Circuit Court of Appeals, Second Circuit.
Pennie, Davis, Marvin & Edmonds, of New York City (William H. Davis, Baldwin Guild, and Curt von Boetticher, Jr., all of New York City, of counsel), for appellant.
Samuel E. Darby, Jr., of New York City, Thomas G. Haight, of Jersey City, N. J., and Darby & Darby, of New York City, for appellees.
Before L. HAND, SWAN, and CHASE, Circuit Judges.
L. HAND, Circuit Judge.
This is an appeal from a decree in equity dismissing a bill to enjoin the infringement of claims 1, 5, 6 and 10 of patent No. 1,879,863, issued on September 27, 1932, to Harold A. Wheeler and assigned by him to the plaintiff. Judge Galston's opinion below, reported in (D. C.) 7 F. Supp. 908, describes the invention accurately and fully, as well as the more important parts of the prior art. We reach the same result as he, and though we follow a somewhat different course, we adopt his statement in general. He thought that the claims should not be limited to receiving sets and for that reason he found Affel's patent, No. 1,574,780, an anticipation. Necessarily he also thought that they should not be limited to automatic control by impressing the voltage of the plate circuit in the detector upon *330 the grid of an earlier radio frequency amplifier; that is, by sending it backward instead of forward. No doubt verbally the claims are general and apply to transmitting units in which the controlling connection is between the plate circuit of the detector and a subsequent amplifier. But all the figures are of receiving sets and the language of the specifications fits them; throughout it is apparent that the invention was primarily intended to be embodied in receivers, and it might be justifiable to interpolate by reference these limitations, if there were a genuine invention to be saved in that way. At first it seems to have been supposed that such implications were proper only when the claims had the conventional phrase, "substantially as described." Seymour v. Osborne, 11 Wall. 516, 547, 20 L. Ed. 33; Westinghouse v. Boyden Power-Brake Co., 170 U.S. 537, 538, 18 S. Ct. 707, 42 L. Ed. 1136. But this has long been recognized as unnecessary, and courts now construe the patent as a whole according to its predominant intent. Mantle Lamp Co. v. George Bowman Co., 53 F.(2d) 441, 444 (C. C. A. 6); Smokador Mfg. Co. v. Tubular Products Co., 31 F. (2d) 255, 257 (C. C. A. 2). Even if patent claims were not especially ductile, as the doctrine of equivalents is witness that they are, the parts of any legal document always in some measure mirror the whole and cannot be otherwise understood. However, we do not find it necessary to decide whether these claims should be so limited, because even if we read them as the plaintiff wishes, we think that there was no invention.
It is of course true that if they incorporate the element of a receiving set and a feed back, Affel's patent, No. 1,574,780, is not an anticipation on all fours. The automatic control was to be used in a transmission set, and the voltage from the plate circuit of the rectifier was fed forward, not backward; although, as all the amplifiers were of radio frequency, it does not appear to us how that choice was significant. In a receiving set it turns out for one reason or another to be impracticable to control the audio frequency amplifiers, but it certainly could not have required much ingenuity to learn so. At any rate both in principle and means the invention anticipated Wheeler; it automatically controlled the amplification of a received signal by feeding the voltage of the rectified current from the plate so as to bias the grid of an amplifier; this through a direct connection between the two. The bias on the grid increased with any increase of amplitude in the incoming energy, so that the control increased with its necessity. Finally, although apparently such a control will operate when the plate is positive, provided that it changes proportionately with the amplitude of the signal, Affel's disclosure chances to disclose a negative potential on the plate of the rectifier and the grid of the amplifier. The important thing is that the voltage impressed upon the grid shall be itself proportionate to the amplification which it is to control. With Judge Galston we cannot regard it as critical that a receiving set must be tunable to various frequencies. Though the set is fixed for each frequency, a change to a new frequency of different power is like a corresponding change of power at the same frequency. Certainly whatever are the engineering difficulties of applying such control to a receiving set, it took no ingenuity merely to carry over the idea from a transmitter. The problem of the automatic control of amplification to correspond with variations in amplitude of the carrier current had been solved, and the basic means was known; Wheeler used that means. Björnson, No. 1,666,676, added very little to Affel; the purpose was to cut out line noises from a telephone system. Like Affel he used a rectifier in shunt with the amplifier; the rectified current carried forward, not backward, the negative voltage of the plate to the grid of an amplifier and biased it in proportion to the degree of amplification. For its purpose it was apparently well adapted; but again the unit contained no audio frequency amplifier. Heising, No. 1,687,245, is important only because instead of a triode he used a diode which requires no independent "B" battery. His too was a transmission system, and the control was introduced apparently only to insure that the grids should remain always negative to save power, not an important factor in a receiver. As it stood it could not have been successfully used for Wheeler's purposes.
