August 24, 1926.
Charles Neave, Stephen H. Philbin and Abel E. Blackmar, Jr., all of New York City, and Harry E. Dunham, of Schenectady, N.Y., for plaintiffs.
Clifton V. Edwards and Lawrence K. Sager, both of New York City, and A.D.T. Libby, of Newark, N.J., for defendant.
Walter G. Winne, U.S. Atty., of Hackensack, N.J., Herman J. Galloway, Asst. Atty. Gen., and Harry E. Knight, Sp. Asst. Atty. Gen., amicus curiæ.
In Equity. Suit by the Radio Corporation of America and others against the Splitdorf Electrical Company. Decree for plaintiffs.
The patent in suit is United States letters patent No. 1,173,079, to E.F.W. Alexanderson, assignor to the General Electric Company. The specifications in part state:
"The present invention relates to the selection of oscillations of a given wave length from mixed oscillations, and comprises systems suitable for tuning out interferences in radio-telegraphy. * * *
"In accordance with the present invention, selective tuning is secured by the use of a plurality of resonant circuits arranged in cascade in such a manner that the selectivity of the system increases in geometric ratio with the number of circuits employed. The selective circuits are respectively interlinked by a relay controlling a separate source of energy to initiate oscillations corresponding to potential oscillations impressed upon the relay. As each tuned circuit is more or less opaque to disturbing oscillations differing in frequency from the oscillations to be selected, a certain percentage of the disturbances is eliminated in each circuit of the series, so that the purity of the incoming train of oscillations progressively increases as it is successively relayed. The relay preferably used for this purpose is an electron discharge tube having an incandescent cathode, an anode and a grid."
The claims in suit are 1, 2, 3, 6, 7, and 9 to 12, inclusive. For convenience, the claims fall into two groups 1, 2, 9, and 12, and 3, 6, 7, 10, and 11. It was conceded by the defendant that 1, 2, 9, and 12 read upon its device. It was also conceded at the argument that claim 3 was substantially like claim 2. Counsel for the plaintiffs was content that the court should withdraw from consideration, without prejudice, claims 6, 7, 10, and 11. Claims 1, 2, 3, 9, and 12 will be considered.
The claims are as follows:
"1. The method of selecting sustained oscillations of a given frequency from disturbing oscillations differing therefrom in frequency which consists in impressing all the oscillations upon a circuit, resonant to the frequency of the oscillations to be selected, thereby reducing the effect of disturbing oscillations in accordance with the degree of tuning of the resonant circuit, and controlling by means of the oscillations in said circuit an independent source of energy to initiate oscillations in step therewith and impressing the second set of oscillations upon a second circuit resonant to the frequency of the oscillations to be selected.
"2. A receiving apparatus for electromagnetic waves comprising a plurality of tuned circuits largely opaque to oscillations of other than a given frequency, means linking adjoining circuits, said means comprising a source of energy and an energy transmitting apparatus varying in conductivity with impressed oscillations for initiating oscillations in step with received oscillations and means associated with the last circuit of the series for detecting the oscillations.
"3. A tuned receiving system for detecting sustained oscillations of a given frequency comprising a plurality of circuits resonant to the frequency of the oscillations to be detected and arranged in cascade, relay devices joining each of said circuits to another comprising an evacuated envelope, an electron-emitting cathode, a cooperating anode, and a grid, said device being connected to one of said circuits at the cathode and grid and to another circuit at the cathode and anode and a local source of energy in the second circuit."
"9. The combination of an electrical discharge device comprising an envelope, a cathode adapted to emit negative charges, an anode and means for controlling a discharge between said electrodes, a circuit electrically resonant to variable current of a given frequency, electrical connections between said circuit and the discharge controlling means of said device, a second circuit resonant to variable current of the same frequency, a source of electrical energy in circuit with the electrodes of said device, and means for impressing a variable current initiated by said source of energy upon the second resonant circuit."
