history of communications - comsoc.org · ieee communications magazine • september 2009 21 and a...

IEEE Communications Magazine • September 200920 0163-6804/09/$25.00 © 2009 IEEE

INTRODUCTION

Edouard Branly, a son of France, maybe regarded as one of the very first pio-neers in the field of wireless transmis-sion. However, he himself denied beingone of the fathers of radio communica-tions; he is more accurately the discover-er of a physical effect (the Branly effect)that gave him the opportunity to devisean efficient wave sensor (the coherer)that permitted the invention of radio.Interestingly, he also contributed to theidentification of the role of the aerial.Finally, what is called today the Branlyeffect is still largely unexplained andconstitutes a subject of investigation.

THE GENESIS OF A DISCOVERYEdouard Branly was born in 1844. Afterhaving performed brilliantly in classical

secondary studies, he chose sciences andentered the Ecole Normale Supérieure(at that time Louis Pasteur was the headof studies of this university). In 1868 hewas appointed as engineer (later deputy-director) of a physics laboratory at ParisSorbonne University. He there obtaineda doctoral degree in physics in 1873. Hisresearch field was then electrostatic phe-nomena in batteries [1]. In 1870, duringthe Franco-Prussian War, he served as amilitary engineering officer (during thiswar, the French Second Empire wasreplaced by the Third Republic). Hethen left the Sorbonne in 18751 for thenewly organized Catholic University ofParis, the law of 12 July 1875 now per-mitting private universities. In 1880 itbecame Institut Catholique, as the termuniversity was then reserved again forstate institutions. Leaving the Sorbonne

for a Catholic university raised the riskof marginalization in the secular FrenchRepublic of that time.2 Moreover, whilea modern research laboratory installationwas initially expected by Branly [2], hewould only benefit from the equipmentof a teaching laboratory located in a for-mer dormitory subject to vibrationslinked to the traffic of the nearby rue deVaugirard. Branly would have to wait 60years before benefiting from a modernbuilding for his laboratory. One possibleexplanation is that after the 1870 war,the building of the Sacré Coeur Basilicawas decided on and diverted a lot of thediocese of Paris funds from institutionslike the Institut Catholique [2].

In 1877 he was again a student, butat the faculty of medicine of Paris, con-sequently interrupting his researchactivities. In 1882 he was awarded aphysician’s degree after preparing athesis on hemoglobin concentrationanalysis in blood by optical means.

At the end of the 1880s, Branly wentback to his research in pure physics, con-centrating on the influence of irradiationon the electrical conductivity of varioussubstances. In June 1890 he used aWimshurt machine to create sparks andto study the electrostatic discharge ofvarious substances submitted to the lightfrom the spark (i.e., UV-rich radiation)[3]. Branly had devised a first circuit tocreate sparks using the Wimshurtmachine, and then a second very simplecircuit: a Daniell battery, a galvanometer,

HISTORY OF COMMUNICATIONSEDITED BY MISCHA SCHWARTZ

Critical to the success of Marconi’s pioneering work on radiotelegraphy was the device needed to detect the radio telegraphmessages on reception. Marconi finally settled on the “coherer.”The invention of the device is often attributed to Oliver Lodge, aBritish physicist of the late 19th century. The detection propertiesof this device were, in fact, discovered by Edouard Branly, aFrench experimental physicist of the same period. As Dr. Dilhacmakes clear in the following, Lodge himself attributed the signaldetection properties of the coherer to Branly, coining the term

coherer, and calling the device the “Branly coherer.” The articleclearly and succinctly discusses Branly’s life, and describes thecircumstances around which the coherer detection property wasdiscovered by Branly. As pointed out by Dr. Dilhac, the physicsbehind the coherer long eluded scientists, and is just now, hope-fully, becoming understood. All this makes for a fascinatingstudy. I, personally, have enjoyed reading the article. I am sureall the readers of this column will find it interesting as well.

