Dubilier Condenser Corporation v. Radio Corporation of America, 34 F.2d 450 (D. Del. 1929)

*451 Clifton V. Edwards and Kenneth S. Neal, both of New York City, and James I. Boyce, of Wilmington, Del., for plaintiffs.

Charles Neave, Maxwell Barus (of Fish, Richardson & Neave), and Abel E. Blackmar, Jr., all of New York City, and William G. Mahaffy, of Wilmington, Del., for defendant.

MORRIS, District Judge.

These three suits, each based upon a patent having to do with the radio art and now upon final hearing, were tried together. In each, Radio Corporation of America is defendant. In two, the plaintiffs are Dubilier Condenser Corporation, Percival D. Lowell, and Francis W. Dunmore. In the third, they are the same, save that Lowell is not included. The patents sued upon are No. 1,455,141 to Lowell and Dunmore, No. 1,606,212 to Dunmore and Lowell, and No. 1,635,117 to Dunmore. The corporate plaintiff is the exclusive licensee under each. The claims in issue are 3 and 14 of the first, 2, 3, 4, and 6 of the second, and 9 of the third. The defenses are invalidity; noninfringement; ownership by the United States, in that the alleged inventions were made by the patentees in the employer's time, with its tools and facilities, while the patentees were employees of the Bureau of Standards; and, lastly, dedication by the United States to the public.

Patent No. 1,455,141 discloses and claims means permitting the use of alternating current from the standard residence-lighting power in lieu of direct current from A and B batteries and storage cells in radio receiving sets having radio frequency amplifiers, a detector, and audio frequency amplifiers. The advantages of a radio receiving set adapted to use alternating current from the house-lighting circuits over one whose circuits are supplied with energy from batteries or storage cells are obvious. But the ordinary 110-volt alternating current used for lighting purposes has, usually, a frequency of 60 cycles a second. Unless prevented from doing so, these continual reversals or alternations produce, in the sound output of a radio receiving set, distortion of the modulation and a strong hum that make the reception unsatisfactory. Consequently the problem with which the patentees were confronted was the *452 elimination in sets of the type described of the disadvantageous effects of alternating current. Their solution of that problem is set out in claim 14 thus:

"14. In an apparatus for the reception of radio signals the combination of a source of signal energy, a plurality of vacuum tubes having grid, filament and plate electrodes, circuits interconnecting said electrodes whereby said tubes amplify said signal energy at radio frequencies, a detector for rectifying said energy, vacuum tubes for amplifying said energy at audio frequencies, a source of alternating current, means for supplying said current to the filaments of said tubes, means for rectifying said current and supplying potential to the plate circuits of said tubes, and separate means in each of said circuits for eliminating the hum of the alternating current in said apparatus."

The pith of the claim is found in its concluding words, "and separate means in each of said circuits for eliminating the hum of the alternating current in said apparatus," for the invention, if any, resides, as plaintiffs state, in the combination of separate hum-eliminating means of suitable type in each portion of the type of receiver specified.

The patent discloses separate hum-eliminating means for the radio frequency section, for the detector section, and for the audio frequency section of the receiving set. In the radio frequency section, these means comprise a connection of the grid circuits of the tubes in that portion of the set, through a grid biasing means, to a slider of a resistance or potentiometer bridged across the filament circuit, and, in addition thereto, radio frequency transformers between the radio frequency amplifying tubes and between the last of these tubes and the detector. In the detector part, the means consist of a crystal detector and a condenser placed in a connection between the end of the primary winding of the first audio frequency transformer and the filament circuit. The crystal, though it passes such frequency as is imposed upon it, does not itself create indirect hum, as does a tube. The specification states: "The employment of a crystal detector in place of the electron tube detector reduced the 60-cycle hum very considerably," but adds that, though the crystal is the preferred type of rectifier, "other forms of rectifiers may be readily employed." In the audio frequency section, the hum-eliminating means comprise a connection of the grid circuits of that section, through a grid biasing means, to a slider of a resistance bridged across the filament circuit. These grid connections are made to the filament circuit, not at points of maximum potential variation but at points having the average potential of the whole filament. This connection, says the specification, "serves to keep the grid voltage steady with respect to the filaments," and "forms another of the means for eliminating the alternating current hum from the reproducing telephone receivers." The filaments of all the tubes are on the same heating circuit. The plate circuits are connected to one leg of the filament circuit.

The hum-eliminating means of Radiola 17, manufactured by the defendant and charged to be an infringement, differs from the specific means of the patent in several particulars, one of which is that the defendant employs in its receiver three types of tubes, each requiring voltage different from the other two. The filaments of each type are heated by voltage supplied by three different circuits from three separate secondary windings of the transformer. One circuit supplies the energy for heating the filaments of the three radio frequency amplifying tubes and the filament of the first audio frequency amplifying tube, all of which are of the UX 226 type, and require the same voltage. The grid returns of these four tubes are connected with the midpoint of a potentiometer bridged across the heating circuit for the filaments of these tubes.

Another difference is that the detector in defendant's set is a tube, of the UY 227 type, instead of a crystal. In tubes of that type, the cathode is not the filament, but is a thimble, surrounding, but electrically insulated from, the filament. The filament is but the heater for the separate cathode. The grid return of the detector is connected with the midpoint of the resistance across the heating circuit for the filament of that tube. The cathode of the detector is connected with the midpoint of the potentiometer across the heating circuit for the UX 226 tubes. The last audio frequency amplifier or power tube, of the UX 171 type, like the detector, has a separate heating circuit of its own. The grid return of this tube is connected with the midpoint of a potentiometer in its heating circuit. The plate circuits of the several tubes of defendant's set are connected to the filament circuits at the same points as are the grid circuits of the respective tubes.