"12. A system for selecting electrical oscillations of a given frequency from oscillations differing therefrom in frequency comprising a series of resonant circuits largely opaque to oscillations of other than the given frequency, and means linking adjoining circuits, said means comprising a source of electric energy and an energy-transmitting apparatus varying in conductivity in accordance with oscillations impressed thereon from one circuit whereby oscillations are set up by said source of energy in another linked resonant circuit of said series having the same frequency as the impressed oscillations."
The questions for determination are, first, Is the patent invalid for lack of invention? and, second, Is the patent completely anticipated by the prior art? These questions are so closely related that they must be considered together.
Broadly, the patent provides, in the radio art, for an arrangement of parts in such a way that the tuning is done in geometrical progression or by successive filtration, with audion tubes used in the way specified in the patent. Somewhat more specifically, Alexanderson repeats the signal currents tuned in the grid circuit into the plate circuit, by means of an audion system, and then selects in another tuned circuit, thereafter detecting. This selection by repetition may be in two tuned circuits so connected, or in three tuned circuits, separated by audions, as in Fig. 2 of the patent, shown below.
The invention may be better understood if some slight reference be made to the wireless art. Through the air electromagnetic waves are transmitted. These rapidly oscillatory waves are similar to light and have the same velocity. To receive them a metallic conductor is erected which intercepts the waves from all stations passing over a given region. The function of the receiving system is to render intelligible to the senses the waves received, and also to separate the waves from different stations. This is the problem which the patent in suit meets, and Alexanderson was the first to do it.
The method of the patent is to receive the signal in a circuit of practical and operative sharpness and to repeat the signal and again select, and so on, repeating and selecting as many times as may be necessary to obtain the desired result.
Mr. Frank Waterman, the well known engineer, stated the operation of the Alexanderson system as follows:
"In the Alexanderson patent the idea of transference of the received energy from one circuit to another is entirely abandoned. Alexanderson proceeds on the theory that the thing necessary to obtain the result is that the received signal should be selected in a circuit of such selectivity, in accordance with the then knowledge of the art, as was suitable and convenient for the purpose, that that signal should then be repeated without abstracting energy from the circuit in which the first selection occurs, and that the so selected product of the first circuit should be again submitted to the selective action, the energy of the signal so submitted the second time being not the energy of the first circuit which was derived from the air, or from the original transmitted signal, but energy derived from a local battery, so that it was, in fact, a repetition, and so that the disadvantages which come from the transfer from circuit to circuit need not have to be encountered and overcome. The difference is fundamental, and the difference in result very marked indeed."
Mr. Waterman's oscillograms clearly show the accomplishment of the Alexanderson patent as well as the unsatisfactory results obtained by the Stone method, admittedly the best reference.
Fig. 2 of the Alexanderson patent, here shown below, illustrates the operation of the various circuits described by him. It is also illustrative of the defendant's receiver circuits but for some immaterial details:
Fig. 2 herewith attached:
The radio waves passing over the region and broadcast from many stations are intercepted by the antenna ( 1). The radio frequency currents are then transferred to the grid circuit of tube 1 by means of the transformer coupling 2. The grid circuit is tuned by the secondary coil of the transformer and by the condenser to the right. The radio frequency currents in the grid circuit are then repeated by the action of the tube and its circuits, and appear in the plate circuit. They are then transferred to the grid circuit of the next tube by means of the transformer coupling 12, and so on through the system. The system permits the receiving apparatus to select a faint signal, even though powerful signals are upon the air, and to build it up to the desired signal by resonance; that is, the increase in energy by successive applications of a force recurring at suitable and intermitting intervals.
Alexanderson does not pass received energy from circuit to circuit, but he repeats his signal variations from one circuit to each succeeding circuit without abstracting energy, and obtains selectivity by the repetition. Alexanderson in his first circuit selects the desired signal; that signal is repeated to the next circuit where undesired currents are further filtered out, and so on. By repetition, without loss of energy, further and greater selectivity is obtained.
Mr. Alexanderson is the consulting engineer of the General Electric Company. For years he has been engaged in electrical engineering research. In the latter part of 1912 he visited Mr. John Hays Hammond, Jr., at Gloucester, and discussed with him problems of selectivity in the radio art. He was given by Mr. Hammond an audion some time in February of 1913. Hammond's system of selectivity consisted in radiating a wave of special character so that only a receiver adapted for that particular combination of waves could receive the signals.
