The bill alleges infringement of letters patent No. 1,019,482, granted March 5, 1912, to Patrick Kennedy, on application filed March 17, 1908, for an improved method of charging storage batteries used in car-lighting systems wherein the generator is driven by the car axle, having means for maintaining its output con*128stant, and being, provided with a predetermined voltage relay to control the regulator for automatically charging the battery to its full capacity and then protecting it from excessive overcharge.

The predetermined voltage relay E1, to which reference is hereinafter made, is asserted by defendant to function substantially like that described in the prior patent to Creveling, which was heretofore considered by this court and held valid. Safety Car Heating & Lighting Co. v. U. S. Light & Heating Co., 222 Red. 310. And it is also asserted that the Kennedy patent in suit was intended as an improvement thereon.

Before Creveling, the car-lighting art had adapted means for regulating the generator current, either for constant potential or constant current, and the existence of such different systems was recognized by the patentee. Creveling’s invention was the addition to the constant current system of a battery protective device consisting of a potential stop charge relay placed across the battery terminals, so as to make it responsive to back voltage of the battery during the charging. When the battery voltage rose to a predetermined maximum, say 42 volts on a 16-cell battery, the relay, which until then had been inert, became energized, and the circuit was closed by a regulating coil, and owing to resistances in the circuit the dynamo output was decreased, the electromotive force of the battery was opposed, and overcharging prevented. Referring to the object of the invention and to prior lighting systems generally, the specification states:

“In railway car lighting, various means have been proposed to prevent an overcharge of the storage battery. These systems which have been most widely used in practice are (a) the system in which the generator is regulated to produce an electromotive force of constant value, so that the battery is charged at constant potential, and as the electromotive force of the battery rises it opposes that of the generator, thereby preventing an overcharge; (b) the system in which there is inserted between the generator and the battery a circuit breaber which is operated to cut off the charging current whenever the potential across the battery terminals reaches a predetermined maximum.”

There were numerous objections to the prior devices which were adjustable for regulating the generator to a constant potential or a constant current, arising, principally, from variations in potential or resistances originating within the storage batteries from the back voltage, which made it difficult to apply the usual tests for determining at any time the condition of the battery charge. These difficulties were of serious import, and the patentee designed to surmount them by adding to the system a meter device which in the early part of the charge delayed the action of the predetermined voltage relay and in the latter part of the charge, or after a slight overcharge — after charging to its full capacity regardless of internal variations — functioned automatically to prevent a long-continued overcharge of the battery. The object of the meter or time-measuring device was to ascertain and record the condition of the battery charge.

In describing the details for operating the device the specification refers to a plurality of small magnets, each controlling a star wheel and separately connected in series with one of the solenoid coils (Figs. B', C', D, and E') comprising part of the meter and functioning to *129increase the dynamo output to permit an uncounted quantity to go to the battery. Such uncounted or additional charge, complainant contends, .was a corrective factor, to make up for battery losses from leakage or other causes; but whether it did so in fact is questioned by the defendant. When the system is in0 operation, the lighting of the lamps closes the lamp switch, and thereupon the circuit of the small magnets also closes, with the exception of one, which becomes excited and brings into position its star wheel together with another star wheel —one star wheel indicating battery charge and the other battery discharge. To insure clearness of description, I quote from complainant’s brief:

“The lamps are divided into banks which, in the particular instance chosen, are assumed to be of 10 amperes’ consumption in each hank, so that, when any bank switch is closed, thereby drawing 10 amperes of current for the lamps, the corresponding magnet 5, 6, 7, or 8 is energized and its star wheel is given one reverse actuation in each unit of time. The effect is, of course, to indicate a discharge of 10 amperes. It is thus apparent that the construction of the meter is such that, for each unit of time during which the generator is acting to charge the battery at the normal rate of 30 amperes, the meter will indicate a charge of 30 amperes going into the battery, and for each unit of time during which any bank of lamps is burning, the meter will indicate a discharge of 10 amperes. If the generator is charging the battery when no lamps are lighted, the indication of charge will properly be 30 amperes. If lamps are lighted when the generator is not running, the indication of discharge will properly be 10 amperes for each lamp bank.”

