Lee

Volume 22, 2009 1

ABSTRACT

AWA Review How Dunwoody’s Chunkof ‘Coal’ Saved bothde Forest and Marconi

©2009 Bartholomew Lee

Fig. 1. This undated photograph of Brigadier Gen-eral H. H. C. Dunwoody, circa 1904, shows him in thedress uniform of the era. [1] The yellow sash is adecoration for bravery in the Spanish American Warin Cuba.

United States Army Brigadier GeneralHenry H.C. Dunwoody is little known today.Yet he invented the crystal detector of wire-less spark signals about March, 1906. In theearly twentieth century his device came to beknown as the “heart of radio.” Dunwoody’scrystal detector saved Lee de Forest from a fed-eral injunction sued out by ReginaldFessenden, and then saved the Marconi Com-pany from the long-wave static that challengedthe first transatlantic circuit from Ireland toNova Scotia in 1907 and subsequent years.

General Dunwoody had a distinguishedmilitary career, serving in Cuba during the

The success of wire-less telegraphy pioneersMarconi and de Forestowes much to the inven-tion in 1906 of a simple,reliable detector of wire-less signals by a then-re-tired Army General,Henry H.C. Dunwoody.He showed thatcarborundum could actas a stable and sensitivedetector, permitting thewireless operators of theday to hear even trans-atlantic signals. Thecarborundum detectorgot de Forest out fromunder what would oth-erwise have been a com-pany-killing injunctionobtained by Fessenden,whose electrolytic detec-tor patent de Forest in-fringed. It also facilitatedthe development of anoise cancelling circuitessential to Marconi’slong wave transatlanticsuccess as early as 1907and up to the vacuumtube era. For a decade itwas state -of -the -art,stabilizing the circuitsand operations of wire-less telegraphy and thusfostering new directionsof innovation.

AWA Review

Dunwoody

2

Spanish American War, and liv-ing to 90. He was widely honoredfor his leadership of the U.S.Weather Bureau in the 1880s. In1883 he wrote a book WeatherProverbs (still available) compil-ing folk wisdom about weather inorder to make better predictions.The Arlington National Cemetery[2] summarizes his military ca-reer:

“Henry Harrison Chase

ger-Journal August 30, 1928. In-ventor of ‘Heart of Radio’ Is aFormer Fairfield Citizen. GeneralH. H. C. Dunwoody went fromhere to West Point; was in Chargeof Weather Forecasting for theWeather Bureau for Years, andheaded the Signal Corps of theAmerican Army in the CubanCampaign. Old timers rememberGen. Henry H. C. Dunwoodywhen he was a boy in school here

Dunwoody of Ohio [:]Appointed from Iowa,Cadet, United States Mili-tary Academy, 1 Septem-ber 1862[;] Second Lieu-tenant, 4th U. S. Artillery,18 June 1866... Major,Signal Corps, 18 Decem-ber 1890[;] LieutenantColonel, 15 March 1897[;]Colonel, Chief Signal Of-ficer of U. S. Volunteers,20 May 1898 [Cuba][;]Colonel, Signal Officer, U.S. Army, 8 July 1898 [re-tired as a Brigadier Gen-eral, 1904][;] Died 1January 1933.”

West Point was a lead-ing engineering school ofthe day; Dunwoodygraduated 19th in hisclass according to itsrecords. After he retiredfrom the Army and aftera brief interlude with thede Forest enterprise, hemanufactured arma-ments in several compa-nies, part of the originalmilitary-industrial com-plex. Arlington Cemeteryposts [2] a newspaper fea-ture article (and obituary)telling of his life and suc-cess as a meteorologist aswell as the inventor ofwhat it called the Heartof Radio:

”Fairfield Weekly Led-

Fig. 2. Dunwoody 1906 carborundum detectorpatent, number 857616. Note that circuit figure oneuses DC bias. Marconi used this as a basis forhis detector circuit. Note figure four displays thecarborundum between two steel plates, whichcame to be the usual configuration because of itsmechanical stability. One of the patent claims isthat the detector will not burn out when in proximityto a transmitter, an important advantage shipboardas well as in shoreside installations with co-lo-cated receiver and transmitter equipment. Thepatent omits tuning circuits.

