Western Electric Co. v. Wallerstein (51 F.2d 529, 1931)

GALSTON, District Judge.

This is a patent infringement suit in which infringement of patents to Lowenstein, Mathes, Colpitts, Arnold, and Blattner is. urged.

The defendant operates the Broadway Theatre located in Buffalo. The infringing apparatus is a sound-on-dise talking motion picture reproducing mechanism, which was installed in the defendant’s theatre some time prior to May 7, 1929. The infringement-complained of is the use of two type 10 and two type 11 audio amplifiers manufactured by Paeent Electric Company, Inc.

The five patents were issued to the American Telephone & Telegraph Company, one of the plaintiffs. Another of. the plaintiffs is the Western Electric Company, wMch is the manufacturing subsidiary of the Telephone Company. It is controlled by the Telephone Company through its ownersMp of more than 98 per cent, of the capital stock of the Electric Company. The third plaintiff, Electrical Research Products, Inc., is a subsidiary of the Western Electric Company, which owns all of its stock and which is engaged in the commercial distribution of sound picture apparatus and some other commercial products manufactured by the Western Electric Company.

Western Electric and Products are the exclusive licensees of the Telephone Company under the patents in suit.

The defense herein was undertaken by Warner Brothers Pictures, Inc., for the reason that Warner Brothers control the chain of Stanley theatres which have installed Paeent talking picture reproducing systems similar to that of the defendant herein.

The patents involved are of extreme importance, some of them, perhaps, going to the very heart of the sound moving picture industry, although at the time of the conception of these inventions their association with the reproduction of pictures was not even remotely imagined.

Rather, the early researches of the Telephone Company were directed to the improvement of transcontinental telephonic communication. Such a line was completed in July, 1914, and was open to the public in January, 1915. Repeaters were required for satisfactory long-distance telephone service. Such development was undertaken by the Telephone Company as early as 1903, but it was not until the De Forest three-electrode vacuum tube, and the associated' circuit in wMch it was to be used, had been developed, that marked progress was made.

Extraordinary progress in the art of sound reproduction followed De Forest’s disclosure of the three-electrode vacuum tube. The problems involved were intricate. Sound is composed of a wide range of frequencies. A range of frequencies audible to the human ear extends from about 16 to 16,000 vibrations per second. The piano range is estimated to be from 30 to 5,130 cycles per second.

It is not necessary, as appears from the record herein, to reproduce so wide a range for effective telephone communication; but it is necessary to embody such wide range for successful reproduction of sound pictures and kindred purposes. In other words, much greater power is required to drive the loud speakers employed for public broadcasting and sound picture reproduction than is necessary for radio or telephonic reception, for the reason that the output of these loud speakers must fill large auditoriums and large outdoor areas. Distortion in the reproduction of the sound must, therefore, be zealously guarded against.

It is in these large power applications of audio amplifiers that the inventions, which are the subject-matter of the patents, are concerned, though, of course, they may be used in other fields.

All of the inventions relate to circuits in which the three-electrode tube is employed. These inventions are capable of conjoint use, and are, therefore, properly embodied in this one cause of action.

The scientific principles and the highly technical and abstruse problems involved, at all times presented to the court an intellectual challenge. It is with a feeling of marked appreciation that the court acknowledges the assistance and co-operation given throughout the long trial by counsel and experts alike.

Patent No. 1,231,764, issued to Fritz Lowenstein July 3, 1917, on an application filed April 24, 1912, and renewed April 26, 1917, is for an improvement in telephone relays. With the exception of the third claim, all of the claims are alleged to be infringed. They are resisted on the ground of invalidity, and also because of an alleged estoppel arising out of a license from the Electrical Research Products, Inc., to the Yitaphone Corporation.

The plaintiffs urge claims 1 and 7 as typical, and, therefore, attention may be focused upon them. They read as follows:

“1. Telephone apparatus comprising the combination, with a talking circuit, of a suitably energized relay circuit including an anode and a cathode separated by a conductive gap, a modulating device interposed in said gap and electrically connected with said talking circuit, means for impressing upon said modulating device a potential more negative than that of said cathode, and a translating device arranged to be energized from said relay circuit.”

“7. The combination, with an audion having its anode and cathode included in a suitably energized circuit, of means for impressing upon the audion grid a potential more negative than that of the audion cathode.”

This patent was adjudicated in the Southern district of New York, and these claims, the only claims therein involved, were held valid. Radio Corporation of America et al. v. J. H. Bunnell & Co., Inc., (D. C.) 25 F.(2d) 847.

It is contended by the defendant that the result reached by Judge Winslow, who decided that ease, was due to errors of fact.

The object of the invention was a relay whereby potential differences of incoming speech currents were sought to be maintained in the telephone receiver, with the reproduced sound composed of waves of the same frequencies as those of the incoming sounds, and having the same relative amplitudes as in the original sound waves.

In the- long distance telephone art as then known to Lowenstein, he found that under certain conditions the speaker’s voice was plainly heard in the receiver, while at other times it was unintelligible. This defect he ascribed to the fact that speech currents of higher frequencies become attenuated to a greater relative extent than those of low frequency. Such results were due to electrical properties of a long line. He further observed that the distortion was but little affected by terminal conditions, and concluded that relays or receiving apparatus functioning according to current values could not remedy the distortion nor compensate for the effects; but he said terminal voltages are affected by terminal conditions, and hence argued that it was possible to select such terminal conditions as would enable the relative amplitudes of the original voltages to be well maintained at the incoming end of the line. He reasoned' that a relay apparatus which functions according to terminal voltages was well adapted for the production of receiver operating currents of the desired absolute strengths, and in the relative strengths required for successful operation of the receiver.

