Opinion
July 11, 1927.
Sheffield Betts, of New York City (Charles Neave, Stephen H. Philbin, and Abel E. Blackmar, all of New York City, and Harry E. Dunham, of Schenectady, N.Y., of counsel), for plaintiffs.
Miller, Otis, Farr Henderson, of New York City (C.D. Ehret, of Philadelphia, Pa., George F. Scull, of New York City, Paxson Deeter, of Philadelphia, Pa., and H.B. Farr, of New York City, of counsel), for defendant.
In Equity. Suit for infringement of Alexanderson patent, No. 1,173,079, brought by the Radio Corporation of America and the Westinghouse Electric Manufacturing Company, as licensees, against E.J. Edmond Co., Inc., involving claims 1, 2, 3, 9, and 12. Decree for complainants.
See, also, 18 F.2d 281.
The patent and claims in suit were held valid and infringed in Radio Corporation et al. v. Splitdorf Electric Co., 14 F.2d 643, by Bodine, District Judge, whose conclusion finds strong support in the thoroughly well considered opinion of MacLean, J., of the Exchequer Court of Canada, in Canadian General Electric Co., Ltd., v. Fada Radio, Ltd., decided April 14, 1927, where Alexanderson's Canadian patent, No. 208,583, covering the same invention, was held valid and infringed.
The patent relates to radio receiving apparatus, and particularly to a system for the selection of oscillations of a given wave length from mixed oscillations. The problem of selecting oscillations of a given wave length from the confusion of oscillating currents affecting the antenna of radio receiving apparatus had long been known in the prior art of radio reception. The Stone patent, No. 714,756, of December 2, 1902, and the Marconi-Franklin British patent, No. 12,960, dated December 23, 1907, disclose tuned circuits inductively coupled, arranged in cascade to obtain selectivity. Stone and Marconi attained high degrees of selectivity through the use of successive tuned circuits, each of which to a degree screened out the interfering oscillations with which they were not in resonance, and passed on the desired oscillations with which they were in resonance. Selectivity in such circuits was, however, dependent upon loose inductive coupling between the successive circuits, and with loose coupling the strength of the desired signal was greatly diminished. If close coupling was employed to increase the signal strength, the undesired oscillations were carried from one circuit to another and the system ceased to be selective. Thus in Stone and Marconi it was extremely difficult, if not impossible, to detect the weak signals of fardistant stations, when the antenna was affected by the interference of powerful nearby stations.
In the fall of 1912, Alexanderson first learned of the De Forest audion from Hammond. Immediately the thought occurred to him that it might be used as a relay for radio frequency currents in a series of tuned circuits. It was his conception that by a series of tuned circuits, connected by suitable relays, which would effectively impress upon the succeeding circuit the selected oscillations of the preceding circuit, a very high degree of selectivity could be attained without loss of signal strength. This conception may have occurred to others, but, whether it did or not, Alexanderson was the first to use the audion as a relay for this purpose. That this was invention I have no doubt. When Alexanderson told Hammond of his intention to try the audion as a relay of radio frequency oscillations connecting tuned circuits, Hammond said he believed the audion was too sluggish for this purpose; but Alexanderson and the engineers of the General Electric Company working under his supervision constructed a series of resonant circuits connected by audions as relays, and attained, without loss of signal strength, a greater degree of selectivity than had theretofore been possible. The audion was known in the art as early as February, 1908, and the fact that no one had used it for this purpose prior to Alexanderson is persuasive evidence of invention.
That the prior art does not disclose anticipation of Alexanderson's invention is very clearly shown in the Splitdorf and Fada Cases, cited supra. Indeed, the only references claimed to be in anticipation of Alexanderson which are discussed in the defendant's brief are Schloemilch and Von Bronk, patent No. 1,087,892, and Armstrong, patent No. 1,113,149. I find in the disclosure of Schloemilch and Von Bronk nothing which can be said to advance the art of selectivity in radio receivers or anticipate Alexanderson. The purpose of this invention was to amplify radio frequency currents, and nothing is said about selecting desired signals, or excluding interfering signals. In one of the patent drawings there is shown a tuned antenna circuit coupled by induction with the grid circuit of an audion amplifier, the plate circuit of which is in turn coupled by induction with a detector circuit. Neither the grid nor the plate circuit of the audion is tuned. Connected with the detector circuit, an intermediate tuned circuit is shown. There is no association of successively tuned circuits connected by relays, and all that is said of the tuned circuit which is associated with the detector is: "An intermediate circuit N syntonized to the oscillations will preferably be provided." For what purpose is not stated. No doubt because it was well known that a circuit tuned to resonance of a desired signal would exclude interfering signals to some extent. This casual reference is not shown to have any relevance to the disclosure. Apparently it was merely a mechanical suggestion for the insertion in the receiving instrument of a tuned circuit, so that the system might be given some slight measure of selectivity by the insertion of a means to that end which was well known in the art. It was certainly not a disclosure or anticipation of Alexanderson's arrangement of successive circuits, all tuned to the same frequency and linked together by relays. The precise date of Alexanderson's invention becomes unimportant, and need not be considered, because the invention of Schloemilch and Von Bronk, even if prior in time, does not anticipate the invention of Alexanderson.