Lee

Volume 22, 2009 3

and up to all the pranks of a livelylad. He was born in HighlandCounty, Ohio, Oct. 23, 1842, andso is in his 86th year. He now livesat Ovid, New York.”

Gen. Dunwoody employedcarborundum (which he calledcrystalline silicid of carbon), un-der mechanical pressure and elec-tromotive pressure (i.e., with abias voltage) to detect wireless te-legraphy signals. [3] His patentappears in Fig. 2.

The substance Dunwoodyworked with is now known as sili-con carbide. Nineteenth centurychemist Edward Acheson acci-dentally created this artificial com-pound in his search for a way tomake diamonds. [4] Dunwoodydiscovered its semiconductor prop-erties. It was for this discovery thathe was later said to have invented“the heart of radio.” As this notewill show, his invention sustainedthe success of the two most im-portant early wireless telegraphycompanies.

Prior to Dunwoody’s invention,Chandra Bose, the Indian physi-cist, discovered the rectifying prop-erties of galena (lead sulfide) andpatented a detector of Hertzianwaves (microwaves) that em-ployed these and related proper-ties. (The priorities of Bose arewidely discussed in recent litera-ture [5]). Karl Ferdinand Braunalso developed a rectifying crystaldiode at least as early as 1898.

A little after Dunwoody, an-other wireless experimenter,Greenleaf Whittier Pickard, in-vented a complete wireless receiv-ing system (a “crystal set” that wastunable) in August, 1906, usingsilicon alone and what came to beknown as the “cat’s whisker” in-terface. His patent [6] appears inFig. 3.

Before the crystal detectors,commercial, naval and amateur

wireless signals were detected, i.e.,made hearable or readable, by useof the Edouard Branley filings co-herer, the John Ambrose Flemingvalve vacuum tube diode, theMarconi magnetic detector (collo-quially known as the “maggie”),and by electrolytic processes (pri-marily discovered by CanadianReginald Aubrey Fessenden).

Lee de Forest employed GeneralDunwoody in 1906 in New York,after he retired as the Army’s Sig-nal Officer. In his autobiographyFather of Radio, [7] de Forest saysof the carborundum device that itwas: “... a simple rectifier whichhad been discovered by GeneralDunwoody, former chief SignalOfficer, and now vice-president ofthe company”, i.e., the de ForestWireless Telegraph Company, Inc.De Forest brags about how muchbetter the carborundum detectoris than the Marconi magnetic de-tector for “reading through inter-ference or static.” He delights incalling Marconi Company wirelessoperators “limey Sparks.” He saysthat many a “limey Sparks,” inorder to effect better reception ofMarconi stations’ spark signals,took to “concealing about his per-son a small chunk of ‘coal,’ as theDunwoodie [sic] carborundumwas later called.”[8]

The wireless companies of theday prohibited use of non-com-pany devices, to avoid patent in-fringement claims. Hence the needfor concealment. But a cleverSparks could substitute the bettercarborundum detector for aMarconi detector at sea, with noone the wiser.

At the time (and as always) deForest was challenged by litiga-tion. The Canadian wireless pio-neer Fessenden in particular hadsued de Forest about commercialuse of an electrolytic detector thatFessenden had invented about

AWA Review

Dunwoody

4

Fig. 3. Pickard “crystal set” patent of 1906, number836581, employing silicon and what came to be calleda “cat’s whisker” contact with the silicon. Pickardclaimed silicon as a wireless detector that did not re-quire a bias voltage with the inconveniences of batter-ies and the like. His circuit isolates the detector and itstuning from stray capacitance and body proximity ef-fects.

