Armstrong v. De Forest Radio Telephone & Telegraph Co.

MAYER, District Judge.

This is a suit of major importance. It concerns an invention of high merit, and the cause has been presented *446ably and comprehensively. While the record is voluminous, it differs from some long records in that, by reason of the issues of fact involved, its length is fully justified. The defenses are many, but the principal attacks are directed against- the priority of Armstrong. It is claimed for Armstrong that his date of invention is at least as early as January 31, 1913, and thus antedates Schloemilch and Von Bronk, infra, Meissner, infra, and De Forest, infra. Before discussing the •questions involved in the priority contest, it is desirable to ascertain ■what the patent is and what is its accomplishment.

At the outset, it should be stated that the Armstrong feed-back circuit, as it has come to be known familiarly, must be recognized as a contribution of marked value to the practical art. Its employment has so greatly increased both the loudness and the definition of the sounds •heard in the receiver that long-distance radio communication has been •remarkably improved and thus greater reliability has been attained.

“The present invention,” Armstrong stated in his specification, “relates- to improvements in the arrangement and connections of electrical apparatus at the receiving station of a wireless system, and particularly a system of this kind in which a so-called ‘audion’ is used as the Hertzian wave detector; the object being to amplify the effect of tire ¡received waves upon the current in the telephone or other receiving •circuit, to increase the loudness and definition of the sounds in the telephone or other receiver, whereby more reliable communication may be established, or a great distance of transmission becomes possible. To this end I have modified and improved upon the arrangement of the receiving circuits in a manner which will appear fully from the following description, taken in connection with the accompanying drawings. As a preliminary, it is to be noted that my improved arrangement corresponds with the ordinary arrangement of circuits in connection with an audion detector to the extent that it comprises two interlinked cir■cuits, a tuned receiving circuit in which the audion grid is included, and ■which will be_ hereinafter referred to as the ‘tuned grid circuit,’ and a circuit including a battery or other source of direct current and the ‘wing’ of the audion, and which will be hereinafter referred to as the |wing circuit.’ As is usual, the tw'o circuits are interlinked by connecting the hot filament of the audion to the point of junction of the tuned .grid circuit and the wing circuit. I depart, however, from the customary .arrangement of these circuits in a manner which may, for convenience of description, be classified by analysis under three heads: Firstly, the provision of means or the arrangement of the apparatus, to impart ■resonance to the wing circuit, so that it is capable of sustaining oscillations corresponding to the oscillations in the tuned grid circuit; secondly, the provision of means supplementing the electrostatic coupling of the audion to facilitate the transfer of enerp- from the wing circuit .to the grid circuit, thereby reinforcing the high frequency oscillations in the grid circuit; and, thirdly,- the introduction into the wing circuit of an inductance through which the direct current of the wing circuit ■flows, and which is so related to the grid circuit that the maintaining ■electromotive force across the terminals of the inductance, due to reduction of the direct current, is effective in the tuned’grid circuit to *447increase the grid charge, and consequently to further reduce the current in the wing circuit and in the telephones.”

The “firstly” and “thirdly,” supra, were in the original specification; the “secondly” was inserted during the prosecution of the patent application. The first statement is illustrated in the drawings of the patent, Figs. 1, 2, 3, and 6. An illustrative claim is No. 1, which specifies, inter alia, “a resonant wing circuit” and reads as follows:

“1. An auction wireless receiving system having a resonant wing .circuit interlinked with a resonant grid circuit upon which the received oscillations are impressed, the resonant grid circuit having a capacity so related to the grid as to receive and retain the charge which accumulates thereon.”

The second statement, which defines broadly the instrumentalities of the. first and third statements, is represented by each of the figures of the patent drawings and illustrated by claim 9, which reads as follows:

“9. An audion wireless receiving system having a wing circuit interlinked with a resonant grid circuit upon which the received oscillations are impressed, and an inductance through which the current in the wing circuit flows, the grid circuit including connections for making effective upon that circuit the potential variations resulting from a change of current in the wing circuit.”

The third statement is illustrated in the drawings of the patent, Figs. 1, 2, 4, 5, and 6. In one form of language or another, what is set forth in the third statement is embodied in claims 3, 5, 8, 12, 14, and 17. Claim 17 will suffice for illustration:

“17. An audion wireless receiving system having a wing circuit interlinked with a resonant grid circuit upon which received oscillations are impressed and an electrostatic coupling between the circuits supplementing the coupling of the audion to facilitate transfer of energy from the wing circuit to the grid circuit, whereby the effect upon the grid of high frequency pulsations in the wing circuit is increased.”