While Mr. Alexanderson was visiting Mr. Hammond and discussing the problems of selectivity, he conceived his invention. Returning to the laboratory of the General Electric Company, he set his assistants to work to check his conclusions. There is no doubt that by early January, 1913, the invention was in existence. Mr. Alexanderson is corroborated by Dr. Langmuir, the eminent scientist of the General Electric Company, by the laboratory assistants' letters, notebooks and drawings. Plaintiffs have clearly established a date at least as early as February 4, 1913.
It has never been said that it is necessary for an inventor to broadcast his invention in order to establish a date. It is enough that testimony and documents obviously genuine be offered for this purpose.
The audion was somewhat new at the time of the Alexanderson discovery, but placing the audion in Alexanderson's circuit arrangement was certainly novel and required far more than skill in the art. Tuned circuits were, of course, old, but Alexanderson so assembled the instrumentalities of the art as to attain a new result. He made a selective arrangement which possessed great advantages over anything theretofore known.
The Stone patent, No. 714,756, is undoubtedly the best of the defendant's many references. Stone used tuned circuits connected with each other, but the energy applied to the first circuit was transferred to the second, and so on. If the coupling between the circuits is adjusted so as to obtain selectivity, the desired current becomes so attenuated as to give an unsatisfactory signal. The defendant's expert firmly believes that the Stone reference is a complete anticipation. Mr. Waterman does not. The physical structure of the Stone patent is different, and the results obtained were clearly shown by the oscillograms to be different. Although ample opportunity was offered to examine the method used in making those graphic representations, the defendant did not avail itself thereof. The Stone patent was free to the public. The defendant admittedly does not follow its construction, but does follow the Alexanderson patent. If Stone's invention is the same thing the defendant may use it. It seems extraordinary that it prefers the Alexanderson circuit if it has no novelty. It gives the tribute of its praise to the Stone patent, but imitates Alexanderson. Stone transferred the current from one tuned circuit to another by a transformer. Alexanderson repeats the signal from the first tuned circuit to the second tuned circuit, and so on. His tuned circuits do not react upon one another, and the selective efficiency is built up from circuit to circuit in geometric progression. Any number of circuits may be used with increasing efficiency, but if the Stone circuits are increased in number a point will soon be reached where there is a complete loss of signal.
Of the prior art references other than Stone, the Marconi patent, No. 763,772, was mentioned. The inventor states in his specifications:
"My invention relates to apparatus for communicating electrical signals without wires and by means of Hertz oscillations or electric waves; and the object of the invention is to increase the efficiency of the system and to provide new and simple means whereby oscillations or electric waves from a transmitting station may be localized when desired at any one selected receiving station or stations out of a group of several receiving stations."
Granting that the patent discloses the advantage of obtaining selectivity by tuning successive circuits to the same frequency, the current was passed in the same way and with the same disadvantages of Stone's method. Alexanderson passed his currents through by means of the audions — unidirectional couplings. His arrangement of circuits obtained the better results.
The De Forest patent, No. 879,532, discloses an audion system used as a detector. It does not show a circuit capable of doing the work of the patent in suit. The De Forest patent, No. 1,375,447, effective August, 1912, antedates the patent in suit. Mr. Waterman said of this patent, and his statement was unchallenged:
"The De Forest patent inquired about relates to a method and apparatus for reproducing sounds from magnetic sound records by the instrument known as a telegraphone. The telegraphone is an instrument which subjects a fine wire moving between the poles of a small magnet to the voice currents and leaves some alternations of residual magnetism in that wire, so that when it is again passed in front of the magnetic poles, those sounds may be feebly reproduced. The purpose of the De Forest patent is to amplify those feeble voice currents. It has nothing whatever to do with radio frequency or with selectivity. On the contrary, in so far as it is an audio frequency apparatus, it is subject to the general rules of audio frequency, the circuits of which are not arranged to exclude, but are contrived with all the skill in the possession of the designer so as to include as wide a range of frequencies as possible. It does not seem worth while wasting time on the patent, as it has absolutely nothing to do with tuning in geometrical progression or selectivity of any kind."