—Mischa Schwartz

INTRODUCTION BY EDITOR

EDOUARD BRANLY, THE COHERER, AND THE BRANLY EFFECTJEAN-MARIE DILHAC, UNIVERSITÉ DE TOULOUSE (LAAS-CNRS AND INSA)

1 Partly because he refused to marry the daugh-ter of his faculty head.

2 This may explain the fact that, despite the pro-posals made as early as 1904 and later in 1915by the Stockholm Academy of Science, Branlywas never awarded the Nobel Prize, the Frenchexperts consulted by Swedish academicianshaving made other proposals [2].

IENYCM2471.indd 1 5/12/09 3:25:45 PM

LYT-HISTORY-September 8/21/09 1:16 PM Page 20

IEEE Communications Magazine • September 2009 21

and a metallic disk all wired together inseries. The disk was initially electricallycharged. The two circuits were initiallyclose to each other, so the light from thespark illuminated the disk. Following thespark, in some cases a dramatic increasein disk conductivity could be detected bythe galvanometer. In November 1890 hereplaced the disk by a tube filled withoxidized Zn particles.3 Just after thespark, it was again found that the con-ductivity of the tube was increased byseveral orders of magnitude. He theninserted various obstacles between thespark and the tube in order to evaluatethe properties of the propagation of theinvisible part of the light (i.e., UV)between spark and tube. At one point heinserted a piece of cardboard and noticedthat the effect persisted (Fig. 1). He thenput the circuit made of the battery, gal-vanometer, and tube in another room, 20meters away, separated by thick wallsand a courtyard from the Wimshurtspark-emitting circuit: the effect persistedwhile neither the light from the sparknor its sound could be seen or heard byBranly sitting by the tube while his aideRodolphe Gendron was operating theelectrostatic machine. A small shock wasfound to restore the initial conductivityvalue, while a new spark allowed thephenomenon to be repeated. His paper

on this work was published in French [4]on 24 November 1890.4

Branly called the detecting device (thetube filled in with metallic particles) aradioconductor: this is likely to have beenthe first use of the prefix radio in the con-

text of wireless communications [5]. ForBranly, the term radio referred to the elec-tric radiation from the spark. However, asearly as January 1891, Branly followed theNovember 1890 paper with a publicationon the role of aerials (he did not use theterm) in relation to the distance of trans-mission, together with the effect of a Fara-day cage [6]. The transmission distance wasthen (1891) on the order of 80 meters [3].

BIRTH OF THE COHERERWhat is referred to today as the Branlyeffect is an electrical conduction instabilityappearing in a slightly oxidized inhomo-geneous metallic conductor (usuallyplaced in a glass or ebonite tube and sub-mitted to moderate mechanical pressure)when an external disturbance is applied:the initial high electrical resistance ofmany megohms due to the naturally oxi-dized surfaces then falls to a relativelylow value of a few ohms. A mechanicalshock restores the initial high resistancevalue [7]. The disturbance can be a highcurrent (phenomenon discovered in 1835by P. S. Munk using the discharge currentof a Leyden jar), but in 1879 D. E. Hugh-

HISTORY OF COMMUNICATIONS

3 He observed the same phenomenon with Fe,Al, Cd, and Bi.

4 At that time, publishing in Comptes-Rendusde l’Académie des Sciences was a fast processas there was no refereeing. Reports of Branly’sresults were published in English in the (Lon-don) Electrician in June and August 1891[16].

�� Figure 1. Schematic of Branly’s experiment.

Cardboard screen

Oxidized metallic particules

Glass tube

Galvonometer

Daniell battery

Wimshurt machine

Spark gap

LYT-HISTORY-September 8/21/09 1:17 PM 1

es appears to have discovered that theabove conduction instability could also beinduced by an electrical spark at a dis-tance. However, the Royal Society ofLondon was not convinced, and his resultswere only published in 1899 [8]. Finally,in 1884 T. Calzecchi-Onesti discoveredthat the electrical conductivity of metallicpowders increases after having been sub-mitted to a sufficiently high voltage.