For so difficult a subject-matter, the specification and claims as originally filed fared very well in the Patent Office. The first 14 claims were allowed as filed, and they constitute all of the original claims, except one originally numbered 13, which was rejected on reference-to the Schloemilch and Von Bronk patent. The amendments to the specification (page 1, lines 56-60; page 2, lines 47-54; page 3, lines 33-47) and the amendment to the claims by way of addition of claims 15-18, inclusive, were concerned only with statement of the mode of operation, and did not add to the instrumentality described and claimed in the original application and drawings. Out of the mass of testimony and argument, too extensive to quote or to discuss in complete detail, it is well to settle one proposition at the start.

The Armstrong specification and claims show that the invention was for an instrumentality. The feed-back circuit was well defined in the record on a number of occasions. Professor Hazeltine, plaintiff’s expert, stated that the fundamental principle of Armstrong’s invention was:

“The provision of an arrangement for transferring oscillating current energy from the plate circuit to the grid circuit whereby oscillations present in the grid circuit are assisted.” Answer to XQ39.
*448“Any arrangement by which oscillating current energy is transferred from tbe output or plate circuit of tbe audion to the input or grid circuit to sustain the oscillations in the grid circuit is included in the principle of the Armstrong invention.”' Answer to .XQ41.

Defendant asserts that the question is whether the invention resides in the reamplifying audion or in the oscillating audion or in both or in “some more basic idea,” and then contends that the patent is limited by its own terms to the reamplifying audion.

The amicus curise, through its counsel, put forward substantially the same contention, urging that the Armstrong disclosure was nothing more than the use of the audion for reamplifying purposes.

[1] All of these arguments and all the analysis of the Armstrong patent language and claims come down to a single proposition: If Armstrong invented a new instrumentality, he is entitled to the fruits of all its uses, whether he understood them or not, and whether his theory of operation was right or wrong, comprehensive or limited. Given the new instrumentality, the question is what it does, not how or why it does something. It is urged that the error of the argument for Armstrong is—

“primarily in tbe assumption that the invention is a specific recognizable thing called a ‘feed-back,’ a definite group of mechanical or electrical elements which like a tool may be used for various purposes. That is not the invention. The invention is * . * * a particular use of an otherwise old circuit by, an adjustment of the constants so as to produce reamplification.”

The discussion of the limitations upon the patent by its own terms includes quotations from the patent at those places where Armstrong pointed out that, if the ratio of feed-back coupling exceeded a certain amount, the audion would become a high frequency generator, setting up disturbing oscillations in the grid and wing circuits, and informed the art how maximum amplification of damped wave signals could be obtained below oscillation. See particularly page 4, lines 51-77. But, the fact is that defendant’s oscillating audion does regeneratively feedback energy from the plate circuit to the grid circuit to amplify cumulatively the received signals. This seems satisfactorily shown by .Hazeltine, by Dr. Austin’s paper, and by Armstrong’s paper, Exhibit 20, and his testimony in the Interference Record. The patent does not indicate any use of the audion in the oscillating condition; but, when the audi-on is oscillating, due to feed back, it is coincidently regenerating. XQ87 and answer.

Leaving, now, the construction of the patent as it is written, and the action and results which occur from following the teaching of the patent, it is next desirable to ascertain Armstrong’s date of invention.

[2,3] This case is another contribution to the-romance which has so often characterized the history of forward inventions. Armstrong was graduated at about 22 years of age from Columbia University in June, 1913, as an electrical engineer. As a boy of 15 he became interested in radio and erected a radio station at his home. He obtained his first vacuum tube detector, a so-called Fleming valve, in 1908, from Mr. Charles R. Underhill. Later, about 1910 or 1911, he obtained a so-called De Forest audion or three-element vacuum tube detector. .In the spring of 1912 he began a close study of the funda*449mental action of the audion and read all the literature on the subject. Some time during this period, he connected a condenser across the telephone of a simple audion receiving system and noticed that on some bulbs an increase in signal strength would result. It is important, at this point, to realize that Armstrong is a remarkably clear thinker. His modest demeanor belies his extraordinary ability. His achievement was not the result of an accident, but the consummation of a thoughtful and imaginative mind.

Step by step he proceeded with study and experiment from the summer of 1912 until the fall of that year. The apparatus which he used— the long-wave apparatus and short-wave apparatus — are in evidence. He was obtaining what seemed to him remarkable results, and he showed his apparatus to his father, and let him listen to his signals, and asked his father to advance money for a patent application. Mr. Armstrong, Sr., was evidently not impressed and refused his son financial help. On December 7, Í912, Armstrong told his college mate, Burgi, that he had succeeded in improving the sensitiveness of the audion by means of a new connection, and Burgi noted this in his diary on that date as follows:

“Armstrong told me lie liad a connection for intensifying sound.”