None of these patents being complete anticipations, the question is as to the importance of the step from them to Wheeler. His application was filed on July 7, 1927; Affel's No. 1,574,780, on October 5, 1921; Björnson's on June 27, 1925; Heising's on December 30, 1922. Affel's patent issued on March 2, 1926, a little more than a year before *331 Wheeler's application, and the other two some time after he filed. We have no reason to assume that Affel's invention was known before his patent issued, except in so far as his earlier patent, No. 1,511,015, issued on October 7, 1924, can be said to have disclosed it. Assuming that it did, the interval was only about two years and a half; the art did not have to wait long for the change. Nor does it seem to us that when it came, it involved invention. As we have said, the mere notion was obvious enough, for the defect was apparent in receiving sets; what was loud enough at one frequency was deafening at another. What would correct variations of impressed energy upon a transmission system quite naturally suggests itself as perhaps feasible for a receiving set; one would try it out at least. Great store is set by the fact that the feed was back to the radio frequency amplifiers instead of forward; this the plaintiff's expert after the manner of his kind chooses to Latinize, "regressive" as against "progressive." But the choice seems patent; if feeding the voltage from the rectified plate circuit to an audio frequency amplifier distorted its output, nothing was nearer at hand than to try feeding it backward to the radio frequency amplifiers which had been satisfactorily controlled before. Indeed this is just what Friis had already done in a receiving set. True Affel's rectifier did not rectify the whole current; it was in shunt, and had to be, for a transmitter must emit alternating pulses. But in a receiving set where the whole current must become unidirectional, there was no need of this; and it seems a natural expedient to use as rectifier the detector already there. Again as to Wheeler's triode used as a diode, it was an old device; Heising had used a diode in a not dissimilar setting, and that is not to be distinguished from a triode used as a diode. When, as in a receiving set, compactness becomes an end, why not eliminate the "B" battery necessary to a triode? Finally there is the question of the "time delay," or "time constant" (page 3, lines 86-130). This in figure 1 resulted from the combined effect of the resistance, 53, and the shunted condenser, 54; the delay was a function of the product of these two; we may assume that it is necessary to preserve the lower audio frequency modulations of the carrier current in the radio frequency amplifier. The record tells us nothing more about it except that none of the prior patents contain this feature. We cannot treat it as the basis of the invention, when, so far as we can see, there could be nothing else for it to rest upon. Nothing of the sort is even intimated in the claims and the feature can be imported into them only on the theory that it is essential to the operation of a receiving set. But this goes further than we are willing. Though it be fair to assume that the claims should be generally limited to such sets, the specifications do not themselves declare that the "time constant" is an inevitable essential, and it is entirely possible that the claims meant to leave it as an optional adjunct. Moreover, the mere fact that it was necessary, if it really was, tells us nothing as to the engineering skill necessary to devise it. At best Wheeler could not have regarded it as an important feature; and we find no reason to use the stone which he rejected as the head of the corner. Thus we can find no evidence that the steps he took, given the idea of transferring Affel to a receiving set, demanded more than competent designing.
The plaintiff invokes the test of success; quite naturally because the device has been pretty widely used. But the conditions were not such as to make the evidence impressive. The plaintiff owns more than one hundred and fifty patents which it licenses en bloc, giving to its customers in addition, as they may need it, the help of its corps of engineers. They are free to use all its patents for a single fee, so that it costs nothing to add any new ones as they appear. At last accounts more than one half of them were using Wheeler's invention, and the treasurer of the plaintiff estimated that four-fifths of the whole trade used it; an estimate however which was no better than a guess. A new device need be only a minor advance, for people placed like the plaintiff's customers to exploit it when in substance they could have it for nothing. Again, the plaintiff argues that there had been earlier efforts to get automatic volume control in receiving sets and that they had followed other lines. These are embodied in the patents to Friis, No. 1,675,848, and Falknor, No. 1,698,014. The dates of these were as follows: Friis's application November 29, 1924; patent July 3, 1928; Falknor's application March 13, 1924; patent January 8, 1929. Friis fed back the voltage of a rectified plate circuit to a radio frequency amplifier, as we have already said; so much of Wheeler at any rate he anticipated. But *332 the rectifier from which the control voltage proceeds was added and in shunt with the output circuit of the detector proper; he did not use the detector. It was not indeed an anticipation, and Wheeler is entitled to whatever inference there may be from an earlier effort which followed the old pattern. Little is to be inferred from the fact that Falknor adopted quite another method. As to both we may concede that when an invention would have been welcome for a long time after all obstacles to its appearance had been satisfied, or that when the period is short, if a number of others have tried and failed, it makes more probable the conclusion that the change demanded more than common ability. But neither alternative is true here. Wheeler stands upon Affel; and he cannot point to a number of later efforts, or to the lapse of much time before his own work. Especially in the radio art is it dangerous to be impressed by new details; the subject is all very unfamiliar to us; we must proceed quite in the dark, guided only by the interested advice of those whose conclusions we are personally unable to check, as we sometimes can in the mechanical arts. The industry has been the object of an amazingly assiduous ingenuity; and we are to suppose that many permutations will appear spontaneously from the constant efforts of numerous competent experimenters. True, this may be used as especial evidence of invention when the need is old and the success striking, but it counts strongly against novelty merely as such. We have had occasion to speak of this before in this very art. Technidyne Corp. v. McPhilben-Keator, 72 F.(2d) 242 (C. C. A. 2). We put this patent down as one of those step by step advances, not beyond the compass of capable investigators who run down every lead, and cull out those which appear advantageous. It might be desirable to promote such activities by limited monopolies, but that is not the law; patents do not go to patient and exhaustive experiment; they are the reward of exceptional talent.
The defendants, the Emerson companies, secured the dismissal of the bill as against themselves, because the court lacked personal jurisdiction over them. In the view we take of the patent this error, if any, is harmless and the point has become moot; for, if we should reverse the decree for that reason and hold that the court in fact had jurisdiction over these companies, we should at once have to dismiss the bill upon the merits as against them. That would create an estoppel against the plaintiff, which would then be left in a worse position than before the appeal, for theoretically at least, it will be possible to sue the Emerson Companies again. As to the individual defendant, Abrams, the court dismissed the bill upon the merits, holding that because he resided in the Eastern district of New York he was subject to its jurisdiction. This was plainly correct; and the confusion which has arisen seems to be between jurisdiction to determine whether Abrams infringed the patent, and the correctness of the decision that he had not. The court had jurisdiction over him, however it decided the issue of infringement. Therefore, as in the case of the companies if the judge was in error in holding that Abrams did not infringe, the error is moot; for the plaintiff is in rather better position with the decree affirmed for non-infringement, than if it were affirmed for invalidity.
Decree affirmed.