Do the hum-eliminating means of the patent constitute invention, when employed with the remaining elements of the claims in *453 issue, and, if so, does defendant's receiver infringe those claims? The solution to these problems can be ascertained, not in vacuo, but only in the light of the theories and principles upon which the hum and its elimination are accounted for, which are the theories lying at the foundation of the radio art, of which the most fundamental is, probably, that sound is due to vibration within certain frequency boundaries. Below 16 vibrations or cycles a second, the frequency is too low to be heard. Above 16,000, it is too high for detection by the normal ear. Because waves of low frequency lack velocity and momentum, the range in distance over which they may be transmitted is limited. The higher, or radio frequency waves, which begin, with some overlapping at the end of the upper range of the audio frequency and run upward through eight, nine, or more octaves, with a possible frequency of several million cycles a second, though usually perhaps from 10,000 to 100,000 cycles a second, have the velocity and momentum of light, and, consequently, a great traveling range. To transmit tones great distances, the radio frequency waves are used as carriers of the sound or audio frequency vibrations. To bring about this result, the radio frequency waves, which are usually electromagnetic vibrations, must first be created. This is done by a generator producing and forcing alternating current of the desired frequency into or upon an elevated conductor or aerial connected to the ground. Direct current forced directly onto the antenna will not produce these waves. Alternating current, which reverses in direction at regular intervals of time, is employed. From zero, it builds up in strength gradually, reaches a maximum, and returns to zero. It then builds up to a maximum in the opposite direction, and returns to nothing again.

To visualize this result, it is but necessary to think of a circuit supplied with current by a revolving battery. In one revolution, each pole of the battery makes a contact with each end of the wire completing the circuit. At each revolution, the current flows in the circuit first in one direction, attains its maximum, returns to zero as the poles break their contacts with the ends of the wire, and then flows in like manner in the opposite direction upon contact with the ends of the wire by the poles in reverse position. The double swing of the current, which may be likened to that of a pendulum, is called a cycle. The number of cycles that occur in a second is the frequency of the alternating current in cycles a second. If, when these waves pass from the transmitting antenna, they are all of the same intensity and amplitude, as they will be, if not modulated, they will produce no sound in the telephone receivers or loud speaker of the receiving set, for their frequency is such that their effect upon the diaphragm-operating magnet in the telephone connected with the receiver is practically that of a uniform, direct current, which keeps the pull of the magnet constant and unvarying, and so permits no vibration of the diaphragm, or, if there is any vibration, it is inaudible, for its frequency is above the audible range. But if, as is the practice, sound variations are impressed upon the radio frequency current that produces the electromagnetic or carrier waves, the tone or sound current adds its values to and subtracts them from the average strength of the radio frequency current, and thus produces variations of an audio frequency in the amplitude of the radio frequency waves. It is the rising and falling fluctuations at audio frequency of the intensity of the transmitted radio frequency waves, when reproduced in the plate current of the last audio amplifier in the receiving set, that cause the diaphragm of the telephone or loud speaker of the receiver to vibrate back and forth, and reproduce the frequency and wave form of the rising and falling amplitude of the modulating influence or sound.

Possibly the simplest illustration of modulating means is an ordinary telephone transmitter, electrically connected to the transmitting aerial ground-wire circuit to which an alternator is delivering power at a uniform rate. As the resistance of the telephone transmitter varies with the vibrations of the diaphragm produced by the sound waves impressed thereon, and as the amount of current flowing in a circuit is equal to the voltage divided by the resistance, the vibrations of the telephone-transmitter diaphragm vary the strength or values of the current impressed upon the transmitting antenna. These values fix and determine, in their turn, the amplitude or intensity of the radiated waves.

The continuous, unmodulated waves are symbolized and illustrated by an uninterrupted succession of joined V-shaped figures of uniform height or amplitude, bisected by a straight line. The modulated wave is represented by the same symbol or illustration, save that the V-shaped figures depicting the radio frequency are of a varying height or amplitude, limited by the irregular boundaries fixed by the modulation, or tone transmitted.

*454 Although the radio frequency waves, which have velocity and momentum, are made to carry sound waves which have neither in any large degree, the carrier wave loses strength as it travels. If the distance from transmitter to receiver is great, the signal received may be weak. The weak, modulated, transmitted waves are "picked up" by the aerial of the receiving station and passed to the grid interposed between the filament and plate of the first radio frequency amplifier. The signal current so received upon the grid is, of course, alternating. Hence the grid changes its polarity, or, if given a negative bias, at least its potential, many thousands of times a second. If the receiving set is being operated, a direct current or stream of electrons is flowing from filament to plate, while these changes in polarity or potential are taking place in the grid. The volume of this flow, the strength of this current, which is the output of the tube, is controlled by at least three basic factors the temperature of the filament, the potential or strength of pull of the plate, and the variations in the input voltage upon the grid. No electrons are emitted by the filament, and hence no plate current flows, unless the filament is heated. This is the "Edison effect." When heated, the emission increases and decreases with the temperature. The electron is, or carries, a charge of negative electricity. To cause the emitted electrons to flow across the tube, a positive charge must be impressed upon the plate. Other things being equal, the proportion of the emitted electrons drawn to the plate varies with the plate potential.