In 1894 British scientist Oliver Lodgerepeated Hertz’s 1887 experiments onradiation from sparks, but using a Branlytube as a detector. This was much moresensitive than Hertz’s technique using awire loop containing a small gap withinwhich electromagnetic waves wereinduced, having themselves been createdby a first arc. Hertz’s experiments, carriedout after the publication of Maxwell’s the-ory, were designed to verify Maxwell’s the-ory and to demonstrate the propagation ofelectromagnetic waves over a few meters.These tiny (much less than a millimeter inlength) sparks could only be seen in dark-ness, and in some cases a magnifier waseven required. Consequently Hertz men-tioned in a letter that electromagneticwaves were unsuitable for remote signalingdue to the difficulty of detection [9].

Lodge improved Branly’s tube by addinga relay automatically triggering a shockafter a decrease in electrical resistance,making the device usable for wireless trans-mission. Lodge coined the term Branly’scoherer (from Latin cohaere, to stick) refer-ring to his first understanding that the mod-ifications in resistance were related to smallmovements of the particles, considered asdipoles, due to electrostatic effects. Thesemovements were supposed to form conduc-tive fragile chains, therefore allowing thepercolation of electricity through the pow-der, a shock easily destroying these chains.Edouard Branly did not believe this expla-nation, arguing that the effect persistedeven if the oxidized metallic particles werestuck in paraffin, resin, or wax, and there-fore were unable to move. For this purpose,he also used large steel balls a few centime-ters in diameter arranged in a single chain,and demonstrated there again that theeffect persisted. The term coherer was notaccepted by Branly but was neverthelesswidely accepted.

Lodge and Branly later collaborated inan attempt to identify the underlying phys-ical mechanisms.5 Neither of them reallyfocused on wireless telecommunicationsapplications. It must also be mentionedthat it was only in response to Lodge thatBranly — mainly an experimenter — ten-

tatively devised a first theoretical explana-tion of the effect he had discovered, basedon ether between grains. He did not referto Hertz waves before 1895 [10], but it isworth noting that Hertz’s work was notfully accepted until 1892 after its reproduc-tion by other scientists [11].

On the other hand, it must be stressedthat without the coherer, the birth of wire-less communications would have beengreatly delayed into the next century. Thecoherer was the sensitive sensor required bythose following Branly: in contrast to thedetecting loop, the coherer was an on/offsensor delivering an electrical output, welladapted to binary codes, and likely to beconnected to electromechanical printingdevices through an electrical relay [3]. In1895 both Popov and Marconi started theirradio transmission activities using coherers.More generally, most of the experimentalwork in the field of wave propagation dur-ing the last decade of the 19th century andthe start of the 20th used a coherer as awave sensor [5]. When the first commercialwireless links came into operation at theturn of the century, reliable coherers werecommercially available [5].

In 1898 Ferdinand Braun improved theusual architecture of transmitters by imple-menting a sparkless antenna loop magneti-cally coupled by a transformer to thepower spark loop.6 The previously insur-mountable 15 km limit for transmissionrange with which Marconi was confrontedcould now be exceeded [9]. However,pulsed waves were still used, and thecoherer remained the best detector, surviv-ing the first generation change in wirelesstelecommunications (i.e., the transitiontoward long distances). In 1899 GuglielmoMarconi realized the first wireless trans-mission over the Channel [6] betweenEngland and France. The text of the dis-patch is as follows [6]: M. Marconi envoie àM. Branly ses respectueux compliments parle télégraphe sans fil à travers la Manche, cebeau résultat étant dû, en partie, aux remar-quables travaux de M. Branly. (Mr. Marconisends to Mr. Branly his regards over theChannel through the wireless telegraph, thisnice achievement being partly the result ofMr. Branly’s remarkable work.)

In 1906 Braun replaced the coherer bya crystal of Galena used as a rectifier. Aswith the coherer, it was a cheap device,easy to build, very sensitive, and alsobased on unknown physics principles.However, the crystal rectifier was verywell adapted to the continuous waves


22 IEEE Communications Magazine • September 2009

5 It is worth mentioning that J. J. Thomson onlydiscovered the electron in 1897.

6 In 1909 Braun received the Nobel Prize forPhysics together with Marconi for his contribu-tion to wireless telegraphy.

(Continued on page 24)Garner, NC, 27529Tel. + 1 919 460 8800Fax + 1 919 460 0141E-mail [email protected]

Take advantage of a great choice of CONEC indu-strial interface connectors with IP67 protection.The ideal solution for rough environments.