Referring to the drawing, we need consider only the circuits concerned with the three-electrode tube 14. The controversy as to the scope and the meaning of this patent arises out of the construction to be given to the following paragraph of the specification: “The potentials created in secondary 13 are made to control the current flowing through the ionic field and originating in battery 21 by connecting the one terminal of coil 13 to the modulating member 18 and the other terminal of said coil to a point on the battery 21, which is located ultranegatively relative to the negative point of the battery connected to the filament.”

The plaintiffs’ position is that these words must be interpreted as meaning that the grid was connected to a point on the battery 21, more negative than any part of the filament. The defendant’s position is that a grid biasing potential, more negative than the most negative potential point of the filament, is not disclosed.

The battery connected to the filament, as interpreted by the plaintiffs, refers to the battery 19. The defendant, on the other hand, insists that there is nothing in the specification to indicate that the negative end of the filament battery 19 is that which is connected with the filament through the lead from the battery 21.

The passage is not unambiguous. In the sentence preceding the quoted passage, it is the battery 21 which is referred to when reference is made to the connection of the filament with “a negative point of the battery.”

The defendant contends that whether the grid biasing potential should be more or less than the most negative potential point of the filament depends upon characteristics of the vacuum tube not disclosed in the Lowenstein patent, and, moreover, that such a negative grid potential was not suitable for the three-electrode tubes known at that time.

The defendant urges that the poling connection of the filament with battery 19 is important, because on it depends whether the grid will be more negative than the negative side of the filament. It is argued, that if the positive end of- battery 19 is connected to the lead then the grid can never be more negative than the -negative end of the filament, unless that portion of the battery 21 connected between the grid and filament has a greater potential than the battery 19.

As has been intimated, the defendant interprets “to the negative point of the battery connected to the filament,” as referring to the battery 21, and argues that such interpretation harmonizes with the phrasing of claim 3, whereas plaintiffs’ interpretation is contradictory of claim 3. Claim 3 reads: “Telephone receiving apparatus comprising the combination, with a talking circuit, of an ionic controller having an anode and a cathode suitably spaced apart, a battery, connections from said anode and cathode to the battery, a grid electrically connected to said talking circuit and interposed between said anode and cathode, a connection from said grid to said battery, the battery connection of the grid being negative with respect to the battery connection of the cathode, and a telephone receiver arranged to be energized from the controller circuit.”

The phrase “a connection from sdid grid to said battery,” the battery connection of the grid being negative with respect to the battery connection of the cathode, certainly would indicate that but one battery was referred to- as descriptive of the connections, to wit, the battery 21, as was admitted by Mr. Waterman in his cross-examination.

Plaintiffs seeking to avoid this interpretation draw attention to the showing in Lowen-stein’s drawing of more cells'in the C battery section of the battery 21 than in battery 19, and conclude that the grid is ultranega-tive with respect to the whole cathode, no matter how the battery 19 is poled.

It seems to me that this conclusion is unwarranted, for, as was admitted at the trial, Patent Office drawings are not working drawings. Therefore, one cannot safely venture' an opinion that, because the battery representation in a Patent Office drawing indicates a given number of cells, the inventor intended in practice such number of cells to be employed.

In the conflict of interpretation,'it is helpful to refer to the original application and prosecution thereof in the Patent Office. The original specification set. forth:. “The potentials created in secondary 13 are made to control the current flowing through the ionic field and originating in battery 21 by connecting the one terminal of coil 13 to the modulating member 18, and the other terminal of said coil to a point on the battery 21, preferably located ultija-negatively relative to the negative point of the battery connected to the filament. Both the positive and the ultra-negative points of connection to battery 21 are adjustable to enable variation of the driving potential of 21 for the ionic field and of the ultra-negative potential.”

The three claims forming part of this specification were rejected on letters patent to De Forest, No. 841,387, January 15, 1907, and No. 995,126, June 13, 1911. In the argument in support of the amendment, filed in response to those citations, it was admitted that a device like De Forest’s which, while it was contended operated proportionally to current variations and not proportionally to potential variations, “could still be said to be actuated by potential variations of the main current, because in a circuit a potential causes a current flowand it was argued that the amendment to claims 1 and 3 brought out the fact that the applicant’s device “is one in which the variations of potential in the main circuit are relayed into the relay circuit in substantially proportional form, so that a true ‘potential actuation’ of the translating device as distinguished from a ‘current actuation’ of said translating device is obtained.” This, applicant thought was new, and differentiated his device from De Forest.

It may be noted that in the claims as originally presented, the term “ultra-negative” was not employed. It appears in the claims for the first time, in claim 4, as filed with that amendment of May 28, 1913, wherein it is said that there are means provided for maintaining an ultra-negative potential on the modulating device, i. e., the grid 18.

The attorney for the applicant wrote: “New claim 4 brings out the additional feature of the ultra-negative potential connection of the battery 21 to the grid 18. As this is new with applicant, it is thought this claim should be allowed since it is due to the ultra-negative potential of applicant’s grid or separator 18 that the sensitiveness and quick response of applicant’s relay is obtained.”

On August 8, 1913, all claims were rejected, claims 1 and 2 by reference to Wein-traub, No. 921,930, and Von Lieben, No. 1,038,910. The Examiner noted that both of those citations showed devices operating according to variations of potential rather than variations of current. And as to claim 4, referring to Fig. 2 of Von Lieben, the Examiner said: “It is clear that the fall of potential across the resistance Rw will allow of any gradation of potential difference found to be desired.”

On August 6,1914, Lowenstein’s attorney filed an argument analyzing the patent cited. He maintained that the Weintraub device is a current intensifier, and not a potential current translator.

As to Von Lieben, the attorney pointed out that his grid H is positive, as against the negative portion of the filament, to the grid H, and hence concluded that Von Lieben was not a disclosure of “a potential-operated device.”

And in this argument appears now for the first time insistence by the applicant that claim 4 as filed expressly sets forth that the grid, being ultranegative, is “more negative than any part of the filamentand attention is called to lines 51 to 54, page 1, of the Von Lieben patent, “from which it appears that the potential of the grid electrode is chosen between that of the cathode and the anode.”