About the end of 1912, lie explained to Mr. Thomas Ewing that he wished to make tests with a new receiver he had invented, and obtained Mr. Ewing’s permission to string the-antenna (Armstrong having erected a mast at his home) into Mr. Ewing’s adjoining property.

About the same time, Armstrong showed the receiver to his uncle, Mr. Smith, and tried to get him to advance money for the patent application. This Mr. Smith did not do, but advised Armstrong to make a drawing of the connections and have it witnessed by a notary. During the fall of 1912 and the winter of 1912-1913, Armstrong showed his receiving results to a number of radio amateurs without disclosing the circuits. The description by the amateurs of the result corresponds with what the court heard at the trial during the demonstrations at Columbia University. The amateurs testified that they knew now (i. e., when they testified) that this result must have been produced by the Armstrong feed-back. (Styles; Royce; Russell; Shaughnessy.) Wallace noted the similarity of the hiss to heterodyne reception. The testimony of the amateurs is impressive as to its truth and as to accuracy of recollection.

On January 31, 1913, Armstrong went with his college mate, Burgi, to Goodwin’s Real Estate Office at 123d street and Eenox avenue with a drawing on tracing doth of his circuit, connections and had the drawing witnessed by the notary. The drawing is Armstrong Exhibit E in the interference and Plaintiff’s Exhibit No. 37 in this suit. It was shown to Armstrong’s uncle, Mr. Smith, at about the time it was made, although Mr. Smith did not have such knowledge of yyireless apparatus or systems as would enable him to understand the circuits. Burgi testified that he might have understood the circuits at the time, but would not be able to reproduce them from memory; but he identified the sketch as the one shown to him and witnessed by the notary in his presence. This drawing is reproduced in the patent as Fig. 2, and it *450includes both the adjustable tuning inductance in the plate circuit and the telephones located in the common path.

I have no doubt whatever, either as^to the existence of the apparatus on January 31, 1913, which contained Armstrong’s feed-back circuit arrangement and operated practically and successfully, or of the accuracy of the date on Exhibit 37 of January 31, 1913. That drawing was a record of an invention then completed and reduced to practice. Armstrong had one objective upon which he concentrated all there was in him of ability, of perseverance, and of courage against what were great odds to a young man. That no one of the witnesses saw or understood the circuits is, in the circumstances of this case, immaterial and unimportant. I see no legal obstacle to fixing January 31, 1913, as at least the date of the invention, and I am so completely satisfied with the evidence that I am not embarrassed by any question of fact. The letter of February 17, 1913, to Underhill would be helpful as corroborative if needed.

' [4] After examination of the authorities cited, I am unable to see how the Mason deposition can be received in evidence in this case. Defendant has not had any opportunity in this or any other proceeding to examine Mason, and his deposition, therefore, in this case, would be nothing more than an ex parte affidavit. If the rule is to be different, it must be so declared by an appellate court, and not by the court of first instance.

[5] But Mason’s deposition is hot necessary to plaintiffs’ success in this case. On questions of invention date, much is said about corroboration ; but I do not undertand that corroboration, in the sense of full knowledge by a witness of the inventive conception and an understanding of the apparatus, in addition to the inventor, is necessary. The question always is as to what will satisfy the trier of the facts; and, even in criminal cases (except where there is some statutory requirement to the contrary), the testimony of one witness uncorroborated is sufficient to establish guilt beyond a reasonable doubt, if believed by the jury — i. e., the trier of the facts. In the case at bar, there is not only Armstrong’s testimony, to which I accord absolute credibility, but in addition the apparatus and the sketch — physical things — the testimony of the amateurs, and all the surrounding circumstances leave no doubt that Armstrong’s date is at least as early as January 31, 1913.

Considering the difficulties with which the young student was confronted, he moved diligently, and finally succeeded in filing his application for the patent on October 29, 1913. The merit of the invention was soon recognized, and the very apparatus with which Armstrong made the invention was subsequently utilized commercially at Sayville, Long Island, shortly after the outbreak of the war in 1914, to overcome difficulties in the reception of signals from Ñauen, Germany, and this apparatus continued in use at Sayville for some time. As indicative of the appreciation of the commercial value of the invention, licenses were taken at an early date by the Atlantic Communication Company, the Goldschmidt Company and the Marconi Company.