Obviously, the patent relates to audio frequency circuits. Further, the circuits are not tuned.
The Von Lieben et al. patent, French patent, No. 13,726, was before Judge Campbell in Westinghouse Co. et al. v. Royal-Eastern Co. et al. (D.C.) 9 F.2d 397, in the Eastern District of New York; he stated with respect to the Von Lieben et al. patents (page 403):
"They disclose two stages of audio frequency amplification."
"In the apparatus constructed according to the circuit diagram of the Von Lieben French patent, amsco chart 6, by the defendant, and demonstrated by it, there was no tuning shown, and tuning would be an utterly undesirable thing in an audio frequency amplifier."
The somewhat vague language of the patentee cannot alter the situation. Nowhere does he teach how to avoid undesired currents. It is needless to say that no receiving set would to-day be marketable that was not capable of tuning out undesired waves. De Forest and Von Lieben et al. were seeking to show how to amplify uniformly the received sounds. They did not have in mind Alexanderson's problem.
The Armstrong patent, No. 1,113,149, antedates Alexanderson. Major Armstrong, called to prove the effective date of his patent, when made a witness for the plaintiffs, testified as follows, after first describing his circuit:
"Q. 32: Are you familiar with the effects obtained in radio circuits by a succession of tuned circuits connected together by unidirectional couplings such as an audion? (Alexanderson method.) A. Yes.
"Q. 33: Do those effects resemble the effects obtained in such an arrangement as fig. 3 of your patent? A. They resemble it to this extent, that that arrangement also gives greater selectivity; but it is a different kind of selectivity and it is a selectivity which is most useful against types of interference where the regenerative circuit selectivity is not very good. I might illustrate that in this way: If you were out in the country where you were a considerable distance from all stations and if the signals received from these different stations were of the same order of strength, that is, one perhaps was ten times stronger than the other, then even though the stations were very close together in wave lengths the regenerative selectivity would be very effective. On the other hand, if you were in town with a regenerative receiver and you had a high-powered station near you, then, even though this station was at a very considerable difference in wave length from some other out of town station which you wanted to receive, on account of the fact that it was so much stronger it would give you a type of interference which the regenerative circuit would not overcome. "Now, in that last-mentioned type of interference, that is, when you are in town and near a strong station, the cascading or geometric tuning is at its best provided the station which you want to receive is at a considerable difference in wave length from the local station which is causing the interference. The best type of selectivity or the best results are obtained by using both geometric tuning and regenerative tuning in the same instrument.
"Q. 33. Is there any cascade or geometric selectivity in the arrangement of fig. 3 or in any other figure of your patent? A. No.
"Q. 34: Will you please explain why you consider that there is no such selectivity? A. In order to realize such selectivity it would be necessary to couple to the coil L' in the plate circuit of fig. 3 another tube system which would be substantially the same as is illustrated in fig. 3, with the exception, of course, of the antenna. That is, a coil which would correspond to coil S of the second audion system would be coupled to coil L' of the first.
"Q. 35: When did you first use in a radio circuit audion tubes arranged in cascade or series? A. I think it would be in February or March of 1914."
Armstrong built a regenerative set. He obtains selectivity by regeneration, which increases the desired signal. Alexanderson by successive filtration obtains his selectivity.
The Circuit Court of Appeals for the Second Circuit said in Armstrong v. De Forest Co., 280 F. 584:
"The Armstrong invention consists in applying the variations in the current, which are of radio frequency, in such manner as to transfer energy back into the grid circuit.
"Then he found, not only that the radio frequency waves could be carried over into the plate circuit, but that they could be there amplified by the energy derived from the local battery in the plate circuit without change of frequency or wave form, and that they could be fed into the grid circuit, where they increased the potential variations on the grid."
The plaintiffs antedate the Reisz patent, No. 1,234,489, effective April 9, 1913, so that it is unnecessary to consider this reference.