But, as like charges of electricity repel and unlike charges attract each other, a negative charge upon the grid, interposed between filament and plate, decreases the flow from filament to plate, while a positive charge thereon, but not high enough to stop the electrons at the grid, increases the flow. The effect of a unit change in grid voltage upon the plate current in a particular tube is, usually, about eight times that of a like change in the plate voltage. The number representing this relative effect is known as the amplification coefficient of the tube. It varies in different tubes, but remains constant for the particular tube. The relation between grid voltage and plate current, however, is not uniform and proportional. A unit increase or decrease in grid voltage produces at some voltages greater effects upon the plate current than at other voltages. This relation may vary in different tubes. It is shown by the characteristic curve of the tube. This curve is obtained by plotting plate current against grid voltage. It is not straight. It resembles the integral sign or symbol; that is, at the lower voltages, its tendency away from the horizontal is not great. As the grid voltage increases, becomes less negative, it turns more or less sharply toward the perpendicular. After the voltage becomes positive, the curve turns again more or less sharply toward the horizontal.

Every factor that affects the plate current is of moment, for any pulsation impressed upon it, in any tube, is amplified by each succeeding tube, and the pulsations of the plate current in the last audio frequency amplifier operate the diaphragm of the telephones or loud speaker, by which the plate current pulsations are converted into sound. Consequently it is obvious that, if the sound output at the receiving station is to be an exact reproduction of the sound input at the transmitting station, the plate current must remain uniform, steady, and undisturbed, save as it is affected by the changing grid potentials produced by the incoming signals.

But, in order for a three-electrode tube to function, the input or grid circuit and the output or plate circuit must be connected to the filament circuit. It follows that any variation of the normal potential of any one of the three electrodes with respect to the other two will, unless overcome or offset, affect the plate current, and consequently the sound output of the telephones or loud speaker. If such variation is continuously recurring and its effect upon the plate current is substantial, it is patent that the sound output will be unsatisfactory, or even intolerable. In battery operated sets, in which the filament is heated by the direct current of an A battery and the plate is given its positive potential by a B battery, it is clear that the potential relation of the electrodes to one another is practically constant, save as affected by the incoming signal and such changes as the operator may make in the potential of filament or plate. This is true even if, as is usually done in such receivers, the grid and the plate circuit connections to the cathode are made to one leg or side of the filament, and, as a consequence, all parts of the filament are not at the same potential with respect to the plate or grid, for this difference of potential is constant, and hence produces no variations in the plate current.

The employment, to heat the filament, of 60-cycle, raw alternating current, with its constantly recurring reversal of direction through the filament, and the consequent changing polarity of the legs of the filament, *455 brings about variations of the normal potential of the grid and of the plate with respect to that of the filament as a whole, if the connections of grid and plate circuits to the filament circuit are the same as in battery operated receivers. Moreover, when alternating current is used to heat the cathode, the plate and the grid circuit connections to the filament remaining as in battery operated sets, three additional disturbing effects upon the plate current arise because of variations in the electronic emission of the cathode. One of these is due to changes of the temperature of the cathode occurring because of the zero to maximum current variations in the current flow through the cathode, brought about by the reversals in the direction of flow. The remaining two effects are caused by fluctuations of the magnetic field and of the electrostatic field, respectively, surrounding the filaments.

The consequence of these four disturbing effects of the use of alternating current to heat the filaments varies in the different sections of a receiver of the type called for by the claims. In the radio frequency amplification section, the result of the 60-cycle changes, which are in the audio frequency band of waves, is the tendency of these vibrations to distort the modulation upon, or to further modulate, the weak, incoming, carrier-wave, for the radio frequency transformers which connect the radio frequency tubes bar the passage of any audio frequency pulsations that are produced in the plate current in the radio frequency tubes by the 60-cycle current. The consequence in the telephones or loud speaker is a gurgling or gargling sound. In the detector section of the receiver, where the current is rectified or cut in half by the separation of that flowing in one direction from that flowing in the other, and by the passage of one of these parts only, the effect resulting from the use of a tube and from the direct heating of its cathode by alternating current is an exaggeration of the tendency found in the radio frequency part. From the detector, the current goes through an audio frequency transformer to the first audio frequency amplifying tube. The audio frequency transformer, or transformers, if there are two or more audio amplification stages, are designed to pass waves or plate current pulsations of audio frequency, and to deny further passage to those of carrier wave or radio frequency. In the audio frequency amplification portion of the receiver, the effects of the alternating current there introduced are not a further modification of the transmitted sound waves, for at this stage they have acquired through amplification a self-preserving intensity, but is a strong hum, produced by the 60-cycle frequency pulsations impressed upon the plate current, which vies with or drowns out the tone effects of the transmitted sounds.

Professor Moye tersely summarized the matter in 1921, by the statement that, if one heats the filaments of tubes by alternating current, "without special precaution, one ascertains that the receiving tube lights up well, but the listening devices are the seat of intense hummings, which have a rhythm based on the cycles of the alternating current, and which cover up the signals absolutely. These parasitic hummings seem to be difficult to abolish, for they arise from the very fact of the alternations of the heating current, whose polarities, constantly changed, affect the grid of the tube and impose upon the plate current a periodic rhythm which the telephones transmit as sounds in a manner that is as faithful as it is deplorable."

Lowell and Dunmore were not the first to enter or make accomplishments in this particular branch of the radio art, nor was their contribution a complete solution of the whole difficulty. There are few true pioneers, and it is but seldom that an inventor supplies at one stroke a full answer to the whole problem, leaving nothing more to be done. On the contrary, an invention is usually but a short step, sometimes an almost imperceptible step, along a lengthy pathway. So it is here. The patentees were not the first to substitute for the B battery direct voltage derived from rectified and filtered alternating current. They were not the first to substitute alternating current for the A battery. It was old to bring together the substitutions of alternating current for the A battery and direct current derived from rectified and filtered alternating current for the B battery. The contribution of the patentees did not eliminate all difficulties or disturbing effects resulting from the use of raw alternating current. Their advance, if any, is found, as I understand it, not in a remedy or means to prevent variations in the electronic emission of the filaments or cathodes arising from their constantly recurring changes of temperature, and from fluctuations in the magnetic and electrostatic fields surrounding the filaments, but in their organization of means for the elimination, partly or wholly, in receivers of the three section or division type, of the pulsations in the telephone or loud speaker operating plate current that were due to the variations of the normal grid and plate potentials with respect *456 to the filaments by the elimination, partly or wholly, of those variations.