In Armstrong's feed back audion circuit, patent No. 1,113,149, to which priority of invention is conceded, selectivity is attained by regeneration. Prior to his invention the presence of radio frequency oscillations in the plate or wing circuit of an audion used only as a detector was not suspected. Armstrong tuned the plate circuit of an audion detector to radio frequency in resonance with the grid circuit, and through his feed back connection between the two circuits found that he could feed back from the plate circuit into the grid circuit radio frequency oscillations of the same frequency as that of the desired incoming signal, and thus obtain marked amplification of the desired signal. Armstrong v. De Forest Radio Tel. Tel. Co. (C.C.A.) 280 F. 584, 587-588. Armstrong, who was a witness for the defendant, testified that the regenerative selectivity attainable in his feed back circuit is due to the fact that energy is fed back from the plate circuit to the grid circuit in phase with and assisting the desired oscillations in the grid circuit; that the energy which is fed back passes only one way, from the plate circuit to the grid circuit; and that consequently selectivity in this system is due solely to the amplification of the desired signal, with no appreciable screening or exclusion of the undesired signal. His plate circuit is tuned so that the energy fed back will reinforce the desired oscillations in the grid circuit. The interfering signal is not depressed, diminished, or decreased. Geometric tuning by means of a succession of resonant circuits coupled by relays is neither disclosed nor employed in the Armstrong feed back audion circuit. There is superficial resemblance between the circuit drawings of the two patents, but when the operation of the two devices is understood it is quite apparent that there is no similarity of invention. In Alexanderson's tuned circuits the audion is not used as a detector, detection occurring at a later stage, while in Armstrong the set has but one audion, which is used as a detector and as a regenerator of radio frequency oscillations. The two inventions are utterly different, and Armstrong does not disclose or anticipate Alexanderson.
It is contended that a device, built in accordance with the disclosure of the patent, with such knowledge only as was common to those skilled in the art at the time of its issue, would be inoperative and useless. Upon this issue the burden rests heavily upon the defendant. Remington Cash Register Co. v. National Cash Register Co. (D.C.) 6 F.2d 585, 618. And of course perfection in operation is not the true test of operativeness. Engineer Co. v. Hotel Astor (D.C.) 226 F. 779, 783. A patented device is not inoperative because, when built or operated in a certain way, it will not work, if other ways of building and operating it are fairly indicated by the patent or the prior art. A patent is addressed to men who know the art, and if such men can build and operate the device, so that it functions as the patent says it will, although imperfectly and only after trial and error, that is enough. A.B. Dick Co. v. Barnett (C.C.A.) 288 F. 799, 801; American Stainless Steel Co. v. Ludlum Steel Co. (C.C.A.) 290 F. 103, 108.
Here the claim of inoperativeness is based upon the contention that Alexanderson's arrangement is subject to the regenerative influence of feed back from the plate circuit to the grid circuit of each audion, by which self-oscillations are produced which prevent the functioning of the device. That such oscillations may occur in operation is quite clear from the tests and testimony. The cause of these oscillations was first disclosed by Armstrong in his patent No. 1,113,149, issued October 6, 1914, covering his regenerative feed back system of amplification. Modern methods for the control of regeneration and the undesirable self-oscillations which result therefrom were unknown in the art when the patent in suit issued. These improved methods have perfected Alexanderson's system of geometric tuning, and are employed by the defendant. Their importance cannot be denied (see Hazeltine Corporation v. Electric Service Eng. Corp'n [D.C.] 18 F.[2d] 662), and it may be conceded that the Alexanderson device, without such improvement, would be of little commercial value to-day.