In the next action of the Patent Office, while the claims are rejected for various reasons, it is admitted by the Examiner that Von Lieben in Fig. 2 does not make it clear that his grid is connected to the battery in such a way as to render it negative with respect to the filament. The Examiner adds: “As shown it would appear to be connected so as to be negative with respect to one side of the filament and positive with respect to the others, and is probably intended to be a neutral connection. The advantage of making the grid slightly negative rather than neutral has not been clearly pointed out by applicant and it would seem that this point should be made clear before claim 4 is allowed.”

This led to an amendment on September 11, 1915, in which the advantage of making the grid ultranegative was stressed, and was accompanied by the affidavit of the applicant of September 8, 1915.

The amendment to the specification was this: “This ultra-negative connection is especially desirable and constitutes an improved feature of my invention in its preferred embodiment. The volume and clearness of speech as heard in the receiver in the arrangement shown is materially greater than where the grid 18 and coil 13 are connected, for example, to a point at the same potential as the filament 16. The theoretical considerations involved are somewhat abstruse and not fully understood; but the advantage of the ultra-negative connection is fully established by repeated tests which I have carried out in actual practice. I therefore note the fact without attempting to explain it.”

In the affidavit in support of the amendment, Lowenstein, tracing the history of his invention, states that claims 1, 2, and 3 of his invention were conceived in the year 1906. It seems that he was engaged in carrying on some experiments involving the use of a Cowper-Hewitt mercury rectifier, in the course of which he discovered that, by touching a certain portion of the connected anode arm of the evacuated rectifier, the current was interrupted. This was evidenced by the fact that the anode arm ceased to glow.

Further tests led to the discovery that the effect produced depended upon where the experimenter stood in the static field conditioned by the circuit wires of the apparatus. He found that, when he stood in one position and touched the anode arm, the rectifier would be totally interrupted; sometimes also when the rectifier tube was under potential, but with no current flowing, if he touched the anode arm the current was thereby started; He concluded that this phenomenon could be made use of in connection with telephone repeaters or relays.

More experiments were made in the winter of 1909-1910, wherein the grid was connected merely to one side of the filament, that is, at the same potential as the filament. It was in these later experiments that he substituted, for purposes of comparison, the ion control for the ordinary receiver of the usual telephone system.

In 1911, Lowenstein constructed vacuum tubes of different shapes and dimensions to determine those most advantageous for his purpose. These tests showed that the ion control receiver was far superior to the results obtainable with the ordinary telephone receiver.

It may be noted that thus far in his affidavit Lowenstein made no reference to the ultranegative condition of the grid; but in a talk in 1911 between New York and Chicago, when he employed a condenser in the grid circuit, finding that the talk over the line was poor, the current through the controller tending to choke or stop, he discovered that the trouble could be remedied by touching the grid binding post repeatedly to discharge the grid and to open up the talk. He says: “This gave me the idea of connecting the grid to a point ultra-negative in potential relative to the filament.”

Nevertheless, the succeeding action of the Commissioner of Patents was to reject claims 1, 2, and 3, on the ground that they defined nothing more than the ordinary audion disclosed by De Forest, 995,126, or Stone, 884,-110.

As to the new claims, the Examiner made this interesting comment: “Claims 4, 5 and 6 bring out the idea of making the grid ultra-negative with respect to the filament. In the first place, experience has shown that in use the audion sueh as shown in the De Forest patent acquires a negative charge, due principally to the discharge of negative ions from the cathode. Hence there is a means in De Forest’s audion for making the grid ultra-negative with respect to the cathode.”

Then, as evidencing how the art develops through error, the Examiner says: “In Stone this situation was treated as a defect and was overcome by including a source in the grid circuit which rendered the grid positive with respect to the cathode,” and concludes with the observation: “It does not appear that it is of any material difference whether the grid be made positive or negative with respect to the cathode and it is not believed to involve invention to change the polarity of the source of the current z in the patent to Stone so as to render the charge upon the grid negative instead of positive.”

The Examiner then rejected claims 4, 5, and 6 on reference to the Von Lieben patent.

Following a personal interview with the Examiner in charge, the claims were canceled, and reference to the Examiner’s contention is thus made: “The Examiner’s position seems to be that in the arrangements illustrated in certain of the citations, a charge more negative than that of the audion cathode may now and then be impressed upon the intermediate grid by reason of its circuit connections.”

From the foregoing history of this application, I am persuaded that Lowenstein did not disclose a negative poling of battery 19 to the filament, and moreover, that what he sought and stressed was merely to make the grid more negative than the filament. That he accomplished by connecting each to the battery 21, as illustrated in the patent drawing. I believe also that defendant’s reading of claim 3 is sound.

For the conclusion that what Lowenstein set forth is not invention, the defendant urges, a number of reasons.

It is insisted that amplification by a vacuum tube with minimum distortion requires a grid bias, but that the value of that bias depends upon the degree of evacuation of the tube, the value of the voltage applied to the plate, and the geometry of the electrodes. The defendant contends that it is only because the modern tubes are so highly evacuated and are designed to be employed with so high a plate voltage, and because of the structural arrangement of the electrodes, that the grid bias to be applied should always be more negative than that of the most negative point of the filament; and the defendant argues that those conditions were not characteristic of the vacuum tubes available to Lowenstein and in nse either at the time of his invention or at the date of the filing of his application.

Though it may be admitted that a passage in the Lowenstein affidavit shows that he considered the geometry of the tube, there are expressions employed by Lowenstein in the specification which show that he was not concerned with a highly evacuated tube. This is particularly indicated in his description of the ionic controller. He says the incandescent filament produces ionization of the gas and makes the gas a conductor, and “the potential between filament 16 and anode 17 derived from the battery 21 creates a static field between such filament and anode which exerts a driving power on the ionized gas ■partidles which come in contact with incandescent filament 16 and which are thereby electrically charged * * * and the amount of negative current which is, therefore, conveyed through the field is dependent on the ionization power of the filament.” (Italics mine.)