The star wheels are connected with the magnets to comprise four separate lamp banks; each bank taking an equal quantity of current. When the lamps are lighted the dynamo supplies unmeasured current to the storage battery to equalize for losses, and the pointer travels forward three steps, indicating a charge of 30 amperes to the battery and each reverse step a discharge of 10 amperes.

The corrective factor interferes, no doubt, with determining the precise condition of the battery charge, and in this respect complainant asserts its device is not dissimilar to defendant’s; but, however that may be, the principal object of the inventor was to provide a device for ascertaining at least the approximate state of the battery at any time. Did he succeed in producing a meritorious invention? Has he solved the problem of easily determining the condition of the battery charge — a difficulty due to varying load conditions? And has he devised means for preserving the maximum efficiency of the battery?

The defenses are limitation of claims, inoperativeness, and nonin-fringement as to both systems marketed by defendant, viz. the so-called standard and double relay systems.

The claims relied on are 1, 2, 3, 4, 7, and 8, but it will suffice to set forth the first and third, representing the method and apparatus claims respectively:

“1. The method of charging storage batteries which consists in supplying a charging current to the storage battery for a predetermined number of ampere hours regardless of the electromotive force of the battery, and thereafter causing a predetermined maximum potential difference across the battery terminals to discontinue the charging current, substantially as described.”

“3. In a train-lighting system, a generator driven from the car axle, a stor*130age battery connected to said generator to be charged thereby, and mechanism for regulating the generator to a constant current output, in combination with a controlling device in said circuit which discontinues the charging current when the potential thereof reaches a predetermined limit, and mechanism for rendering the controlling device inoperative until the battery had been charged to a predetermined number of ampere hours, substantially as described.” •

In claim 1 two definite steps for practicing the invention are set forth, namely: (1) Charging the battery for a predetermined time, regardless of the amount of current flowing thereto; and (2) stopping the charging current to'cause a predetermined maximum to flow across the battery terminals. The first step provides for regulation of current supply to the battery by a meter device, the voltage relay F 1 being dormant; while the second step requires that switch F be closed when the charge is nearly completed, to permit relay P1 to come into activity when the back voltage of the battery rises to the stage for which the relay has been set, namely, “the predetermined maximum potential difference.” There was evidence showing that when P 1 was set at a high voltage, and the battery was not co-operating with the meter, the action of coil F 2 was delayed until switch P became excited, but the charge then was quicker than if the 'relay F 1 had been set at a lower voltage-.

There was divergent discussion among the experts as to the meaning of the method claims; that is, as to whether relay F 1 acted like the voltage coil in Creveling’s patent for shutting' off the charge to the battery or whether coil F2 necessarily performed that function. I think that, after the meter contacts were closed in Kennedy’s structure, the charge was indebted to the effectiveness of coil F 2, which, as said, remained inactive until the charge was brought up to a predetermined quantity. Any limitations of the claims in controversy must find justification in the evidence relating to the inoperativeness of complainant’s device. Whether or not such testimony complies with the standard of proof required to overcome the prima facie presumptions arising from the grant will receive consideration hereinafter.

Claim 3, relating to the apparatus, includes the combination for driving the dynamo, the storage battery, means for regulating the generator to a constant current output, a stop charge relay or controlling device, and means for making the controlling device inoperative until a predetermined number of ampere hours have been let into the battery. Claim 4 is substantially the same as claim 3, except that it refers to a movable indicator; while claim 3 refers to the meter as a controlling device. Claim 7 is limited to traveling mechanism and a specific relay having a “second circuit controlling a device responsive to the difference of potential across the battery terminals” for stopping the charge.