Even in this restricted field they were not pioneers, for before their advent Heising had disclosed in his patent No. 1,432,022 the specific potential variation elimination means that the patentees made use of in the several parts of their receiver. More than that, Moye had pointed out and Dr. "P. C." had explained that radio frequency transformers or other tuned circuit transformers prevent the passage of direct hum from the radio frequency tubes to the detector and telephones. Still again, Dr. "P. C." had shown that the modulation effect in the radio frequency tubes can be greatly reduced by impressing upon the grid a bias or voltage that will cause the tube to operate on the straight part of its characteristic curve; that is, in the range of voltages indicated by that part of the curve, for in that range the relation of grid voltage to plate current is proportional. He taught that, if by a C battery, or other grid biasing means, the grid is given the voltage indicated by the midpoint of the straight part of the curve, the fluctuating increases and decreases made in that voltage by the incoming signal energy, if not too great, will not extend to or around the knees of the curve where modulation is produced by a more or less abrupt change in the grid-voltage, plate-current ratio. In fact, plaintiffs concede that every element of their claims was, in itself, old. But they correctly assert that no one before them ever applied raw alternating current to the filaments in the audio frequency portion of a receiving set, and that there was no prior disclosure or suggestion that separate hum-eliminating means installed in the several parts of the receiver would be advantageous.

Defendant makes its attack by pointing out many asserted faults or deficiencies in the disclosed "hook up" or circuits in which the separate hum-eliminating means are installed, and by taking the position that there is no advantage in using a separate grid-potentiometer connection for each part of the receiver, if all the tubes have the same voltage requirements. But it is not denied that the main circuits are operable. Consequently it is immaterial that the patentees show a half-wave rectifier instead of a full-wave rectifier for the plate current, a rheostat at a particular place in the filament circuit, instead of elsewhere, a grid leak and condenser, instead of a C battery, or other improved biasing means in the grid circuit of the radio frequency tubes, or fail to show tuned coupling transformers.

Nor can I agree with defendant's conclusion that there is an infirmity in the patentees' disclosure of a grid circuit connection to the potentiometer and a plate circuit connection to one side of the filament circuit, called by defendant's expert the "method of cure," instead of a connection of both grid and plate circuits to the potentiometer, called by defendant's expert the "method of prevention," for both methods and their respective advantages and manner of use were set out in patent No. 1,432,022 to Heising, and were, consequently, known to those skilled in the art to whom the specifications of patents, including the patent in suit, are addressed. The art having been taught by Heising, among other things, that if the plate connection is made to one side of the filament circuit, and the grid connection to the slider of a potentiometer across that circuit, compensation for the variation of the plate potential with respect to the filament must be made by moving the slider to the other side a distance governed by the amplifying coefficient of the tube, Lowell and Dunmore were not called upon to repeat that instruction.

Whether a separate grid potentiometer or the alternative separate grid-plate potentiometer connection for each part of the receiver has any merit or utility, if all the tubes have the same voltage requirements, is a matter with respect to which the parties differ sharply. The defendant, contending that the patentees' arrangement of separate potentiometers is of no practical utility and does not involve invention, asserts that in separate, fixed, potentiometer connections for tubes requiring the same voltage, the points of connection all have the same potential, and so have the same effect as if but a single potentiometer were employed, and that separate variable connections are productive of beneficial results only if there is a variable connection for each individual tube, whereby adjustments may be made for the eccentricities of any tube varying from standard.

It seems to me this contention has two destructive infirmities. In the first place, the potentiometer connections of the patent are variable, not fixed. If the tubes employed require it, the connection on one potentiometer may be made at a point having one potential, and that on the other at a point having a different potential. As I understand it, tubes made at the date of the patent were less uniform in their characteristics and functioning than are the tubes of to-day. Again, quite satisfactory adjustments for tubes of slightly varying characteristics may readily be made, if the tubes are divided into groups having *457 separate adjusting means, for tubes having certain characteristics may be put in one group, and those having other characteristics put in another. This is particularly true where, in receivers of the type called for by the claims, variations in tubes and accuracy of the connection adjustment is of less importance in one section than in another.

In any event, the preponderance of the evidence supports the claims. The patent carries a presumption of validity. The Bureau of Standards deemed the disclosures of the patent of such importance that it caused a receiver embodying the invention to be made, and published its Bulletin No. 450 describing it. The technical magazines discussed it. Some licenses have been granted. Royalties have been paid, though the specific design of the structures made under the licenses is not shown by the record, and it does not appear that they have had any marked commercial success. Persons skilled in the radio art, who heard one or more of the sets made by the patentees operate, concluded and testified that something had been accomplished. Dr. Lee testified that the patentees' arrangement of the hum-eliminating means "prevents the hum from becoming cumulative through the sets. It also has a great benefit, in that you can operate the various sections of each set adaptable to their maximum efficiency by selecting tubes suitable for each purpose and rectifying the hum in those sections to the best advantage." Other witnesses testified to the same general effect.