Again: “The individual charge is a function of the negative potential of the filament and also of the degree of evacuation * * * Therefore the current in the ionic field circuit will increase with the incandescence of the filament, the potential difference of the battery and, up to a certain point, with the evacuation.” (Italics mine.)

All of these expressions would justify the inference that Lowenstein was dealing with a vacuum tube of low evacuation containing ionizable gas.

Defendant then argues that a grid biasing potential more negative than the “most negative point of the filament” was not suitable for the De Forest tubes of Lowenstein’s day, as further evidence of the fact that Lowen-stein did not teach that the grid biasing potential should be more negative than the "most negative point of the filament.”

Moreover, the defendant contends that the results of the tests described in the Low-enstein affidavit, filed during the prosecution of his application, in view of the tubes then in use, can be explained only on the assumption that the plus terminal of the filament heating battery was connected to the filament lead that runs down to the battery 21, so that its potential was opposed to the biasing potential, and that in effect the grid bias was more positive than the negative end of the filament.

It may be noted that both of the experts, Mr. Waterman and Professor Bowles, agree that about twenty-five or thirty volts were the average plate voltage for amplifier tubes of the De Forest gassy type; that is, those used at the time of Lowenstein’s invention.

In consequence, a study of Plaintiffs’ Exhibit Ho. 29 affords considerable interest. It is a chart of characteristic curves, plotted with respeet to the grid biasing potential of a three-electrode vacuum tube. It seems to be agreed by the experts in the case, that to get a signal substantially free from distortion the tube must be operated with a grid bias that corresponds to the middle point on the straight line portion of the characteristic curve. Interpreting the lower chart of this exhibit, it appears that, with the 200 volt curve, a zero grid potential is required to transmit without distortion the positive and negative voice of the incoming signal. Likewise it appears that, with the 250, 300, 350, and 400 volt curves, the grid bias in order to avoid distortion would have to be progressively negative; but that with the 150, 100, 50, and 25 volt curves the grid bias would be progressively positive.

The inferences deduced from a reading of these graphs were corroborated by tests made in the courtroom during the trial under the direction of defendant’s experts, Professor Bowles and Dr. Fcrrand. These courtroom demonstrations included tests on tubes of the old De Forest type, as well as modern tubes, and the showing seemed to be that whether the grid bias potential was to be positive or negative, with respeet to clarity and intensity of the reproduced signal, depended upon the plate voltage and the type and geometry of the tube.

Lowenstein may, nevertheless, have made an invention; hut it is in the light of this understanding of the action of a vacuum tube and in the light of the prior art that we must consider the nature of his contribution.

Patent Ho. 1,038,910 to Von Lieben and Reisz shows a grid bias intermediate between the positive and negative terminals of the filament, and indicates means for making this bias more negative than the average potential of the filament. This, too, depended for operation upon gas ionization, and one must deduce that Von Lieben and Reisz disclosed a causal relationship of a biasing potential on the grid to the filament. That being the ease, was invention involved in determining whether in a given tube the biasing potential lay positive, on the zero axis, or negative?

In this connection, the comment of the Examiner referred to hereinbefore that “it is not believed to involve invention to changa the polarity of the source of the current in the patent to Stone so as to render the charge upon the grid negative instead of positive,” challenges attention despite the fact that the Examiner clearly was wrong in contending that “it does not appear that it is of any material difference whether the grid be positive or negative with respect to the cathode.”

But it seems to me it does follow that if the art showed there was a causal relation between the potential of the grid and the potential of the filament, there could hardly be invention in selecting a proper bias, unless at the same time it was pointed out how that bias was to be determined by the characteristics of the tube. .

Mr. Justice Stone, in De Forest Radio Co. v. General Electric Co., 51 S. Ct. 563, 569, 75 L. Ed. 1339, in his opinion concerning the validity of the Langmuir high vacuum tube patent, said: “There was little or no practical use for a high vacuum tube in 1913. The De Forest audion was not in general use and Langmuir did not see one until that year.”

This offers additional reason for believing that the defendant is correct in urging that, at the time of the filing of the Lowen-stein affidavit in support of his invention, Lowenstein at most was referring to the low vacuum tube which the commercial art then knew.

It is for the foregoing reasons that I believe the Lowenstein contribution, however valuable it may subsequently have proved to the art, was not invention.

The next patent to be considered is that to Robert C. Mathes, No. 1,426,754, issued August 22, 1922, on an application filed October 23, 1916. This patent relates also to' improvements in circuits for electron discharge devices.

Only claim 25 is in issue. It reads as follows: “In a vacuum tube amplifier circuit wherein the space current of said amplifier depends upon the voltage of a variable source of current connected to an anode and a cathode, the method which comprises supplying said circuit with at least two impulses, one a compensated impulse derived from an effect produced by variations of said source and the other due to an impulse to be amplified.”

Claim 8 of the patent was held to be valid and infringed in Radio Corporation et al. v. J. H. Bunnell & Co. (D. C.) 22 F.(2d) 847. The subject-matter of claim 8 is disclosed in Fig. 1. The voltage drop across the resistance 9 in the grid-filament circuit maintains on the grid a potential more negative than any part of the filament, but because the defendant’s grid biasing resistances are shown each connected in a path coming to the grid and the plate circuits instead of between the grid and the filament, infringement of this claim is not alleged. It is the compensating aspect of the Mathes invention, therefore, that is^ involved in this ease. The inventor states: . „

“A further object is to furnish a method and means for compensating for fluctuations in the potential of the output circuit battery of the vacuum tube.”

“Specifically these objects may be realized by the indirect use of the heating battery or of some other source, but in either case associating said source with the output circuit battery.”