Fessenden’s patent.De Forest gloatedabout the unhappyconsequences forFessenden’s litigationof his directors’ deci-sion: “... happily forme as it later turnedout.” [9]

Dunwoody’s in-vention of thecarborundum detec-tor thus saved deForest’s wireless en-terprise. Fessendenhad persuaded acourt to issue an in-junction against deForest’s use of anyelectrolytic detector.Dunwoody’s detectorworked as well andlikely better, thusavoiding the sanc-tions of the injunc-tion and permittingthe company to re-main in business.[10]

The judge’s find-ings of fact and legalopinions in the detec-tor injunction UnitedStates Circuit Courtlawsuits were offi-cially reported fortheir precedentialand technical value,and may be found inthe Federal Reporter

1903. De Forest in his autobiog-raphy goes on to say that theAmerican De Forest Wireless Tele-graph Company’s directors haddecided to use, at the company’sstations, the Dunwoodycarborundum detector. They ineffect decided to forego de Forest’slegally problematic “spade” elec-trolytic detector at issue withFessenden. De Forest almost byaccident no longer infringed

series of law reports.[11] DeForest’s detector litigation reachedan acme in the Audion andFleming Valve dispute in 1916[12]. This U.S. District Court ju-dicial opinion discusses detectors,including Dunwoody’s, at length.

One of the further advantagesof the Dunwoody carborundumdetector in marine service, beyondits sensitivity, was that it wasmechanically very stable. This

Lee

Volume 22, 2009 5

was so because in its initial imple-mentation two metal plates heldthe “coal” firmly. De Forest alsoconcluded that its sensitivity wasindependent of the pressure of thecontacts. [13] There is some dataon relative sensitivities of earlydetectors. Carborundum appearsto be about as good as the“maggie,” and not as good as sili-con, see Fig. 4. [14]

The year 1906 was said at thetime to have delivered the worstatmospheric static that the na-scent wireless art had yet faced.Sunspots also peaked in 1906 - 07[15] although no one at the timeunderstood any connection withradio propagation or the auroraleffects on it. Atmospheric static,especially on what we now call thelong waves, was the bane of op-erators world-wide. In masteringthis challenge, the carborundumdetector played an important rolein the development and success ofMarconi’s international wirelesscircuits as early as 1907.

Marconi’s first working trans-atlantic circuit connected Clifden,Ireland with Glace Bay, NovaScotia. A Nova Scotia archive [16]describes the role of Dunwoody’s

chunk of “coal” in this initial andsuccessful first transatlantic cir-cuit and some subsequent refine-ments:

”The construction of a receiv-ing station in Louisbourg, NovaScotia and a similar one inLetterfrack, Ireland in 1912-1913represented the final phase of theestablishment of the firsttransAtlantic radio communica-tions service.

“The first Louisbourg detectorwas the carborundum detector, arugged crystal detector inventedby General H.C. Dunwoody of theU.S. Army. Marconi used a circuitcalled the “balanced detector,”inwhich two carborundum diodeswere connected and electricallybiassed [sic: biased] in such a waythat strong impulses produced bylightning discharges would tend tocancel out, whereas the weakersignal that the operator was try-ing to copy would be detected.”

The balanced carborundumcrystal detector circuit appears tohave been developed primarily byMarconi engineer H. J. Round.He later, on behalf of the RoyalSignal Corps, worked with MajorEdwin Howard Armstrong during

Fig. 4. Table of relative sensitivities of wireless detectors from V.J. Phillips. [14]

AWA Review

Dunwoody

6

the First World War. His work onthe balanced detector circuit re-sulted in the Marconi Model 16balanced Crystal Detector circa1916.

Marconi engineer Elmer T.Bucher, in Practical Wireless Te-legraphy, (1918) provides circuitdetails and a wealth of technicalinformation on the carborundumdetector. [17] See the graphicalappendix to this note for Bucher’sillustrations of the characteristiccurve of carborundum, his analy-sis of its workings as a biased de-tector and two Marconi imple-mentations of the steel andcarborundum detector.

The Bucher schematic in Fig.6. correlates with the Model 16. Itworked by listening only on onecrystal by holding the bias voltageon the other too low. The sharpinput voltage of a static crash,however, moved both crystalshigh on their characteristic curves,and the resulting nearly equal de-tected voltages canceled eachother out.