But infringement of the Alexanderson patent of 1913 is not avoided by using his invention with subsequent improvements. International Time R. Co. v. Bundy R. Co. (C.C.A.) 159 F. 464, 469. The question is whether without improvement the invention was operative. The evidence clearly shows that regeneration occurs in the patented device only under conditions dependent upon elements variable in operation. By empirical variation of these elements the device can certainly be made to work. Thus self-oscillation can be suppressed by reducing the amplifying voltage or by loosening the inductive couplings between the circuits and, in general, by changing the ratios of resistance, inductance and capacity in the various circuits. If the strength of the plate current is merely sufficient to sustain the original signal strength without amplification, there will be no objectionable regenerative effect. None of the elements upon which self-oscillation depends are fixed by patent specification. On the contrary, the instruction of the patent is that they are to be varied in operation. When so varied by the operator, the objectionable regenerative effects can certainly be eliminated by trial and error, if too great amplification is not insisted upon.
The tests recited by the defendant's expert are not convincing. In the first place, he omitted variable inductances from the grid circuits and constructed an instrument which he knew in advance would not function. No room was left for trial and error, and little, if any, allowance was made for the knowledge of one skilled in the art in the year 1913. In Loom Co. v. Higgins, 105 U.S. 580, 586 ( 26 L. Ed. 1177) it was said: "When the question is whether a thing can be done or not, it is always easy to find persons ready to show how not to do it." In Rynear Co. v. Evans (C.C.) 83 F. 696, 697, Judge Coxe said that, "when an expert undertakes to prove that his adversary's process or machine is a failure, he always scores a success. It is much easier to make a machine that will not work than one that will." The defendant's tests show that an Alexanderson device can be made which will not work. They do not show that such a device could not have been made to work by one skilled in the art in 1913.
The tests of the apparatus conducted upon the trial convincingly demonstrated that within certain limits a very high degree of selectivity was attained, with substantial amplification of signal strength, and that when the system was caused to oscillate these oscillations were easily eliminated by the variation of inductance or voltage, which any operator of the device in 1913 certainly would have tried in an intelligent effort to make it work, even though he was entirely ignorant of the mysteries of the audion which were later explained by Armstrong.
Alexanderson's system was actually used with success in receiving in Schenectady radio signals transmitted from Honolulu, and since 1920 in the transoceanic service of the Radio Corporation of America, where a high degree of selectivity is essential, 15 receiving sets being connected with a single antenna, and each receiver being used for the selection of radio frequency oscillations of a frequency differing from those received on the other sets. With improved methods of neutralizing the effect of regeneration, the system is extensively used in commercial receiving sets to-day. Alexanderson's contribution to the art was important, and the validity of his patent cannot be disputed upon the ground of inoperativeness.
Nor do I think the point well taken that Alexanderson did not invent the device described in the patent, but that his assistant, Langmuir, did. Langmuir was one of the engineers of the General Electric Company associated with Alexanderson in the research department of that company. What he did was done pursuant to Alexanderson's request, and in execution of the conception which was Alexanderson's. He contributed no inventive thought to the plan which had been fully outlined by Alexanderson. His task was to construct a series of tuned circuits connected by audion relays. It is true that before he undertook to do this he and his assistants manufactured audions with a very high vacuum, so as to exclude the possibility of gas ionization in the operation of the audion. These improved audions are preferred in the Alexanderson patent, but are not specified as essential, nor is there proof that they are essential. If credit is due to Langmuir for the invention of these improved audions, this should not deprive Alexanderson of his invention of a selective tuning system in which they are used as relays.
Claims 1, 2, and 12 are sufficiently broad to cover any relay in place of the audion. Claims 3 and 9 are limited to devices in which the audion is used as a relay between the cascaded tuned circuits. Inasmuch as it does not appear that any effective relay for radio frequency currents other than the audion was known to Alexanderson, it may be that claims 1, 2, and 12 are too broad. This question should, I think, be reserved until it arises in a suit in which infringement is asserted against a device in which a relay other than the audion is employed.
Claims 3 and 9 are clearly valid, and as clearly infringed. A decree to this effect may be entered, upon the usual notice. The decree may contain appropriate recitals of the fact, which appears from statements made upon the trial by defendant's counsel, that this suit was openly defended by the At-water-Kent Manufacturing Company, a Pennsylvania corporation, the maker of the infringing device. The motion to make this foreign corporation a party to the record, which was made upon the trial, is denied on authority of Parsons Non-Skid Co., Ltd., v. E.J. Willis Co. (C.C.) 176 F. 176; Freeman-Sweet Co. v. Luminous Unit Co. (C.C.A.) 264 F. 107; Van Kannel Revolving Door Co. v. Winton Hotel Co. (D.C.) 263 F. 988. In Dick's Press Guard Mfg. Co. v. Bowen (D.C.) 229 F. 193, where such a motion was granted, the motion was unopposed and the decision of Judge Noyes in the Parsons Case, supra, apparently was not called to the attention of the court.