It is quite true that the patentees did not eliminate all causes of hum arising from the use of alternating current. They did not eliminate the hum-producing fluctuations in the plate current arising from the variations in the temperature of the cathode and from the fluctuations of the magnetic and the electrostatic fields surrounding the filaments. They did not invent or disclose a tube detector possessing the advantages of both a crystal and an audion when used with alternating current. And it is quite possible that these further steps needed to be taken before the alternating current operated set could be made commercially successful. The patentees are entitled to no credit for these later improvements. But they were the first to make and disclose a receiver, of the type claimed, operable by alternating current.

Obviously, later supplemental discoveries and improvements by others have no tendency to take from the patentees any credit or benefits to which they are entitled for the particular advance in the art made by them. Fleming's contribution was recognized even after De Forest's improvement converted the Fleming valve into the epoch-making audion. Bell's telephone was made generally useful by reason of improvements by others. Although it is not claimed that the contribution of Lowell and Dunmore is of the same rank as the contributions made by Bell and Fleming in their respective arts, yet the invention sustaining principles of law apply with equal force to all. I think the claims are valid.

Upon the issue of infringement the question is, not whether the defendant has used supplemental or additional hum-eliminating means to prevent or overcome the effect of variations in filament temperature and fluctuations in the magnetic and the electrostatic fields, or even whether it has used supplemental or additional means to prevent or offset the consequences of variations of the normal grid and plate potentials with respect to the filament. On the contrary, it is solely whether the defendant has employed the separate hum-eliminating means of the patent, or their equivalents, to produce, in whole or in part, its result. It is not disputed that the defendant employs as its radio frequency and as its first audio frequency amplifiers tubes of the UX 226 type, designed after the granting of the patent in suit, for the express purpose of overcoming the thermal, magnetic, and electrostatic fluctuations arising from the use of alternating current upon filament cathodes. Nor is it denied that the UX 226 type of tube accomplishes to a very substantial extent the purposes for which it was designed. But the question is, not whether the UX 226 tubes are the means of the claims or are the equivalents therefor, but is whether other hum-eliminating means of Radiola 17, which supplement or are supplemented by the partial hum-eliminating or hum-preventing features of the UX 226 tubes, are such means.

Defendant, however, presents a preliminary question based upon the fact that in the receiving set for which the patentees disclosed their hum-eliminating means, the patentees showed only a crystal detector and a condenser in a connection between the end of the primary winding of the first audio frequency transformer and the filament circuit, and stated that "the employment of a crystal detector in place of the electron tube detector reduced the 60-cycle hum very considerably," while defendant employs a tube detector, no condenser, and connection of the character disclosed in patentees' detector section. It contends that the claims cannot be construed to cover a set employing an electron tube detector. The advantage of the *458 crystal resides in the fact that it has no filament, no filament input of alternating current, and, consequently, does not itself generate hum. The condenser was installed merely as a precautionary means to introduce opposing voltages into the detector circuit to compensate for stray pick-ups from the alternating current circuits.

Defendant, however, elected to employ as its detector a tube which generates hum, and to eliminate the hum so generated by the means used by the patentees in the sections in which tubes were employed by them. It is true that defendant's detector tube is of special design, and that the hum generated by it is less than that generated by other tubes. It is of the UY 227 type. In it the filament is not the cathode, but is merely the heating element for the separate cathode. The cathode has large heat inertia, and consequently is not subject to the wide temperature variations that occur in tubes in which the filament and cathode are one. It is electrically insulated from the filament, so that the alternating heating current of the latter does not pass to it. Like other three-electrode tubes, the UY 227 type amplifies. The patentees did not, however, restrict themselves to a crystal detector. They expressly stated that "other forms of rectifiers may be readily employed." The only "other form" of rectifier then in practical use was the electron tube. Nothing in the specification excludes it from the scope of the claims. Nor did the patentees restrict themselves to the precise form or design of tubes then in use, if their tube hum-eliminating means were employed in supplement thereof. The defendant did not escape the patent by substituting an improved tube detector in place of the crystal and condenser, if in addition thereto it made use of the appropriate hum-eliminating means of the patent.

Defendant's ultimate contention is that it does not employ the patentees' separate means for eliminating hum, or the equivalent of such means. The patent shows two potentiometers. To one, the grids of all the radio frequency tubes are returned; to the other, the grids of all the audio frequency tubes. All the tubes of the patent have the same voltage requirements. Radiola 17 has one potentiometer for the output tube of the UX 171 type, having one voltage requirement, another potentiometer for the detector tube of the UY 227 type, which has another voltage requirement, and still another potentiometer for its UX 226 tubes, three in the radio frequency section and one in the audio frequency section of the set, having still another voltage requirement. In support of its contention that its hum-eliminating means are not the separate means of the patent, defendant makes two contentions. Of these the first is that defendant's arrangement is in accord with the teachings of the prior art, which discloses a plurality of amplifying tubes having the same voltage requirements connected to one potentiometer, and tubes of different voltage requirements connected to different potentiometers. The second is that, by connecting the grids of its radio frequency tubes and the grid of its first audio frequency tube to the same potentiometer, it not only ran directly counter to the teachings of the patent, but by so doing it furnished to each audio tube of its set that is, to the detector, to the first audio amplifier, and to the output or power tube a separate or individual adjustable potentiometer, whereby each tube receives the most delicate adjustment.