Complainant contends that the principal claims herein are entitled to a broad construction, and that, when so construed or interpreted, the standard and double relay devices employed in defendant’s structures are wrongful appropriations; that an analysis of the evidence unmistakably discloses that the Kennedy meter actually measured the battery charge, and not merely the energy required for completing *131the charge; and that corrective allowance is made for inherent battery deficiencies. Defendant rejoins that the claims broadly include a lighting system and improvements which were not invented by the patentee; that, if their wording is descriptive of its structures, the invention was nevertheless incomplete and inoperative when the patent application was filed.

[1] Although the inventor had in mind overcoming serious difficulties to which the earlier battery protective devices were subject, still I think his idea related more specifically to the constant current regulator of Creveling, in which the back voltage of the battery was responsible for actuating the predetermined voltage.relay, rather than to the constant voltage regulator wherein the back voltage was not regarded as a material factor in discontinuing battery charge. In discussing the inoperativeness of the Kennedy meter, it is well to have in mind the law succinctly stated by Judge Hough in Engineer Co. v.. Hotel Astor (D. C.) 226 Fed. 779, to wit, that a patented device need not be perfect in order to escape the charge of being inoperative and incomplete, even though the later structure is the better of the two. In this case, however, it is necessary to ascertain whether the patented device was incapable of accomplishing the object of the inventor. If the patentee designed to construct an ampere hour meter, and it is evident that an operative one cannot be constructed from the teachings of his patent, its utility is negatived; but, as I understand the law, the patent will escape invalidity if the device is operative to a limited extent only, and will be entitled to a narrow construction, excluding a later device for accomplishing the result by an essentially different mode of operation. Herzog v. N. Y. Telephone Co. (C. C.) 172 Eed. 425, affirmed 176 Fed. 349, 99 C. C. A. 623; Westinghouse v. Boyden Power Brake Co., 170 U. S. 537, 18 Sup. Ct. 707, 42 L. Ed. 1136. In Lovell v. Seybold Match Co., 169 Fed. 288, 94 C. C. A. 578, Judge Coxe clearly states the rule which in my view may be applied herein. He says:

“The claims must be construed in the light of the contribution which the patentees made to the art. They should hold whatever of value they have added provided it involved invention to make the addition. It would, however, be grossly unfair to compel the builder of a practical working machine to pay tribute to one who has added nothing o£ substantial value to the art, simply, because the language of his claims is broad enought to include the successful structure. Claims should cover what the patentee has invented and not what he imagines he has invented.”

Defendant’s witnesses Ranphier, Robinson, and Sheldon, concededly having wide experience in the car-lighting art, have positively sworn that the Kennedy meter could not possibly indicate the state of battery charge, owing to the absence of an element responsive to the current flowing into and from the battery; that an automatic indication was impossible owing to the necessity for manually setting switches for co-operation with variations in lamp conditions; that in an ampere hour meter for car lighting it was important that there should be embodied in the structure an element responsive to current changes in the circuit and means for instantaneously multiplying the current by time; that the Kennedy device was inoperative and impracticable, *132and could not record ampere hours to the battery or indicate approximately battery conditions. Lanphier, however, on cross-examination, seemed to agree that it would register correctly during regulation at 30 amperes-as long as the current consumed amounted to 10 amperes per lamp bank. Mr. Lunn, chief electrician for the Pullman Sleeping Car Company, admitted that the device would operate when applied 'to cars having center lights in banks.

There was testimony of a highly technical character relating to the corrective factor. Charts were introduced in evidence of curves showing the percentage of error based upon actual amperes to the battery, which were then contrasted with the indicated input to the battery. It is unnecessary for me to treat at length of the charts. Those upon which defendant relied were criticized because they did not consider the proportional quantity of unmeasured current required by the lamps to compensate for battery leakage and other causes. The club car chart disclosed a closure of the meter contact Fz (Elkhart stop to Toledo and Harmon stop to New York) closing the relay E1 and controlling it by the potential of the battery; the charge meanwhile continuing. But it is not improbable that this condition was due to failure to properly set the principal relay, which if it had been set to close circuit E2 of the regulator would have resulted in stopping the charge. Although counsel for defendant point with confidence to the charts as corroborating the testimony relating to inoperativeness of complainant’s patent, I think the testimony relating thereto is open to the construction that the Kennedy method and apparatus, when contrasted with the modern ampere hour meters, are not as practical or useful.