Mathes said that it was well known” that current flowing through the output circuit undergoes fluctuations owing to changes within the power battery. It was the difficulties caused by these fluctuations in the space current of the vacuum tube that Mathes sought to overcome. The method he proposes, he explains as follows: “Assume that the vacuum tube circuits are so arranged, as hereinafter described, that the current for the input circuit comes from a source in series with the source for the output circuit. For illustration, suppose there is a decrease in the potential from the output circuit battery due to some internal change; this will necessarily tend to decrease the space current between the filament and anode of the vacuum tube. If the input and output batteries are of similar nature, this change will in all probability also decrease the potential of the input circuit battery, which is in series with it, so that the grid potential will become less negative thereby tending to increase the space current to its former value. The compensation would take place in a similar manner for an increase in the voltage of the batteries, due, for instance, to the batteries being charged by some suitable means.”

The circuit defined in claim 25 is illustrated in Fig. 3 of the patent. The negative grid voltage is supplied from the drop of potential across the terminals of the resistance 9. This drop is due to a source of voltage 7, associated with the battery 5 of the ■output circuit. To prevent the grid from reacting to sudden fluctuations in the battery 7, a condenser 18 and an inductance 16 are provided, the latter inserted in series with the battery. Fluctuations in battery 5 are made by a condenser 20 and inductance 21. Provision is also made for a condenser 17 to be shunted across the resistance 9 to provide a path for the oscillations in the input circuit. A vacuum tube 15 is employed as a resistance, having a heated filament 27 and an anode 28.

Batteries 7 and 5 of Fig. 3 are of the same type and are subjected to the same sorts of fluctuations. The inventor says: “If, for example, these batteries are dry cells, they will gradually drop in voltage as they are used and, as explained heretofore, a gradual drop in the voltage of the output battery will be compensated for by a similar drop in the input battery. If these batteries are storage batteries, there will be a gradual drop in voltage of both input and output batteries during discharge.”

From the foregoing description, it should appear that the two impulses referred to in the claim are impulses in the plate circuit. The compensated impulse referred to in the claim is the plate voltage. The uncompensated impulse, that is, the impulse to be amplified, is the signal. Hence Mathes’ method is to supply to the plate circuit compensated source variation impulses and uncompensated signal impulses.

The defendant contends that claim 25 is invalid because of anticipation, and urges noninfringement on the ground that his arrangement was derived, not from Mathes, but from the prior art.

In connection with the defense of invalidity, the defendant endeavors to read claim 25 so broadly as to bring it within the disclosures of the prior art, particularly those of Yon Lieben. The interpretation thus offered is sought to be supported by a reference to another Mathes patent, No. 1,493,-216, on application filed in February, 1919. It was in this later application that claim 25 of the patent in suit originated. In transferring the claim to the application which resulted in the patent in suit, Mathes stated: “New claims * * * are the same * * * as allowed claims 12, 13 and 19, application serial No. 277,566 filed February 17, 1919. It is believed that the Examiner will have no hesitancy in allowing these claims in this application, since at times when generator 19 is connected to the circuit, the energizing current for the tubes is derived from this source.”

Whether, as originally filed in the later application, this claim had a wider connotation seems doubtful; but at any rate the claim reads certainly on Fig. 3 of the patent in suit, and the specification affords a sufficient disclosure of it.

I think the claim essentially defines a method. It is not limited to any specific arrangement or apparatus by which the method may be practiced. Therefore, I reach the interpretation of the claim, as hereinbefore set forth, which in effect is the plaintiffs’ version thereof.

Now, in respect to the prior art, the defendant relies on the Lowenstein patent in suit, on an article in the Eleetrotechnisehe Zeitsehrift by Eugen Reisz of November 27, 1913, the Yon Lieben and Reisz patent, 1,-038,910 of 1912, Langmuir, 1,273,627, and Colpitts and Arnold, 1,388,450.

First as to Lowenstein. In urging this patent as an anticipation, the defendant points out that in Lowenstein the current for the input circuit comes from one part of the battery 21, and the current for the output circuit comes from the same battery, and concludes that thus the source of the input circuit current is in series with the source of the output circuit current.

But it does not appear to me that the two sections of the Lowenstein battery are operative in series. One end supplies the grid, and the other the plate. Mr. Waterman explained that the grid does not draw current, but the plate does. He concluded that the two parts of this báttery 21 would not run down together and would not be subject to corresponding voltage fluctuations. It is possible, of course, that the sections might run down together, and then there would be some compensation. Both compensating actions of Mathes are not found in Lowenstein, even if one admits such accidental eompen-sating result. For example, the more exact compensation attained by Mathes by means of the two element vacuum tube 15'is in no way indicated in the Lowenstein circuit. The impedance of this tube decreases as the current passing through it increases. A current change through the circuit beginning with the right-hand section of battery 7 through resistance 9, impedance 15, and the inductance 16 to the opposite end of the battery will be greater than the voltage change which produced it. At the same time that this occurs, the ratio of the resistance 9 to the resistance of the tube 15 changes, thereby altering the proportion of the voltage of battery 7, which is made available for biasing, and thus increasing the compensating effeet.

The Eleetroteehnische Zeitschrift article by Eugen Reisz may be discussed also in connection with the Von Lieben and Reisz patent. It appears that Von Lieben and Reisz disclosed the normal grid bias as more positive or less negative than the filament, but Professor Bowles argues that there is associated with the output of the B battery a filament-heating and grid-biasing battery, which biases the grid by a drop in potential across a resistance. He infers that this battery and resistance correspond to the filament-heating and grid-biasing battery 6 and the resistance 9 of Mathes. He admits that the Von Lieben tube, shown in Fig. 4 of this article, discloses a desirable grid bias more positive than the negative side of the filament, while in Mathes the grid bias was more negative than the negative side of the filament; but the defendant claims that claim 25 embraces both of these arrangements.