The patents and publications of the prior art do disclose in several instances a plurality of tubes, apparently of the same voltage requirements, connected to the same potentiometer. In fact, it was Lowell and Dunmore's departure from a single hum-eliminating means common to all tubes and sections of a receiver that marked their advance over the prior art. Patent to White, assignor to General Electric Company, No. 1,195,632, granted in 1916, discloses tubes having different voltages connected to different potentiometers. That patent was cited by the Patent Office against the Lowell and Dunmore application, but as it disclosed a transmitting, not a receiving, apparatus, and as it, of course, showed no separate hum-eliminating means for each section of a receiving set, it did not constitute an anticipation of the patent in suit, and, when this was pointed out, was withdrawn. Yet it is quite true that the patent does disclose separate grid-potentiometer connections for tubes of different voltages. But this fact is not in itself determinative of the question of infringement, for a separate grid-potentiometer connection for tubes of different voltages is not, in itself, a denial of a separate grid-potentiometer connection for each section of the receiver. Hence the final question is whether defendant's separate grid-potentiometer connections for tubes of different voltages are, or are not, likewise, separate grid-potentiometer connections for each of the three sections of its receiver.

In the great majority of cases that come before the courts, the defendant has made some variation in one or more of the elements of the claimed invention. But a *459 variation is, of course, no defense, if the substance or principle of the patent is employed. Nor will it avail a defendant to show merely that he has added something to the patented device, enabled it to perform added functions, or made it better or worse. A departure from the letter of the claims will not serve, if their principle and substance are made use of. In the decision of such cases the court acts merely as a jury would, and finds in each case whether on the particular facts the invention claimed has or has not been taken.

Had defendant employed for its radio amplifiers tubes requiring one voltage, for its detector a tube of different voltage requirements, and for its audio amplifiers tubes calling for a still different voltage, and connected the grids, or the grids and the plates, of the tubes to potentiometers across their respective filament circuits, I take it that, notwithstanding the employment of different voltages, infringement could not be successfully denied, if the conclusions hereinbefore arrived at are sound. Structurally, however, defendant departed from that arrangement by placing its radio amplifying tubes and its first audio frequency tube, of the same voltage requirement as the radio frequency amplifiers, in the same filament circuit, and by connecting the grid and the plate of the latter, as well as the grids and the plates of the radio tubes, to the potentiometer across that circuit. Notwithstanding this shifting of the grid and plate connections of the first audio tube from the audio section to the radio section of the set, it employed three potentiometers, the number of which is identical with the number of sections of the receiver. Lowell and Dunmore were the first to disclose the advantages of a multiplicity of potentiometers, one for each section, in a receiver of that type. Defendant presented much testimony and made several demonstrations, which make clear, I think, why the change from the strict letter of the claims in suit could be made in the connections of the first audio tube probably without detriment, possibly to some advantage; but neither the testimony nor the demonstrations disclose in Radiola 17 any departure in principle or spirit from those claims.

Patent No. 1,606,212, granted November 9, 1926, to Dunmore and Lowell upon an application filed March 21, 1922, relates to an amplifying unit and reproducer of electrical signals. The unit includes a plurality of thermionic vacuum tubes connected in circuit for audio frequency amplification and a loud speaker employing a coil armature and an electromagnetic field coil. The unit was old; but it had been operated only with direct current. The patentees' aim was to obtain faithful, humless reproduction from alternating current operation. Their means for accomplishing this result are set out in claim 2 thus:

"2. In an apparatus for the reproduction of sound, a thermionic vacuum tube amplifier, an input circuit, an output circuit, a power source of alternating current, means for rectifying said current, a loud speaker reproducer connected in said output circuit and comprising an armature winding and a field winding and circuits for supplying said rectified current to said output circuit and to said field winding, said windings being connected in such relation that disturbing current effects in said loud speaker reproducer are substantially eliminated."

The function of the loud speaker is, of course, to convert the audio frequency pulsations of the output electrical current of the amplifier into sound vibrations. To produce this result, means must be employed to enable the pulsations to actuate a diaphragm. A telephone receiver, which functions in this manner and for this purpose, consists of a hard rubber case inclosing a U-shaped permanent magnet of hard steel, to the free ends of which are clamped two soft iron pole pieces, each wound with a coil of small-diameter insulated copper wire in the talking or output current circuit. A hard rubber cap or earpiece, with an opening at its center, supports a soft, sheet-iron diaphragm in front of the pole pieces. As the pulsating current flows through the coils on the soft iron poles, it sets up a magnetic field, which either strengthens the constant magnetic field of the permanent magnet, and so draws the diaphragm a litle closer to the poles, or weakens that field by opposing it, and hence allows the diaphragm, which is always under strain of attraction from the permanent magnet, to spring away from the poles. Between pulsations the magnetic field returns to its normal strength, and the diaphragm immediately returns to its original and normal position. Thus the diaphragm vibrates with the pulsations of the current, and reproduces the sound vibrations which at the transmitting station caused the current pulsations. This device employs both the permanent magnet of hardened steel and the electromagnet of soft iron pole pieces, which is magnetized only when electricity flows through the coils and loses its magnetism when the flow ceases.

In the loud speaker of the patent there is no permanent magnet. An electromagnet, only, is made use of. This is in the field coil, *460 10. Again, the armature is not the loud speaker diaphragm, but is an "armature coil," 9, in the output circuit of the vacuum tube amplifier, movably supported and connected to the diaphragm. It is so positioned that it is, likewise, within the magnetic field of the electromagnet. As current electricity always produces a magnetic field, and as the pulsating current of the amplifier output circuit passes through the armature coil, an electromagnetic field is also set up by that coil. The fluctuations of that field correspond with the pulsations of the current by which that field is produced. These fluctuations, acting in conjunction with and in opposition to the field or pull of the electromagnet, impart vibrations to the armature coil, and so to the connected diaphragm, proportionate to the current pulsations in the amplifier output circuit. The audio frequency vibrations of the current are thus converted into sound.