[2-4] There were cited by defendant the prior patents to Wray and Taylor to show that the Kennedy method, viewed broadly, was invalid for lack of novelty; but these patents were granted subsequent to the Kennedy patent, and, though the applications were filed first, they are not anticipatory, and can be considered only to prove prior inventions. Sundh Electric Co. v. Interborough Rapid Transit Co. (C. C.) 222 Fed. 334. As to these devices, however, it is enough to state that Wray lacked the combination of the claims in suit, while Taylor’s-device was not in the train-lighting art. It is true that means in the form, of ampere hour meters for controlling storage batteries in connection with stationary regulators were old; but the devices in which they were applied indicated battery charge only, without preventing overcharge. Kennedy took steps towards complete^ solving the problem of battery protection, but his construction did not go far enough, and its deficiencies and limitations were at once perceived. It was incapable of indicating the number of ampere hours mainly because of the inclusion in his system of lamp banks having a plurality of magnets, star wheels, and supplementary coils on the regulator. He also failed to provide a suitable element for bringing together parts for measuring the current and creating responsiveness to battery variations. He applied for and obtained a second patent embodying a single magnet for operating a plunger in lieu of a number of magnets; but, though such device was used experimentally for about a year, its use was discontinued. To remedy the defects of his system and apparatus something more was required than mere mechanical *133engineering skill; nor did the later inventions inure to his benefit. Computing Scale Co. v. Standard Computing Scale Co., 195 Fed. 508, 115 C. C. A. 418. The Kennedy device' was not entirely useless, in the sense that it would not work at all; but its principle of operation was clearly divorced from that of defendant’s standard and double relay systems and apparatuses.

As in my opinion the patentee had in mind a method for measuring the time during which current passed to the battery, as distinguished from the amount of current, and an apparatus requiring a predetermined voltage, I have reached the conclusion that the claims in issue must be limited to such a method and device, regardless of their broad phraseology.

Defendant contended that in 1911 one Bliss invented the battery protective device in question which eventuated in the modern Sangamo Exhibit meter. The Bliss application for patent embodied a motor supplied with current through a commutator to indicate the charge through the battery. Complainant points to the testimony of Robinson and other witnesses to prove the inoperativeness of the Bliss device ; but in my view of the controversy it is immaterial whether Bliss or Lanphier & Fits solved the problem, bringing into existence defendant’s ampere hour meters. Defendant’s devices do not embody the voltage relay F1 of the Kennedy patent, or any relay set to operate at a predetermined voltage, nor the particular meter — nothing that responds to back voltage of the battery as an initial factor. Its regulation is for constant voltage in the main circuit during battery charge, with gradual reduction of the charging current. Each system has a main coil IS connected between the main leads, which is the principal element in the regulation, with battery coil 11, in the main circuit transmitting current to the batter}?; the meter, upon becoming effective, initiating the rotatable element at a rate of speed proportional to the strength of the current, and thus getting the ampere hours uninfluenced by back voltage of the battery. The double relay shown in Exhibit A does not, like complainant’s relay F1, remain out of action awaiting the back voltage of the battery before becoming active and then stopping the charge. Its operation is due to voltage in the supply mains, but in no sense does it operate only when the battery reaches a prearranged voltage, as in the Kennedy device. This, it seems to me, makes the two systems radically different; one, as heretofore stated, relying on a modified predetermined voltage relay due to the back voltage of the battery, and the other avoiding such characteristics and arranging the elements in combination to make the meter the factor for acting upon the regulator. When the claims in suit are thus construed defendant’s systems are not infringements of the Kennedy patent in suit.