USB 2.0 and RJ 45 Industrial Ethernet connectionsystems available as plastic or full metal versionsfor heavy duty applications.

CONEC offers a broad range of plug and receptaclehousing kits, cable assemblies and accessories.

USB 2.0 and RJ 45 Connector Systems fromCONEC: quality keeps connected!

When the GoingGets Tough.

CONEC IP67 Rated Connectors

USB 2.0 plug and receptacle housing kit, full metal version, M28 thread coupling ring, Partno. 17-200321 (receptacle) 17-200331 (plug)

RJ 45 Cat.5e plug and receptacle housing kit, bayonet coupling, plastic metallized version, Partno. 17-10011 (receptacle) 17-10013 (plug, plastic coupling ring) 17-10044 (plug, metal coupling ring)

RJ45 Cat.5e plug and receptacle housing kit, full metal version, M28 thread coupling ring, Partno. 17-100934 (receptacle) 17-100944 (plug plastic coupling ring)

USB 2.0 plug and receptacle housing kit, bayonet coupling, black plastic version,Partno. 17-200161 (receptacle) 17-200121 (plug)

RJ 45 Cat.5e plug and receptacle housing kit, bayonet coupling, black plastic version, Partno. 17-10000 (receptacle) 17-10001 (plug)

IENYCM2458.indd 1 3/19/09 4:08:15 PM


about to be utilized in radio for analog(voice) transmission, while the cohererwas useless in that case. With the addedinventions of the vacuum rectifier by JohnFleming in 1904 and the triode by Lee deForest in 1906, wireless technology rapidlyevolved, and the coherer fell into oblivion.

BRANLY AND THETELECOMMUNICATIONS INDUSTRY

After his discovery, Branly did not pursuefurther research specifically devoted towireless telecommunications such as try-ing to increase the distance of communi-cation, or identifying the role of the aerialor of wave frequencies [12]. Moreover, hedid not file any patent and freely distribut-ed Branly tubes to all requesting them,concentrating on the improvement of itssensitivity and stability.7 Nevertheless, heparticipated in experiments of long dis-tance wireless transmission, both in 1898at the Eiffel Tower with EugèneDucretet,8 and in 1899 at Ouessant Islandon the Atlantic coast. In 1912 he refusedthe position of technical advisor in one ofMarconi’s companies [12] when Gugliel-mo Marconi himself paid a visit to Bran-ly’s laboratory at Institut Catholique [6].9Both Marconi and Branly were electedmembers of the Academia dei Lincei inRome, Italy (later Pontifical Academy ofSciences), Branly in 1902. However, Bran-ly did briefly participate in the commer-cialization of telecommunications throughthe Société Française de Télégraphes etTéléphones sans Fil created by VictorPopp in 1901 [12]. Unfortunately, in con-trast to Great Britain, the field of telecom-munications was then in France a statemonopoly, and the company had to stopits activities in 1904 despite — or becauseof — successful first deployments.

EDOUARD BRANLY’S ACADEMICACTIVITIES AFTER THE INVENTION OF

THE COHERER

Between 1900 and the second WorldWar, Edouard Branly’s celebrity was high

in France, even if today it is much morelimited, not to mention that he is nearlyunknown in countries outside of France.As an example, he does not appear inthe New Dictionary of Scientific Biogra-phy. The celebrity in France was due tothe fact that Branly was considered oneof the fathers of radio communications.In 1911 he was therefore elected to theFrench Academy of Sciences in competi-tion with Marie Curie (he had previouslybeen a candidate twice). The electionwas a tough one, fought out betweensupporters of both candidates. One ofthe issues under debate was the fact thata woman was running for election. Anacademic clerical group supported Bran-ly, while progressive scientists lobbied forMarie Curie. An unusually large publicmanaged to attend the final session ofthe Academy during which the ballot wasorganized. To prevent incidents, ArmandGautier, who was chairing the session,instructed the ushers to “allow everybodyin… except women of course”! Branlywas elected one ballot ahead of MarieCurie, who never competed again. Thefirst woman to be elected as a full mem-ber was Yvonne Choquet-Bruhat in 1979.