Since the sound output of the loud speaker is the joint product of the two magnetic fields, it is obvious that for accurate reproduction of the transmitted tones the only audio frequency fluctuations in either of those fields that may be permitted to take part in imparting vibrations to the armature coil, and through it to the diaphragm, are those produced by the transmitted-tone variations in the current flowing through the armature coil. So to restrict the magnetic fluctuations in the joint field, when both the amplifier and the loud speaker are powered by 60-cycle alternating current from residence lighting circuits, was, in reality, the problem to the solution of which the patentees addressed themselves. They were the first, so far as the record discloses, to set themselves to this task. Until about 1918, the ordinary telephone receiver was employed. Then the permanent magnet type of receiver, having a weak, but constant, field, appeared. About 1920, the Magnavox, in which the output circuit from the receiving set passed through an armature coil suspended in the field of an electromagnet, the coil of which was energized by a battery, made its appearance. Its power requirements were such that it fell into disuse. Instead of attempting to supply the necessary power, the art directed its efforts to the improvement of the permanent magnet type of loud speaker.

Lowell and Dunmore, however, went in the opposite direction. They turned to the house-lighting, alternating current for power, and employed means to overcome its 60-cycle, hum-frequency effects in the joint field. In the amplifier the means employed consist of a rectifier, a filter, or condenser, and inductance coil, to minimize the pulsations in the rectified current, the old, variable, potentiometer connections in the grid-filament circuit, and a separate, armature-coil circuit. In the field circuit, wherein, also, current variations or pulsations would produce corresponding fluctuations in the magnetic field, the patentees utilized a half-wave rectifier, and the field coil itself to remove, neutralize, or reduce, through its inductance and "skin effect," the pulsations and hum-frequency variations in the current. In addition the patentees pointed out by amendment of March 27, 1926, that pulsations in the field winding and in the armature circuit are opposite in phase, and that there is a consequent tendency for the disturbing pulsations to offset and neutralize each other.

Defendant contends that the amendment was and is invalid, in that it contains new matter and was made about six months after defendant's loud speaker went on the market. It further contends that the claims are based on the amendment, and that they, too, must fall, not only for want of a proper foundation in the specification, but also because of anticipation by Pridham and Jensen's patent No. 1,105,924, issued in 1914, and because, if they are not limited by phase opposition, they are merely for an amplifier, old per se, aggregated with the old Magnavox loud speaker. Defendant likewise urges that its loud speaker is not within the scope of the claims when properly construed.

Plaintiffs take the position that the amendment did not introduce new matter in any invalidating sense, in that, as they assert, the amendment was in effect but an explanation of the theory upon which the device operated or functioned in part. They further contend that the claims in issue, even if construed to include phase opposition, are not anticipated by the Pridham and Jensen patent; that, since other claims not in issue are specific to phase opposition, the claims in issue should not be so construed to be limited by phase opposition, but that, howsoever construed, defendant's loud speaker comes within them.

The most direct approach to a solution of such of these issues as are requisite for a proper conclusion lies, perhaps, through an ascertainment of the meaning and scope of the claims. For this purpose claim 2, above quoted, was accepted by the parties as representative. Their differences have to do only with the last clause "said [armature and field] windings being connected in such relation that disturbing current effects in said *461 loud speaker reproducer are substantially eliminated."

Regardless of the fact that Dunmore and Lowell were the first to obtain a reasonably satisfactory reproduction of transmitted sounds from an amplifier loud speaker unit energized by alternating current, it is, of course, beyond dispute that the claims cannot be given a construction broad enough to confer upon the patentees a right to the exclusive use of alternating current for producing their result. O'Reilly v. Morse, 15 How. 62, 14 L. Ed. 601. Consequently, interpretation requires that the means of their disclosure be resorted to. Separately considered, both the amplifier and the loud speaker were old. "Skin effect," and impedance, which includes reactance and resistance, were known inevitable consequences of the energization of field coils of the Magnavox type by pulsating current. In view of the lack of novelty in the amplifier and in the loud speaker, separately considered, and of its failure to find any co-operative effect set out in the specification, the Patent Office rejected all claims. Thereafter the patentees amended by adding the matter now appearing in lines 65 to 79 of page 2 and in lines 24 to 33 of page 3 of the patent. All claims were likewise amended. As a consequence of these facts and actions, all claims must be held to include and be limited by the interaction or co-operation disclosed in the amendments. If these claims were anticipated by the Pridham and Jensen patent, or if the amendments, the keystone of the supporting arch for the claims, contained new matter in an invalidating sense, the claims must, of course, fall; but, as I can find no infringement by defendant's loud speaker of the claims when limited by phase opposition, it is unnecessary to pass upon the validity of the claims.

Plaintiffs expressly concede that there is no co-operation between defendant's amplifier and loud speaker to neutralize or eliminate directly by phase opposition the hum-producing fluctuations. They assert, however, that defendant, through equivalents, employs the means of the claims to produce the result of the patent. But defendant's device is designed to eliminate from the amplifier circuit and from the field coil circuit any and all hum-producing fluctuations there present, and to leave no residual fluctuations to be neutralized by co-operation between the amplifier and loud speaker. The means used by defendant to make its field coil current steady and pulsationless, and to shield its armature coil from the effect of residual fluctuations or magnet leakage, include, in addition to the impedance and skin effects of the coil itself, a full-wave rectifier instead of the half-wave rectifier of the patent, a filter to smooth out the ripples in the rectified current, a pot magnet, the placing of the armature coil in as noninductive relation as possible to the field, and copper rings to shield the armature coil from the effects of any surviving current pulsations.