As has been said heretofore, I think this interpretation of claim 25 is not justified, and, in consequence, that the reference to the work of Von Lieben and Reisz does not show an anticipation of this claim. In other words, to make these references effective' would compel such a broad reading of the claim as seems to me to be wholly unwarranted.

The references to Langmuir, 1,273,627, and to Colpitts and Arnold, 1,388,450, are mainly for the purpose of showing in the prior art a biasing resistance in series with a source of plate current and supplying a compensating negative potential to the grid.

The Langmuir patent' is for a method of and means for controlling potentials. This patent is one of considerable interest. Langmuir, in his specification, says: “The greater the current the greater the drop in potential in the resistance 13 and the greater the difference in potential between the grid and cathode. Thus as the current increases, the grid becomes more negative until a current value is reached at which further increase is prevented by the negative potential on the grid.”

Professor Bowles was of opinion that in the Langmuir arrangement a variation of potential of the plate battery would be accompanied by the Mathes compensation effeet on the grid; but the defendant seems compelled to make the admission that the Langmuir patent in itself is not a disclosure of claim. 25 of the patent, and for that reason urges that it would require no invention to use the-Langmuir arrangement for securing a compensating negative grid bias, because a like-arrangement for the specific purpose is shown, in the Colpitts and Arnold patent. This latter patent, though, makes no provision for permitting the signals to pass uncompensated.

The defendant complains that claim 25 does not contain the limitation that the signal impulse is not to be subjected to the compensating effeet. I think, however, the claim both in language and in effect means that the signal impulse is not to be subjected to the compensating effeet. The claim refers to two-impulses. One it specifically designates as a. compensating impulse, and “the other due to an impulse to be amplified.” The latter, of course, must refer to the signal, and the-specification as to that, says: “Condenser 17 may be shunted across 9 to serve as an easy path for the oscillations in the input circuit.”'

And Mr. Waterman said:

“The arrangement of fig. 3 provides for-avoiding this difficulty in two ways: first, the alternating current in the plate circuit is,, so far as practicable, kept out of the battery altogether. It is caused to pass through the condenser 12 and the primary of the- transformer 13, and any current flowing in the grid circuit — the input circuit — coming from the transformer 8 is prevented from passing through the resistance 9 by the shunting of the resistance 9 with the condenser 17.”

“The signal current is a rapidly, varying current, varying from say 50 to 5,000 times per second. Such a current will pass readily through a condenser, while the battery current will not pass through a condenser at all; therefore, by putting a condenser such as 17 around resistance 9, a passage is provided for the alternating current, while the direct current flow from the battery is compelled to pass through the resistance 9 and the resistance 15; therefore, the compensating effect is experienced as regards the direct-current supply sources of the system, while no such effect occurs to prevent each tube giving its full amplifying power for the signal.”

“It is this combined aspect of the Mathes invention that I understand is suggested to the engineer by the language of claim 25; and there are two classes of impulses that we have to deal with here: one is the power source for the plate circuit and the source of voltage for the grid circuit. These it is desired to compensate for fluctuation. It is not desired to compensate' out the signal current, yet the signal is likewise supplied to the circuit. The idea is, therefore, to differentiate between those two, compensating one and not the other.”

I conclude, therefore, that Mathes made a valuable contribution to the art, that it was the result of invention, and that claim 25 is valid.

The defendant seeks to avoid infringement on the alleged ground that his biasing resistance is in series with, and derives its current from, the source of the plate current itself; that such arrangement is not shown or suggested by Mathes, but on the contrary is disclosed, as the defendant claims, by Col-pitts and Arnold.

It will be necessary to refer to the wiring diagram of the defendant’s circuits. The grid biasing and plate voltages for the tube Y2 pass upwardly from the filter to the coil P3, thence to the plate, across'the space to the filament, down from the filament to the coil S9, then up through the resistance R2, thence to the left, and down to the other end of the power supply. The flow of current through the resistance R2 causes a drop in voltage and impresses on the grid a potential more negative than any part of the filament, the opposite end of R2 being connected to the filament.

Mr. Waterman contends that the resistance R2 of the defendant’s circuit corresponds to the resistance 9 of the Mathes patent. This resistance is in series with the tube Y2 itself, and to that extent Y2 corresponds with tube 15 of Mathes, which also is in series with resistance 9.

In the defendant’s system, as the power supply voltage varies, the direct current voltage in the plate circuit of the tube Y2 will vary. Such variations must cause variations of the normal current flowing through the tube, and will also cause the resistance of the tube to vary. Increase of current through V2 increases the current through R2, since they are in series, and that in turn increases the voltage drop across R2, and increases the negative bias on the grid. This tube Y2, therefore, in the defendant’s circuit and device performs the function of amplifier and regulator of the flow of current.

Thus in defendant’s circuit there are two compensating activities; and, as in Mathes, provision is made to let the signals pass without themselves being compensated. This function is performed by defendant’s condenser C, and is the equivalent in function of the condenser 17 of Mathes.

It seems to me that the defendant could avoid infringement only by successfully pressing his interpretation of claim 25. Inasmuch as I cannot accept that view, I And claim 25 infringed.

The next patent to consider is that to Colpitts, No. 1,128,292. This patent is perhaps the most interesting of those involved in the litigation. The invention relates to electric wave amplifiers, and particularly to the use of vacuum discharge repeaters for repeating and amplifying, in an output circuit, waves of electric energy received in the input circuit. Here, again, the inventor’s object was to free these waves from distortion.

Colpitts found that a single audion repeater does not repeat in the output circuit a perfect form of reproduction of the electric wave received in the input circuit, and he ascribed that result to the fact that the space between the filament and the grid has unilateral conductivity. Thus, unlike Lowen-stein and Mathes, the distortion which Col-.pitts sought to eliminate was that which arises in the tube itself, as distinguished from one arising in the circuits.