In 1915, during World War I, Branlydevised an optical telegraph operatingin the infrared, allowing transmissionswithout the risk of messages being inter-cepted; experiments with transmissionsup to 20 km were carried out [2].

In 1932, thanks to a grant from thefamous and rich François Coty, whomade his fortune creating and sellingperfumes, a new laboratory for Branly(then 87) was at last inaugurated. Thenew laboratory had been devised byBranly’s son-in-law Paul Tournon, andincorporated special devices such as alarge Faraday cage and stone pillars toeliminate parasitic vibrations. This labo-ratory is today a small museum locatedwithin Institut Catholique de Paris (21rue d’Assas) devoted to Edouard Bran-ly’s research and achievements. Fromthe beginning it was designed for thatpurpose, following Coty’s will.

For a few more years, Edouard Branlypursued his research projects there. In1934 Sara Delano Roosevelt, mother ofFranklin Delano Roosevelt, then Presidentof the United States, visited him. InDecember 1935 he published his last paper[2] about a new design for medical ther-mometers where the mercury-filled glassbulb was replaced by a metallic Ni bulb.

Branly died in 1940 at the age of 96.He was survived by his two daughtersand five granddaughters. The funeralwas celebrated in Notre Dame de Paris,and was attended by the President ofthe French Republic Albert Lebrun.

THE FATE OF THE BRANLY EFFECT

With the absence of subsequent indus-trial applications, the lack of accuracyof the coherer as a scientific instru-ment, and the difficulty of elaboratinga definitive theory despite the compre-hensive experimental work performedby Branly, the Branly effect fell intooblivion. Electrical transport propertieswere mainly studied within solid-statephysics during the first half of the 20thcentury. However, in the 1950s theBranly effect was again utilized inJapan within the first wirelessly operat-ed toy [13].

Several explanations have been sug-gested to explain the Branly effect [7],first at the microscopic scale: electro-static attraction of the grains, electricalbreakdown of the metallic oxide layers,the tunnel effect, and, finally, localwelding of the grains through electro-thermal coupling and melting of microcontacts between grain surfaces. Macro-scopic phenomena were also invoked,such as electrical percolation. New char-acterization methods, such as 1/f elec-trical noise evaluation or infraredobservation of conduction paths, wereused, suggesting new theories [7]. How-ever, difficulties due to the quantitative-ly weak reproducibility of thephenomenon together with the highnumber of influencing parameters suchas aging, temperature, and grain mate-rials and size explain why, at the end ofthe 20th century, the theory of theBranly effect was still to be established[14]. This was despite the fact that, asearly as the 1960s, the emerging physicsof granular materials triggered new sci-entific interest in relation to this oldeffect. Since then, this interest has con-stantly increased, as granular media arepresent in many fields. For instance,mixtures of solid propellant for somerockets, incorporating aluminum parti-cles, may be affected by electromagnet-ic waves created by nearby lightning.10

More recently, and on another scale,nanoelectronics also started using grainsin various structures, either submittedto high currents due to electrostatic dis-charges or exposed to strong electro-magnetic interferences [15].

A recent publication [7] is very con-vincing in demonstrating the key contri-



7 He devised in 1902 a device that was moreefficient than the classical coherer: the trépied.

8 Eugène Ducretet was an industrialist special-izing in the construction of technical equip-ment. The 1898 successful Morse signalingfrom the Eiffel Tower saved it from demolition.

9 Branly and Marconi last met in 1932 duringthe XIth International Telegraphic Conferencein Paris, for which they co-chaired the gala din-ner [2]. Marconi died in 1937. He was survivedthree years by Branly.