[5] As to the counterclaim: Defendant claims that complainant’s commercial lighting systems, type L, and type B3 in evidence, were infringements of the prior Bliss patent, No. 773,917, dated November 1, 1904, for improvement in electrical distribution, and acquired by defendant by assignment before this action was brought. Reliance is placed on claim 8, which reads as follows:

*134“S. In a system of electrical distribution, a primary generator, means for driving it, a storage battery, a translating device, means under the control of the storage battery for automatically reducing the charging of the storage battery by the generator when the battery is charged to a predetermined point, and means also under the control of the generator-driving mechanism for restoring the generator to its normal operative condition, when the speed of the generator-driving mechanism reaches a predetermined point, substantially as set forth.”

And it is asserted that complainant’s lighting systems embody in combination the stop charge relay described by Bliss, which was brought into activity by the low voltage of the battery after partial discharge. The object of the patentee was no doubt to arrange his apparatus (of the axle-driven type), so that, after the stop charge relay was connected, it would automatically drop back or reset whenever the train stopped, to permit the dynamo to resume charging the battery when the train started. Such is the arrangement of the mechanism.

The patentee did. not describe or illustrate a current regulator, and the operativeness of the invention is challenged on that ground. Invalidity of the patent because of inutility depends upon whether more than mechanical skill was required to connect a dynamo- regulator to the system in order to make it useful. The expert witness Bentley substantially testified on cross-examination that a constant current regulator, such as he had located in. the Bliss system in his sketch on direct examination, would perhaps not operate successfully; that it would tend to cut out the added resistance W in the shunt field circuit upon its coming in, with the result that the charging current to the battery would be resumed at normal quantity, regardless of the action of the stop charge relay F, which concededly was across the battery when in circuit.

When defendant rested with its testimony, complainant moved to dismiss the counterclaim because of obvious inoperativeness; but the motion is now denied. I doubt, however, whether at the date of the invention a current regulator could have been added to the combination of the patent without the exercise of invention. The evidence on this point is not convincing. But the question of the validity of the patent is unimportant, in view of my finding, as' hereinafter stated, that the Bliss patent is not infringed by complainant.

It is contended that the Bliss patent in suit is anticipated by the Brush patent,’ No. 281,176, of July 10, 1883; but this has not been proven. The claims of the Brush and Bliss patents, it is true, read very much alike; but the former were applicable to a stationary lighting system — one in which the dynamo was not intended to be stopped and then started again, as in Bliss’s car-lighting system. The Brush specification, it is true, describes short circuiting magnet C when the battery charge is completed by raising the armature lever, and later dropping it to open the short circuit around the said magnet, thus leaving it ready to act upon the starting of the charging current, regardless of the battery charge. But Bliss’ device was nevertheless different from that of Brush in important particulars. For example, the latter does not relate to a variable speed--driving mechanism for the dynamo, and it has the potential relay differently arranged — that *135is, located permanently across the battery terminals and continuously kept under battery control; while in the former the prearranged voltage relay not only controls the output of the dynamo, but it is also automatically transferable during the charge from battery control to dynamo regulation. These were essential differences, and accordingly the Brush patent was not anticipatory.

However, it has not been proven by a fair preponderance of the evidence that complainant's structures were infringements of claim 8 in question. In its systems the dynamo was self-excited, as distinguished from battery-excited, by the action of a potential relay or stop charge device; while in Bliss the said relay operated to change over from so-called self-excitation to battery-excitation because of the action of the relay. Hence, in my opinion, the proof relating to infringement is insufficient. The respective relays did not, I think, function in substantially the same way to produce substantially the same result.

A decree in accordance with this opinion may be entered, viz. decreeing the Kennedy claims in issue not infringed by the defendant company, and dismissing the counterclaim alleging infringement by complainant of the Bliss patent. No costs to either party.