The Meissner United States application was filed on March 16, 1914. The oath recites four German applications, the earliest of which *451was dated April 9, 1913. This application (now German patent No. 291,604, dated June 23, 1919J discloses the regenerative or feed-back circuit adjusted to produce oscillations; but it does not state that it is for radio reception, and the antenna and ground connections are omitted. It is unnecessary, however, to discuss this patent in detail, because by according to Armstrong the date of January 31, 1913, he preceded Meissner by about 2% months

Schloemilch and Von Bronk, United States patent No. 1,087,892, issued February 17, 1914, was filed on March 14, 1913, subsequent to Armstrong’s date; but, in any event, it would not have embarrassed Armstrong, for it did not occur to Schloemilch and Von Bronk to re-amplify the high frequency variations by feeding them back to the grid circuit. In other words, this patent failed to point out, and its authors evidently did not appreciate, the essential point of Armstrong’s invention.

The Reisz patent, No. 1,234,489, filed April 9, 1913, and issued July 24, 1917, does not cia.ll for discussion.

We thus come to the De Forest defense. It is no reflection on De Forest to observe that in a number of instances in this record his memory has proven faulty. This is not strange. During the years 1912, 1913, and 1914 De Forest was confronted with many difficulties and many obstacles. He had much to contend with, and, besides, he was experimenting in numerous directions, and, in such circumstances, independent memory is not always reliable.

The work with the audion from 1912 to 1914 had to do primarily with two uses of the audion: (1) As a telephone relay; and (2) as an amplifier of telephone and detected radio signals and the use of the amplifier to record radio signals on a telegraphone wire. The “howling” of an audion oscillating at audio frequencies would defeat both of these purposes, for the obvious reason that the noise of the. “howling” audion would destroy telephone conversation and radio signals alike. On the other hand, the feed-back circuit would have been-of great service in the amplification of radio signals and in recording them on a telegraphone wire.

The documentary evidence of the work of De Forest and his associates begins with the entry in the De Forest experimental notebook (Defendant’s Fxhibit H under date of June 21, 1912J, where he enters the observation of a beat or high frequency note with the straight audion hook-up. His notes show this to have been transient and incapable of reproduction, and he recognized that it was not the true heterodyne effect. De Forest Ct. Rec. pp. 184, 185. The entry of June 21, 1912, so far as it is pertinent to the present inquiry, shows that De Forest did not then know how to produce a beat note with an audion receiver.

The next entry is in the Van Ftten notebook under date of August 6, 1912. Van Ftten was working on the audion as a telephone relay ánd amplifier. He had looked up the two-way telephone relay in Kempster B, Miller’s well-known work on the subject of telephone engineering, and had learned that a mechanical telephone relay, when used in a two-way circuit, would sing and howl, unless precautions were taken to prevent it. De F. Int. Rec. pp. 60, 77, and 78.

*452On July 23, he entered in this notebook a two-way telephone circuit, using two audions under the heading, “Think the following arrangement (doped out yesterda)'-) would possibly make a good telephone repeater.” It happened, as De Forest testifies, that they did not have two audions available, and on July 24th Van Etten attempted to set up this two-way circuit of his with a single audion having two grids and_ two plates. The arrangement did not work, because, as Van Etten said in an entry dated August 5th:

“Above n. g. because audion does not get in the game — simply two repeater coils in series — see inked ckt.”

The following day, August 6th, he connected the_ input circuit of a double audion to the output circuit, and in this accidental way found that the audion would howl or sing. He pointed out in his, notes that the arrangement gave a beautiful clear tone, which lowered and raised in pitch as the B-battery was varied, and could be wiped out by a magnet, “and then the watch ticks came through as before.” He says r

“This phenomenon is apparently similar to the ‘howl’ produced in an ordinary O B telephone when the receiver is placed against the transmitter, but nevertheless, as shown by variation of number of cells in battery B and by magnetic wiper, is also intimately associated with the audion.”

On the same day he set up a balanced two-way telephone circuit, corresponding to the two-way balanced telephone circuit shown in Miller, which is used to prevent the howling of mechanical telephone relays, and m'ade the entry:

“Tried following circuit, but it was n. g.; could not make it boost at all; could only make it howl.”

From this point on his notes show a continuing attempt to produce a telephone relay circuit in which the audion would not howl, and in this effort he succeeded on or about August 29, 1912. The August 6, 1912, entry of Van Etten was not.copied in the laboratory notebook, nor was anything done which showed that any one appreciated the phenomenon. De Forest’s testimony on the point is not sufficiently clear and definite to be satisfactory.