10 As early as in 1901 Branly was asked by theFrench Ministry of War to examine the effectsof wireless telecommunications equipment ongunpowder stored in warehouses.

(Continued on page 26)

(Continued from page 22)


bution of local welding to the Branlyeffect. The authors, E. Falcon and B.Castaing, strategically decided to elimi-nate side contributors to the Branlyeffect such as percolation, or the alter-native nature of the applied or inducedvoltage, and to devise an experimentalarrangement where a limited set ofparameters would be involved. Similarto Branly in some experiments, theyused a single chain of steel beads sub-mitted to dc currents, and measured thevoltage drop and thus the resistanceacross various numbers of beads. Thus,they concentrated on electrical trans-port at a microscopic level linked to thedc Branly effect. They observed thatwhile initially highly resistive, above acertain voltage the chain resistivity dra-matically decreases. They also derived atheoretical model of this effect, andthereby both experimentally and theo-retically demonstrated that local weld-ing is one of the contributors to theeffect. It originates from an electrother-mal coupling in the vicinity of the microcontacts between each bead. Thesewelded contacts are, of course, likely tobe broken by a shock. However, with-out a shock, this welding is irreversible,and the chain will exhibit high and near-ly constant conductivity whatever thefuture current levels. Unsurprisingly,the whole process was found to bereproducible.

CONCLUSIONDuring his life, Branly constantly deniedbeing the inventor of radio [11], claimingthat as a scientist he had only discoveredan effect that had permitted others, likePopov, to develop wireless transmission.11

On the other hand, he was one of theearliest promoters of the remote controlof equipment. During a largely publicizeddemonstration on 30 June 1905 at Palaisde Chaillot in Paris he demonstrated thepossibility to remotely and wirelessly turnlight bulbs on and off or activate variouselectromechanical apparati includingmotors.12 He made a publication on thattopic in Compte-rendus à l’Académie desSciences de Paris in March 1905.

What is today called the Branly effectrefers to a set of phenomena some ofwhich were discovered before Branly.However, Branly was undoubtedly thefirst to publish about the effect of a sparkthat is, in modern words, the effect of an

electromagnetic wave. Therefore, throughthe coherer and the demonstration thatradio waves could easily be detected overlong distances, he permitted early devel-opments in wireless telecommunications.Finally and surprisingly, because of bothits own intrinsic complexity and its poten-tial involvement in modern technology,the Branly effect is today still a subject ofinvestigation.

ACKNOWLEDGMENTSThe author is grateful to Professor MischaSchwartz, editor of this column, for sug-gesting the idea of this article and for hisnumerous suggestions for improving thevery first version.

REFERENCESMost of the references below are in French.However, [7] and [16] present in English a dis-cussion about past and recent explanations ofthe Branly effect, and detailed studies of thefirst age of radio. For an account of the privatelife of Branly, the book written by Branly’sdaughter [6] is the essential source.

[1] N. Hulin, “Edouard Branly, la formation d’unphysicien parmi d’autres,” Revue d’histoire dessciences, vol. XLVI , no. 1, Jan.-Mar. 1993, pp.7–26.

[2] P. Monod-Broca, Branly au temps des ondeset des limailles, Belin, 1990.

[3] J.-C. Boudenot, Comment Branly a découvertla radio, EDP Sciences Editions, 2005.

[4] E. Branly, “Variation de la conductibilité sousdiverses influences électriques,” Compte-ren-dus à l’Académie des Sciences de Paris, vol.111, 1890, pp. 785–87.

[5] D. T. Emerson, “The Stage Is Set: Develop-ments before 1900 Leading to PracticalWireless Communication,” IEEE GLOBECOM,San Antonio, TX, 2001.

[6] J. Terrat-Branly, Mon père, Edouard Branly,Corrêa, 1941.

[7] E. Falcon and B. Castaing, “Electrical Con-ductivity in Granular Media and Branly’sCoherer: A Simple Experiment,” Amer. J.Physics, vol. 73, no. 4, Apr. 2005, pp.302–06.