It seems that Colpitts, not unlike Lowen-stein, was not fully acquainted with all the theoretical conditions under which the audion operated. So much was admitted by Mr. Waterman. But however imperfect his theory may have been, he sought and obtained a desirable objective in eliminating the distortion created within the tube. Colpitts evidently thought that the audion would repeat only one-half of the signal wave; that is, the positive half wave. Apparently Colpitts knew only the soft tubes. Such tubes as have been indicated could not be operated at higher plate voltages, and of them it is true that, when operated at the lower plate voltages and with zero grid bias, they would much more effectively repeat the positive half wave than the negative half wave of the signal. Colpitts, therefore, sought to have both half waves repeated in such fashion that each half necessarily derived from the grid would disappear in the final amplified product. In effect, Colpitts succeeded in curing the grid defects of one tube by the grid defects of a second by placing the input or grid circuits of the two tubes in divided input relation.

Fig. 1 of the patent represents the system of circuits employed. Two audion elements are combined and included in each of the circuits. Incoming and outgoing conductors are represented by the wires 1 and 2 terminating in the primary winding 3 of an input repeating coil 4, and the wires 5 and 6 terminating in the secondary winding 7 of an output repeating coil 8. The secondary winding of the repeating coil 4 is divided into two coils 9 and 10, and the primary winding of the repeating coil 8 is divided into two coils 11 and 12. The audions comprise two heated filaments 13 and 14, two plates 15 and 16, and two grids 17 and 18. The filaments are heated from the single battery 19, shunted across their connected terminals. In the output circuit, the primary coils 11 and 12 of the repeating coil 8 has included within it a battery 21.

The arrangement of audions in this Col-pitts patent led to its designation as a “push-pull” arrangement. It is the contention of the plaintiffs that, though the term “push-pull” was known to the art, the invention of Col-pitts has given it a new and specific meaning. In effect it means, as they contend, the arrangement of two three-electrode vacuum tubes with divided input and divided output circuits.

Claims 1 and 5 are in issue. They read’// as follows:

“1. An electric wave repeating apparatus comprising divided input and divided output circuits, means for producing two ionized streams, said streams being oppositely included in said output circuit, and two electrodes for controlling said ionized streams respectively and oppositely connected in said input circuit.”

“5. In an electric wave repeating apparatus, the combination of two repeating elements comprising ionizing means, an anode and an input electrode for each said element, an input repeating coil the secondary winding of which is connected at an intermediate point to said ionizing means and at its opposite terminals to said input electrodes respectively.”

The defendant resists Colpitts’ claims on the ground that they are not the result of invention. It is contended that Colpitts stated no new combination and no new mode of operation. It is said that the only thing Col-pitts did was to substitute a new type of telephone repeater, i. e., the audion, in an old telephone circuit, and that the whole and only novelty of the arrangement of Colpitts lies in the audion repeater, and not at all in the combination.

The defendant in support of this proposition relies first on the push-pull telephone circuits shown in the prior art patents of Dean, Kitsee, Stragiotti, and Grissenger. The defendant’s position is that this part of the prior art shows two mechanical .telephone repeaters connected baek to baek and fitting into a common output transformer, and that the operating function of such arrangement was to tend to balance out in the input transformer the'distortions produced by the imperfect operation of the two repeaters, respectively.

One’s first reaction to these mechanical repeaters is that they have no element corresponding to the grid of the audion, and necessarily, therefore, show no provision for connecting such elements in a divided input circuit.

Mr. Waterman said'of the devices of these patents that they have this in common: “That they depend upon compression and decompression, either by air waves or by the vibration of a diaphragm, or reed, in response to telephone currents applied to a magnet, so that the resistance is simultaneously increased on one side and decreased on the other side of the vibrating member.”

'Mr. Waterman contends, and I think with much force, that as illustrated by Professor Bowles’ oscillograms, the current brought about by compression is quite different from the current brought about by release of compression. I think it is too much to contend successfully, as Professor Bowles seeks to do, that the double opposed carbon button devices resulted in cancellation of distortion in individual waves by virtue of the back to baek arrangement through a common transformer circuit or output.

In his discussion of the carbon button repeaters of the prior art, Professor Bowles made much of the Shreeve repeater; but I think the result was unfortunate. It seems to me that the cross-examination on this Shreeve patent made most convincing the proposition that the substitution of audions for carbon button repeaters required invention.

The Shreeve repeater, as appears from the Gherardi and Jewett Paper, in the early days was a useful device and, perhaps, the best repeater available until the audion was developed. These writers, however, point out that “an inherent defect in this type of instrument is the fact that the sensitiveness falls off rapidly when the input energy is below a certain minimum. * *' * Thus at one moment the repeater in a given circuit may be handling the energy from two persons speaking loudly, using telephones very near the terminal of the circuit, while at the next moment it may be called upon to function with the energy which it receives from telephones connected to long circuits, which in turn connect to the circuit with which it is associated. If the persons using the telephones in this latter ease have weak voices, if they do not use the telephones properly by talking directly into the transmitter, or if the apparatus or lines themselves are inefficient, it is clear that the energy at the repeater terminals may be many times less than that in the previous ease. Further under those conditions maximum amplification is the more to he desired so that a repeater element whose performance fluctuates with the amount of energy which it receives and particularly one whoso performance is worse when operating with the smallest energy is subject to a handicap when it comes to applying it generally in the telephone plant.”

Now, Professor Bowles took the Shreeve patent, No. 1,156,636, and sought to place two of these carbon button repeaters back to back, with the thought, doubtless, that all Colpitts had to do was to substitute two audi-on elements for the Shreeve repeater. What he did was extremely ingenious; but Mr. Waterman pointed out that the Bowles’ arrangement would be inoperative because the signal could not get to the carbon button elements at all, as the condensers, which are essential to the operation of the Shreeve repeater, when put together, would short circuit the audion frequency current; and moreover, that the circuits are such that, even though alternating currents were produced in the elements 26-26, those currents would be short circuited within the output coil.