The Schloemilch and Von Bronk patent, No. 1,087,892, is assumed, for the purposes of this case, to have an effective date of February 8, 1913. As before indicated, Alexanderson antedates this patent. He conceived of his arrangement in November, 1912. He spoke of it to Mr. John Hays Hammond, Jr., Dr. Langmuir, and others shortly thereafter, and wrote about it in a letter of February 4, 1913. The patent in question, assuming it to have an earlier effective date, then Alexanderson is not an anticipation. The defendant is licensed under this patent. The United States Department of Justice filed a brief in case as a friend of the court.
Fig. 3 of the Schloemilch and Von Bronk patent is as follows:
The specifications describe this arrangement as follows:
"Fig. 3 shows an arrangement for obtaining such repeated increase by means of several gaseous currents. Here again, a designates the vacuum tube having the oxid cathode c, heated by the battery b, the anode d, and the auxiliary anode e. The oscillations produced by the aerial f in the winding g are supplied, as in the embodiment shown in Fig. 1, to the auxiliary anode e and to the cathode c. The increased high-frequency oscillations then flow in the circuit closed by the source of direct current i over the cathode c and anode d and are supplied from this circuit by means of the transformer k to the detector circuit comprising the detector l and condenser p. An intermediate circuit n syntonized to the oscillations will preferably be provided. The low-frequency impulses or currents supplied by the detector l are now supplied over a second transformer o and a second vacuum tube s1, having the cathode c1, the anode d1 and the auxiliary e1, and are increased again by this tube, and the increased currents of low frequency are finally supplied from the circuit of the source of direct current i1 through another transformer q to the telephone m or other indicating instruments. Obviously, the currents could be increased still further by additional vacuum tubes."
It should be noted that the tube is not used as a detector. A crystal is used as a detector, and the tubes are used to amplify the received currents. There is not a word said in the patent about tuning the circuits to obtain selectivity. Mr. Loftin bridges this ditch by saying that such would be understood from the practice and teachings of the prior art. However, the prior art references and teachings show no such thing. The government cites in support of the teachings of the prior art the De Forest patent, No. 879,532, as showing such a tuning of the grid circuit of the tube to the antenna. The disclosure of this patent is merely of an audion system used as a detector and shows no tuned circuits coupled by an audion. If Schloemilch and Von Bronk had selectivity in mind it is a very obvious assumption that they must have known how to show it and tell about it. This they did not do.
Mr. Waterman said, in disposing of this reference:
"The Schloemilch and Von Bronk patent makes no reference to selectivity, nor does it deal with it in any way directly or indirectly. It is, like most of the other patents that have been referred to, addressed to an entirely different problem, and I agree with Mr. Loftin, if I was correct in understanding him to say that it is on a par with the other references, excepting of course those that deal with selectivity, like the Stone and the Hammond.
"The Schloemilch and Von Bronk patent states this proposition — that whereas the three element tube had been used as a detector, the patentees found that a more distinct signal could be obtained in the telephone if the tube is used to amplify and a crystal is employed to detect. There are those to-day who are of the same opinion.
"The basis of Mr. Loftin's statements regarding the Schloemilch and Von Bronk illustration, as I understand it, is that whereas figs. 1, 2, and 4 of that patent show merely an antenna coupled by a transformer to a tube, there being no variable element in the transformer, fig. 3 does have an arrow pointed at the secondary turns, and he (Mr. Loftin) concludes that that arrow is intended to disclose a method of tuning that secondary circuit, making use of its own capacity and of the capacity of the tube, which of course is very, very small.
"I don't understand what authority there is for such a view, inasmuch as the only circuit which is described as resonant is the circuit possessing the condenser n. In that circuit there is also an arrow, which arrow connects the crystal l with the coil k.
"The patent does not mention either of these arrows unless I have made an oversight, but the purpose of the arrow in conjunction with the crystal detector and coil k is very well known. It is in order to vary the coupling of the crystal to the coil, and it is a logical deduction that the arrow connected with the secondary of the transformer on the grid side of the tube is similarly to vary the coupling of the tube with that coil.
"If a tube were used as the detector, in other words, to replace the crystal, a variable connection symbolized by the arrow would in accordance with this disclosure be used to vary the coupling of the tube with that coil.