Note the top right buzzer cir-cuit to enhance audibility of thedetected almost-continuous-wave signals, known as un-damped spark signals, whichMarconi rotary spark transmitters

Fig. 5. The Marconi Model 16 balanced Crystal Detector circa 1916 at theBellingham, Washington, American Museum of Radio and Electricity of John D.Jenkins. See: http://www.sparkmuseum.com/MARCONI.htm; copyrighted imageused by permission. Note the two sliding linear rheostats at the bottom left, usedto unbalance the bias voltages so that only one carborundum detector workedat a time (until a static crash activated both, cancelling the crash). Note alsothe adjustable tubular capacitor at the top right, which looks like the “Billi con-denser,” the last variable capacitance in the tuning circuit, the “secondary cir-cuit,” of the American Marconi Type 107A tuner, which also used carborundumas its primary detector.

Lee

Volume 22, 2009 7

generated. Thus, by taking advan-tage of the characteristic curve ofthe biased carborundum rectifiersto cancel out strong signals (light-ning crashes) but pass weaker sig-nals of intelligence by one or theother detector, Marconi masteredatmospheric static. In the laterdevelopment of this device, atuned buzzer provided a hetero-dyne to make undamped (con-tinuous wave) signals audible.

It is thus fair to say thatDunwoody’s “coal” may well havesaved the two most importantwireless communications compa-nies in the nascent days of the ra-dio art, before the use of thevacuum tube. Dunwoody’s detec-tor had the advantages of me-chanical reliability, as comparedto galena or silicon detectors us-ing “cat’s whisker” interfaces. Nordid wireless operators have to findor reset a “sweet spot” especiallyafter each transmission.Carborundum’s performancewhen biased permitted the devel-opment of the noise cancelling cir-cuit, for which other detectorswere unsuitable. The manufac-

ture of carborundum presentedno difficulty and it did not have tobe mined and selected the waygalena crystals, for example, did.Carborundum’sensitivity sufficed,especially in view of the tremen-dous power put out by the earlycommercial wireless transmitters,in the range of 300 kilowatts.

Dunwoody’s carborundum de-tector bridged the receiving tech-nology of the filings coherer of theearly experimenters, includingMarconi, and the later and soondominant technology of thevacuum tube triode. Its availabil-ity, reliability and performancemade it the detector of choice fornearly a decade, until the adventof Armstrong’s regenerative cir-cuit for de Forest’s triode circa1916.

In the absence of Dunwoody’sdiscovery, both commercial andamateur wireless operationswould no doubt have employedthe several other minerals anddetector technologies available,but at some cost especially in reli-ability. The great advantages ofcarborundum permitted the ener-

Fig. 6. Round’s circuit for using two balanced carborundum detectors to mini-mize static in CW reception (Bucher’s Fig. 300 B Marconi’s System for theReception of Undamped Oscillations) [17].

AWA Review

Dunwoody

8

gies of innovation to focus on newcircuits and techniques, and thenon the quantum leap of vacuumtube technology for detection,amplification and oscillation,which provided the major elec-tronic advances of the first half ofthe twentieth century.

NOTES1. The source of the photograph of

then Col. Dunwoody is a page inthe Archive of the New YorkPublic Library, No. 1227290,NYPL Digital Gallery.

2. Arlington Cemetary - http://www.arlingtoncemetery.net/hhdunwoody.htm . This goes on:“Fairfield Ledger Tuesday,January 3, 1933 Former FairfieldResident Passed Away Jan. 1. Brig.Gen. Harry [sic; Henry] HarrisonChase Dunwoody, 90, formerchief signal officer of the U. S.Army, died at Interlaken, NewYork, Sunday according to wordreceived here by friends.” It isinteresting to note GeneralDunwoody’s grandson HaroldDunwoody served in combat inWorld War Two, Korea andVietnam, retiring as a muchdecorated Brigadier General. Hisgreat- granddaughter, AnnDunwoody, is now the U.S.Army’s first woman four stargeneral officer.