The entry of February 7, 1914 (Exhibit D, blue page 26), shows that where De Forest duplicated the arrangement he referred to it as “this new and simplified circuit,” and regarded it as useful only for the “trigger effect.” In De Forest’s article in the Electrical World on February 20, 1915, no mention is made by De Forest of the August 6, 1912, entry of Van Etten.

On August 29th, after entering in his own notebook the two-way circuit which did not howl, Van Etten made an entry showing the use of two audion amplifiers in cascade(. He explained this arrangement to De Forest; it was considered of great importance, and was copied into De Forest’s notebook by Van Etten under .that date. This is the invention De Forest afterward sold to the A. T. & T., and about which he said:

“When I made tbis interesting discovery, I lost interest in the oscillating feature.”

In the Van Etten notebook, it is described merely as an amplifier; but in the De Forest notebook it is noted that by reversing certain con*453necfions the arrangement makes all sorts of musical notes in the telephones, and that the notes could be changed by putting very small capacities, such as the capacity of one’s body, or from thumb to finger; between ground and grid, wing, or filament.

It is testified by De Forest that these effects were due to a feedback brought about by the fact that the coils numbered 1 and 5 oñ the sketch were placed close together on the table, although there was no indication of that in the sketch or notes. De F. Int. Rec. p. 128. Van Etten fails to corroborate him as to the existence of any feed-back in this circuit. De F. Int. Rec. p. 74.

Although there was in fact an accidental capacity feed-back between the output circuit of the second audion and the input circuit of the first audion in this cascade system, that coupling was not apparent in the diagram, was not mentioned in the notes, and was not understood by De Forest. The arrangement did not indicate to De Forest or Van Etten how to control the • oscillations, or that they were due to any circuit arrangements, and not due to some peculiar and obscure characteristic of the audion.

Without indulging in further detail, it is enough to say that the proof falls far short of any conviction that in 1912 either De Forest or Van Etten had any realization of what is now the invention of the patent in suit. The first application by De Forest of the oscillating audion to any useful purpose was in a laboratory experiment on April 17, 1913. The entry on that date beginning, “This day I got the long looked for beat note,” indicates that; theretofore, De Forest had not been able to get this heterodyne phenomenon. There was no sketch of the circuit arrangement, nor was the circuit disclosed to any one, scientist or layman.

It is argued that, for several' reasons, the circuit described in the note was not a feed-back circuit (plaintiffs’ brief, p. 69 et seq.); but, whichever way the entry may be construed, it is at least open to debate, and, in view of the lack of corroborating circumstances, a debatable entry, not clearly illustrated and not disclosed to any one at the time, leaves the matter in doubt. The De Forest notebook entries of 1914 need not be analyzed, in view of Armstrong’s date. It was not until September, 1915, that De Forest filed any application showing what he claimed to be a feed-back circuit, and in the meantime he had filed 30 patent applications.

Armstrong’s counsel do not contend that De Forest’s lack of knowledge that the ultraudion circuit was a feed-back circuit would deprive him of whatever benefit might accrue to him from the invention of that circuit in the spring of 1914; but it is their contention that the invention of the ultraudion circuit in February, 1914, the failure to recognize it as a feed-back circuit, and the filing of the ultraudion applications without including any circuit of the 1912 notes, completely negative the idea that he had invented the feed-back circuit in 1912. With this view I agree.

It is not practicable to discuss the testimony in all its elaborate detail. All has been considered, although some has not been mentioned, such as that of Logwood, which needs no serious comment. On the one side is an enthusiastic, never-say-die young student — feed-back Arm*454strong — with but one thought possessing him. He not only discloses to many persons his belief that he has invented something worth while, but he produces his apparatus, and he produces a sketch which is_ extraordinary for its clear and unmistakable description to one skilled in the art, and the date of that sketch is incontrovertibly fixed.

On the other side is a then experienced and able worker in the art, experimenting along certain lines, who is unable to rely solely on notebook entries, which are not clear, but require construing, and who supplements these entries by recollection which is fallible and not certain. If De Forest, in 1912 or 1913, invented the feed-back circuit, the obvious financial reward in store for him would have induced him, notwithstanding all his difficulties, to do one of two things: (1) That which Armstrong did, that is, in some way make a clear memorandum and have somebody know about it; or (2) file an application for a patent in the same way that during the period concerned he filed many other applications.

Holding then, that Armstrong, is the first inventor, and that his claims in suit «should be construed as he contends, the sole remaining question is that of infringement. It will suffice to say that on this branch of the case I accept Hazeltine’s testimony (Court Record, pp. 41-59), from which it appears that all the claims are infringed.

Plaintiffs may have a decree in accordance herewith, with costs.