"When the tube is transferred, therefore, to an amplifying function and connected to the transformer, no change is made which would suggest the omission of that arrow, and I think the best that can be said of the matter, therefore, is that we have a disclosure which says nothing whatever about selectivity, which is, on the contrary, clearly directed to another problem, and we have an ambiguous drawing."
As before mentioned, the reference is of mere academic interest, since Alexanderson antedates this patent. However, it is to be noted that the defendant, though licensed under this German patent by the Navy Department, does not use it. The defendant does not use a crystal detector or a tuned antenna as called for in the patent, and does use a tuned secondary circuit associated with an untuned antenna. There are other differences enumerated by Mr. Waterman and satisfactorily explained by him which are unnecessary to state here at length.
The Lorenz German patent, No. 258,478, seems to have an effective date of April 5, 1913, clearly not an anticipation. However, the system of selectivity used was quite different from Alexanderson's method. Lorenz tuned to audio frequency and not radio frequency.
Even though the physical structures of some of the prior art references bear some trifling resemblance to the structure of the patent in suit, none has been presented which shows a system which operated or could be operated as Alexanderson's was operated. He strove for, and was the first to obtain, a method to secure selectivity by the use of the audion in circuits so arranged as to filter out the designed current and repeat it from circuit to circuit, building it up as it went in geometric progression, a truly great invention.
"In regard to all of the above patents, adduced against Richardson's patent of 1866, it may be generally said, that they never were, in their day, and before the date of that patent, or of Richardson's invention, known or recognized as producing any such result as his apparatus of that patent produces, as above defined. Likenesses in them, in physical structure, to the apparatus of Richardson, in important particulars, may be pointed out, but it is only as the anatomy of a corpse resembles that of the living being. * * * Taught by Richardson, and by the use of his apparatus, it is not difficult for skilled mechanics to take the prior structures and so arrange and use them as to produce more or less of the beneficial results first made known by Richardson; but, prior to 1866, though these old patents and their descriptions were accessible, no valve was made producing any such results."
The defendant sought to attack the invention by the testimony of Mr. Benjamin F. Miessner, now engaged as chief engineer for a firm manufacturing receivers similar to the defendant's. Mr. Miessner was employed in 1912 by Mr. Hammond, who was experimenting with means to control torpedoes. Mr. Miessner made various sketches from time to time in a somewhat dilapidated notebook from which leaves had been torn and others inserted. The notebook drawings show the use of audions in various circuits for torpedo control, and whatever invention was made was undoubtedly embodied in Mr. Hammond's patent, No. 1,318,342. The notebook on page 5 does not state that the two circuits shown were tuned to the same frequency. Crystal detectors are shown and not audions. However, the next page shows the use of an audion. Mr. Miessner, who has now a decided interest in the outcome of the litigation, says that the circuits of this system were tuned to the same frequency. The notes do not show this. Further, Mr. Miessner's problem at Gloucester was not one of selectivity, but one of control.
The Circuit Court of Appeals for this circuit said, in Pyrene Mfg. Co. v. Boyce, 292 F. 480:
"It is well settled that evidence of prior use must be clear and satisfactory and in its consideration every reasonable doubt should be resolved against the one raising the question. Eibel Process Co. v. Minnesota Ontario Paper Co., 261 U.S. 45, 43 S. Ct. 322, 67 L. Ed. ; Barbed Wire Patent Case, 143 U.S. 275, 284, 12 S. Ct. 450, 36 L. Ed. 161; Coffin v. Ogden, 85 U.S. (18 Wall.) 120, 124, 21 L. Ed. 821."
Mr. Miessner said as late as 1916 in his book on Radiodynamics, p. 153, that his efforts in 1912 were experimental only. Such is not invention. Eibel Process Co. v. Minnesota, 261 U.S. 45, 43 S. Ct. 322, 67 L. Ed. 523.
Of course, the circumstance that defendant has a license from the Navy Department to use the Schloemilch and Von Bronk patent (a later patent) does not avoid infringement of the patent in suit. Diamond Match Co. v. Ruby Match Co. (C.C.) 127 F. 341, 344.
The plaintiffs are entitled to an injunction and an accounting.