3. Dunwoody’s patent is number837616 dated December 4, 1906,filed March 23, 1906.

4. http://www.reference.com/browse/all/crystal%20detector.

5. See, e.g., Tapan K. Sarkar, et al.,An Appreciation of J. C. Bose’sPioneering Work in Millimeterand Microwaves, in History ofWireless (2006), John Wiley &Sons, at pps. 291-310.

6. Pickard’s patent is number836581 dated November 20,1906, filed August 30, 1906.

7, 8, 9. Lee de Forest, Father ofRadio, self-published, 1949, atpps, 127, 195.

10. A. Douglas, “The CrystalDetector” IEEE Spectrum April

1981 p. 66.11. National Electric S. Co. v. De

Forest Wireless Tel. Co., 140F[ederal Reporter] 449 (U.S.C.C.N.Y. 1905); National ElectricSignaling Co. v. De Forest W. T.Co., 145 F. 354 (U.S. C.C.N.Y.1906).

12. Marconi Wireless T. Co. v. DeForest Radio T. & T. Co., 236 F.942 (U.S. D.C.N.Y. 1916).

13. Lee de Forest, in ElectricalWorld, September 8, 1906reported in D. McNicol, Radio’sConquest of Space, p. 125, -27(Arno Reprint, 1974)).

14. V.J. Phillips, Early Radio WaveDetectors, (Science Museum,I.E.E., and Peter Perigrinus;London 1980) p. 213, Table 10.1.

15. E. W. Maunder [he of theMaunder Minimum episode of fewsunspots], “The sun-spots of 1906- II” in The Observatory, Vol. 29,p. 348-352 (1906) archived at:http://adsabs.harvard.edu/full/1906Obs....29..348M; see also forthe intensity in 1907: http://a d s a b s . h a r v a r d . e d u / f u l l /1907Obs....30..126M .

16. The Official Research Site for theFortress of Louisbourg (NovaScotia, Canada) - http://fortress.uccb.ns.ca/Marconi/marconi2.htm.

17. Elmer T. Bucher, PracticalWireless Telegraphy, (1918) pp.172, 286-87, 288ff and figs. 153(a,b,c), 157, 157a, 158,163, 198,199, 300.

This article was peer reviewed.

APPENDIX

The following illustrations fromBucher, Practical WirelessTelegraphy [17], provide graphicaldata on carborundum’s performanceas a detector and illustrate workingMarconi Company detectors.

The so-called “characteristiccurve” of carborundum (and othersemi-conductors) is not linear,according to Bucher. (This is alsotrue for other semi-conductors andvacuum tube diodes as well.) An

Lee

Volume 22, 2009 9

ordinary resistance curve is linear,i.e., the more voltage applied, themore current flows in an exactlyproportional way: doubling thevoltage doubles the current. A non-

Assume a bias voltage of twovolts. An alternating radiofrequency (RF) voltage comes to thedetector from the antenna andtuning circuit. Assume it is two voltspeak to peak. The negative(subtractive) one volt peak willdecrease the bias current flowingfrom three micro-amperes to onemicro-ampere, down the curve as itwere. The positive, additive one voltpeak will, however, have a greatereffect. It will move the flow of currentfrom three micro-amperes to sixmicro-amperes. The effect is all thegreater higher up on the curve, atthree volts bias for example. Thusthe detector permits more currentto flow in one direction when underthe influence of the alternating radiofrequency voltage.

Bucher then presents as his nextstage of his explanation a diagramshowing how the differentialresponse to subtractive and additiveparts of the RF input result in anaverage current flowing in only inone direction. He posits a biasvoltage of three volts. The telephonereceiver diaphragm integrates the

linear resistance on the other hand,permits a more than proportionalcurrent to flow as voltage increases.Bucher’s curve for carborundumillustrates this characteristic.

Bucher’s illustration of how the characteristics of biased carborundum make forasymmetrical if not one-way passage of RF and detection of it by way of differ-ential current flows and mechanical integration by the telephone receiver dia-phragm in response to them.

