage I have discovered in musicDVDs and an impressive three-dimen-sional sound. This DVD features anaudio-only DTS track that is a showcaseof superbly executed surround sound.And the music, well, you know the Ea-gles songs.

2. The Moody Blues, Hall of Fame:Live from the Royal Albert Hall.Threshold Records ID9757TEDVD.British band The Moody Blues has al-ways attracted fans of soft rock for itslush, orchestral sound, of which the

hit “Nights in White Satin” is a pre-mier example. The quartet playsalongside the World Festival Orches-tra at the fabled Royal Albert Hall. Thevideo takes full visual advantage ofthe opulent venue. The digital videoproduction is striking; the images areso detailed.

The audio portion features three dif-ferent tracks: stereo LPCM, Dolby Digi-tal 5.0 and DTS 5.0, and I would recom-mend the latter. This is not a typo—5.0is the correct format description, and itmeans that the low-frequency effectschannel is not enabled. The recordingdoes not have killer bass, and the LFEchannel is not really required.

The recording does have a stupen-dous amount of detail and space, al-though image wanders a little. Hall am-bience is fantastic.

3. John Fogerty, Premonition. WarnerRecords 38496-2. At first glance, theopportunity to listen to some old Cre-dence Clearwater Revival songs withjust a member of the original CCR line-up is not very appealing. But JohnFogerty was CCR, and the addition ofnew musicians with renewed spiritprovides for an incredibly energizedrendition of the old songs, in some in-stances livelier than the original ver-sions. Watch for the extremely ener-getic drummer. Encoded in 5.1 DolbyDigital. Lyrics may be subtitled.

4. Rick Wakeman, The Classical Con-nection. Beckmann CommunicationsID8814CLDVD. Keyboard maestrocomes alive in this DVD. WatchingWakeman’s fingers fly over the key-board is a treat in itself. The music in-cludes classical arrangements of pop-ular music. The maestro is supportedby veteran guitar player David Paton.This album is best heard in stereo, asthe 5.1 Dolby Digital track does notadd any significant depth to thesoundstage.

(continued on pg. 95)


a quick, cheap speaker pair for the basement · a quick, cheap speaker pair for the basement ... a...

Documents