Patents to Hewett, No. 749,791; Taylor, No. 953,361; and Arnold, No. 1,118,173, form another group of patents relied upon by the defendant in this art to prove lack of patentability by Colpitts. These patents relate to mercury vapor are devices.

Gherardi and Jewett wrote: “This arrangement, which is essentially a very insensitive unilateral repeating device, has never, so far as is known, been made to function as a telephone amplifier. In addition to its inefficiency, this arrangement is undesirable because of noise, distortion and variable amplification.”

Of the Arnold arc, Gherardi and Jewett wrote: “This gaseous device is capable of good amplification and is fairly free from distortion except such as may be introduced by the fact that impedance of the electromagnet may not be identical with that of a telephone line. It was tried out experimentally on telephone lines but has never been used for any length of time and then only under special engineering supervision. For example, units were installed on the important long distance circuits between New York and San Francisco but were never used commercially.”

Of the Taylor device, as well as that of Arnold, it may be said that it shows no divided input control. The control elements of Colpitts are missing, and it does not appear that the two input circuits could be opposed to remedy the inaccuracies of their control effects.

In brief, it would appear that the mercury vapor arc devices do not anticipate the Colpitts patent.

Finally, in the discussion of the prior art of the Colpitts patent as indicating a third type of reference, it is necessary to consider the Gerdien patent, No. 1,004,012. This patent shows a Geissler tube. This tube requires a high voltage for operation, and produces a characteristic cathode dark space. It has no heated filament. Gerdien sought to control the dark space by causing the current carriers to adopt spiral paths. Changes in these carrier paths changed the resistance within the tubes. The Gerdien tube is a strictly two-element tube. It shows no unilateral conductivity, and the batteries, therefore, may be poled either way. There is no divided input to the control element.

I have great difficulty in seeing how the Gerdien tube suggests the Colpitts invention. Gerdien was apparently interested in controlling or counteracting the sluggishness of the Geissler tube. A Gerdien tube is different in construction from that of the three-electrode vacuum tube. It certainly does not anticipate the claims of the Colpitts patent.

I am of opinion, therefore, that the Col-pitts patent is valid and infringed.

The patent in suit, No. 1,504,537, to Arnold, technically viewed, presents the greatest difficulties. The two scientific opinions, opposed almost on every fundamental matter involved, are in the utmost discord.

The patent is entitled for an improvement in power limiting amplifying devices. The original application was filed September 3, 1915. It was allowed November 18, 1920. On May 15, 1921, the application was declared in interference with an application of Dr. Alexanderson; on February 9, 1922, the application was allowed. On June 28, 1922, the applicant presented a petition for the entry of an amendment. The petition for the filing of the proposed amendment was, on June 30, 1922, denied.. Another petition for amendment was submitted on July 31, 1922. On August 2,1922, another petition for leave to amend was filed. On August 7, 1922, an office action was adverse to the amendment, but contained a statement that proposed claims 31, 32, 36, 37, and 39 appeared to be allowable. Then apparently the application lapsed, but was renewed on September 14, 1922, and finally issued on August 12,1924.

The foregoing detailed reference to the history of the case is made necessary because of the defenses which are vigorously asserted.

The claims involved are in two groups. Claims 17, 18, and 20 form one group, and claims 33 to 36, inclusive, the second. The former group was referred to 'throughout the trial as the “high impedance” claims, while the latter group was designated the “push-pull loading” claims.

As to the former claims, defendant’s position is that they are invalid for lack of novelty ; and that the defendant does not infringe because defendant’s system is derived from the prior art, and not from the disclosure of the Arnold patent.

Claims 17, 18, and 20 read as follows:

“17. An amplifier comprising two electrical discharge devices, input and output circuits therefor, said devices being symmetrically and oppositely disposed with respect to each of said circuits, space current paths for said devices, said paths having a common external portion of high impedance.”

“18. An amplifier comprising two electrical discharge devices, input and output circuits therefor, said devices being symmetrically and oppositely Arranged with respect to each of said circuits, spaee current paths for said devices, a source of space current therefor, said paths having a common external portion, and means associated with said source for preventing fluctuating currents therethrough,”

“20. A repeater comprising two electric discharge devices each having a cathode, an anode and an impedance varying element, a connection between said impedance varying elements, a connection between said cathodes, a connection between said anodes, an electrical impedance in one of said connections, an electrical impedance in another of said connections, other connections from intermediate portions of each of said impedances to one of said first mentioned connections, one of said last named connections being provided with a source of current, and means tending to maintain a constant supply from said source.” .

The Arnold patent discloses a radio receiver consisting of antenna, a radio frequency amplifier stage, a conductor stage, and an audion amplifier. So far as this suit is concerned, it is sufficient to consider only the audion amplifier.

The third or output stage of this amplifier has two tubes arranged with a divided input and divided output circuits such as is shown in the Colpitts patent. This final stage is preceded by means of a single tube amplifier. The specification states:

“The thermionic 'repeater being unilaterally conducting, the repeater element 48, 49, 53 can transmit positive current due to battery 65, only in the direction from 53 to 48. Also, element 50, 51, 52 can transmit positive direct current only in the direction from 52 to 50. If these currents are approximately equal, it follows that the maximum variation in current around the circuit 48, 53, 54, 52, 50 can never exceed the magnitude of the normal current in either element, provided none of this varied current can pass through that part of the circuit of the battery 65 which is common to the two thermionic elements. To prevent any appreciable passage of variable current'the high resistance 58 is used.”

“The variations in the normal currents in the winding 55, which variations constitute the signals to be received, are produced in the usual way by the action of the grids 49 and 51, across which the signal voltage is impressed, so that it is obvious that an impressed voltage of large value, tending to produce a large variation of current in the power limiting device, cannot cause an altex*-nating or varying current in winding 55 larger than