Bucher’s characteristic curve forcarborundum.

AWA Review

Dunwoody

10

oscillating E.M.F. [electromotiveforce or voltage] being impressed onthe crystal is greater than the

additive larger pulses into a virtuallarger steady current. Buchersummarizes the effect this way: “ ...the added voltage due to the

Marconi vertical carborundum detec-tor.

subtracted voltage and that the finaleffect of this is an increase of currentthrough the head telephone circuitover the duration of one wave train.”

Marconi and de Forest had toreduce theory to practice. Bucherpresents two illustrations of workingcarborundum detectors. The first isa vertical configuration, likely in useon a panel in shipboard installations.Bucher suggests use of a steelphonograph needle as the contactwith carborundum in a holder,reminiscent of Pickard’s holder forsilicon as illustrated in his patent.

The second illustration alsoshows adaptive re-use. In 1904Fleming had contributed thevacuum tube diode Fleming Valve

American Marconi’s substituted hori-zontal carborundum detector for aFleming Valve.

to Marconi’s operations. It wasreliable but insensitive. Accordingto Bucher, it was American MarconiCompany engineers who cleverlyaffixed a horizontal carborundumdetector to a screw base for theFleming valve, permitting itssubstitution and use of the filamentvoltage as the bias. AmericanMarconi had been De ForestWireless, so perhaps that company’sfondness for solid-statecarborundum, even over thevacuum-state Fleming valve,carried over to this adaptation. Itmay also be the case that Flemingvalves were hard to get and easy tobreak or burn out.

Lee

Volume 22, 2009 11

Bartholomew (Bart) Lee,K6VK, xKV6LEE, holds anextra class amateur radio license.He has enjoyed radio and radio -related activities in many parts ofthe world, most recently inSingapore, Australasia, and thePapua New Guinea area. Radiohistory and technology have fas-cinated him since he made hisfirst crystal-set with a razor bladeand pencil lead detector more than50 years ago. He is a widely pub-lished author on legal subjects,and most recently on the historyof radio. He has written about ra-dio in intelligence operations(from 1901 forward, including theCIA on Swan Island in the 1950sand 1960s), the history of wire-less telegraphy, especially thework of Marconi and the indepen-dent developments on the U.S.West Coast, short wave radio andits history, radio ephemera includ-ing radio stamps, and radio inemergency and disaster response.Since 1989, he has made almost20 annual presentations to theAWA Conference on his researchinterests, including the develop-ment of television in San Fran-cisco in the 1920s. The AWA pre-sented its Houck Award for Docu-mentation to him in 2003 and theCalifornia Historical Radio Soci-

ety made its ‘Doc’ Herrold Awardto him in 1991 in connection withthe Electronics Museum of thePerham Foundation. In 2001,during the New York City disas-ter recovery operations followingthe “9/11" terrorist enormity, heserved as the Red Cross deputycommunications lead from Sep-tember 12 to September 21. Barthas also served as the San Fran-cisco Auxiliary CommunicationsService Liaison Officer, and as anARRL ARES Emergency Coordi-nator. He presently serves as anARRL Government Liaison andVolunteer Counsel, and as coun-sel emeritus to the California His-torical Radio Society. Bart is a triallawyer by trade, and a formerAdjunct Professor in Law & Eco-nomics at a San Francisco Uni-versity. He is a graduate of theUniversity of Chicago Law School,and St. John’s College. He invitescorrespondence atKV6LEE@gmail.com.

ACKNOWLEDGEMENTI am grateful to the distinguishedSpanish antique radio collector andhistorian Salvador Munoz Gomarizfor stimulating my interest inGeneral Dunwoody; I am happy tohave supplied for his use thephotograph of Gen. Dunwoody andan earlier draft of this article. Seehttp://escibalofilms.blogspot.com/2 0 0 9 / 0 1 / h e n r y - h - c -dunwoody.html.

ABOUT THE AUTHOR

Bart Lee photo by Paula Carmody,taken in Indonesia; copyright Bart Lee2009.

AWA Review

Dunwoody

12