[8] D. E. Hughes, “Prof. D. E. Hughes’s Researchin Wireless Telegraphy,” The Electrician, May5, 1899, pp. 40–41.

[9] A. Kleinert, “Ferdinand Braun et les débuts de laTSF en Allemagne,” Revue d’histoire des sci-ences, vol. XLVI, no. 1, Jan.-Mar. 1993, pp.59–71.

[10] D. Pestre, “Contrepoint autour d’EdouardBranly: trois propositions de révision,” Revued’histoire des sciences, vol. XLVI, no. 1, Jan.-Mar. 1993, pp 83-93.

[11] C. Blondel, P. Carré, and N. Hulin, Editorial,Revue d’histoire des sciences, vol. XLVI, no.1, Jan.-Mar. 1993.

[12] C. Blondel, “Branly face à l’innovation tech-nique : un cas d’espèce?,” Revue d’histoiredes sciences, vol. XLVI, no. 1, Jan.-Mar.1993, pp. 39–58.

[13] E. Falcon and B. Castaing, “L’effet Branlylivre ses secrets,” Pour la Science, no. 340,Feb. 2006, pp. 59–64.

[14] Jean Cazenobe, postscript of [2].[15] Saint Macary L. et al., “Size Effects on

Varistor Properties Made from Zinc OxideNanoparticles by Spark Plasma Sintering atLow Temperature,” Advanced FunctionalMaterials, vol. 19, 2009, pp. 1–9.

[16] H. Aitken, Syntony and Spark — The Originof Radio, Wiley, 1976.

BIOGRAPHYJEAN-MARIE DILHAC [SM] is a professor at InstitutNational des Sciences Appliquées of ToulouseUniversity, where he served as head of theDepartment of Electrical and Computer Engineer-ing (2004–2008). His main teaching fields arerelated to electronics, signal processing, andtelecommunications. A side activity is devoted tothe history of computing and telecommunica-tions. He also serves as a member of the IEEECommunications Society’s History Committee. Heis a senior scientist at LAAS-CNRS research labo-ratory. His present research activities are relatedto energy management in wireless sensor net-works applied to either structural health moni-toring or in-flight testing for aeronauticalapplications. He is in charge of many collabora-tive research programs in connection with Air-bus.


Comments on “Partial-ResponseCoding …,” by H. Kobayashi

G. David Forney, Jr., MassachusettsInstitute of Technology ([email protected])

Thank you for the interesting andauthoritative article by HisashiKobayashi [1] on the history of partial-response, maximum-likelihood (PRML)technology, which became an industrystandard for magnetic recording in the1990s. To expand upon that account,your readers might like to know that, asfar as I am aware, the first commercialimplementation of a version of PRMLwas in 1969 in the Codex AE-96, a sin-gle-sideband 9600 b/s telephone-linemodem that used 1 – D2 partial-response signaling. This technique wasdescribed in a 1971 patent [2], and inthe last section of my 1972 paper [3, pp.373–75]. Practically, it gained about 3dB, and extended the product life of theAE-96 by several years. Theoretically, itled me to think about maximum-likeli-hood sequence detection for generalintersymbol-interference channels,which eventually led to [3].

REFERENCES[1] H. Kobayashi, “Partial-Response Coding, Max-

imum-Likelihood Decoding: Capitalizing onthe Analogy between Communication andRecording,” IEEE Commun. Mag., vol. 47, no.3, Mar. 2009, pp. 14–17.

[2] G. D. Forney, Jr., “Error Correction in Sam-p led-Data Circuits,” U.S. Patent no.3,613,077, Oct. 12, 1971.

[3] G. D. Forney, Jr., “Maximum-LikelihoodSequence Estimation of Digital Sequences inthe Presence of Intersymbol Interference,”IEEE Trans. Info. Theory, vol. IT-18, May1972, pp. 363–78.

LETTER TO THE EDITOR


11 The coherer was the first solid-state deviceused in electronics [16].

12 5000 spectators were present [2].

(Continued from page 24)


history of communications - comsoc.org · ieee communications magazine • september 2009 21 and a...

Documents