Defendant's mode of operation is the direct antithesis of that of the device of the patent, which intentionally employs neutralization by direct phase opposition. Plaintiffs, however, insist that ripples of hum frequency in the current of the field winding cause magnetic fluctuations that produce eddy currents in the copper rings; that the same results flow from alternating current variations in the armature coil, and that these currents are in opposed phase relation to the producing currents and consequently bring about opposed phase relationship. But even if this theory be sound, and its soundness is questioned, the phase opposition of the patent is direct, not echoed.

Patent No. 1,635,117, to Dunmore, is for a signal receiving system. The complete device therein disclosed is one for radio operation of a telegraph instrument. Among the elements of the system, however, are three-electrode, amplifier tubes, operated from an alternating current source of power. The pith of claim 9, the only claim here in issue, is found in the novel feature of supplying a negative potential to the grid of an amplifier tube from a source of alternating current.

When an electron tube is used as an amplifier in any radio signal receiving system or set, it is essential, if it is to function properly, that it operate on the straight part of its characteristic curve, which means, as I understand it, that the normal potential of the grid must be negative with respect to the filament. The means employed by Dunmore to obtain such grid potential, without interference from the source of alternating current, consist of a transformer connected through a primary winding with the current-source, a secondary winding, a rectifier, a condenser, to smooth out the rectified current, connected across the output terminals of the rectifier, and a resistance of desired value shunted across the output circuit of the rectifier tube. It is this resistance that causes the drop in potential and creates the desired negative potential on the grid. In tubes operated by direct current, a negative grid potential had been obtained by a resistance, *462 or by a C battery in the grid circuit having its negative pole towards the grid.

There is no instance, however, in the prior art, of the employment of a resistance to obtain a negative grid potential in an amplifier tube operated from an alternating current source. Yet the use of a resistance for that purpose in connection with a tube operated from an alternating current source would seem obvious, but for the evidence to the contrary afforded by the prior art. White, of the General Electric Company in 1916 was granted his patent No. 1,195,632, which discloses alternating current operated amplifiers, but it disclosed no means for maintaining a negative grid potential. In the same year, Heising, of the Western Electric Company, likewise disclosed in patent No. 1,432,022 amplifiers operated by alternating current, and likewise failed to disclose means for obtaining a negative potential on the grid. In 1917, in patent No. 1,445,278, Heising made use of a resistance in a receiving set operated by direct current. Mathes, of the Western Electric Company, in patent No. 1,426,754, applied for in 1916, and Scriven, of the same company, in patent No. 1,375,739, filed in 1918, both employed resistance to obtain a negative grid potential in direct current operated sets. Moye, who in 1921 showed a set in which A and B batteries were replaced by alternating current, failed to show any means for bringing about a negative grid bias. In July, 1921, Dr. "P. C.," commenting upon Moye's disclosures, showed the necessity of a negative potential on the grid, the difficulties of obtaining it when an alternating current source is employed, and for his solution suggested the old C battery of the direct current sets.

Hartley patent No. 1,442,430, pleaded as an anticipation, shows the use of a special, direct current generator with three brushes to obtain the desired grid bias. Defendant, however, takes the position that, since an alternating electromotive force is induced in armatures of all generators, and that direct current in the external circuit is but commutator rectified alternating current, the Hartley source of power must be considered as alternating current, not direct current. But generators have long been so perfected and arranged that they produce smooth, continuous, direct-current, and in his rebuttal testimony the witness Brigham makes it clear that the problems arising with respect to receiving sets operated from an alternating current source are radically different from the problems attending operation by current obtained from a direct current generator.

Another defense is that plaintiff's disclosure is for a telegraph relay system, not for a loud speaker or receiving set. This contention looks only at the surface differences, and not at the fundamental analogies. In each instance, the only thing the art needed to be told was how to obtain a negative grid potential from a source of alternating current. That the patent disclosed. The art already knew what the potential should be in each instance, and how to vary it from any disclosed values.

Again, defendant asserts that claim 9 is void because Dunmore, in April, 1922, two months after his patent application was filed, published a full description of the subject-matter of his patent, and that claim 9 was not inserted in the application until October, 1925, 3½ years later and two or three weeks after defendant's loud speaker went on the market. In support of this defense, defendant relies upon Webster Electric Co. v. Splitdorf Electrical Co., 264 U.S. 463, 465, 44 S. Ct. 342, 68 L. Ed. 792, and Westinghouse E. & Mfg. Co. v. Jeffrey-De Witt Insulator Co., 22 F.(2d) 277 (C. C. A. 2). But, unlike those cases, in the suit at bar there was no delay of more than two years after others had entered the field, and there is no presumption that Dunmore had knowledge of defendant's loud speaker, or that the amendment was motivated by it.

Defendant denies infringement by its loud speaker 104 and its Radiolas 17 and 18, and asserts that all that it has done is to supply old, direct-current circuits with direct voltage from rectified and filtered alternating current as a substitute for batteries. But it was Dunmore's inventive genius that taught the art that defendant's circuits, old as they are, constituted means for supplying a negative potential to the grid electrode of a tube amplifier from a source of alternating current. I think claim 9 is valid and infringed.

Of the two special defenses set up in the answer in each suit, the first is in the nature of a plea in bar; the second, in the nature of a plea in abatement. The facts alleged in support of these "pleas" are that Lowell and Dunmore were research employees of the Bureau of Standards; that they undertook the work resulting in the patents in suit at the direction of their employer; that such work was done in the course of such employment, in the employer's time, and with its tools and facilities; that the Bureau of Standards, with the co-operation of Lowell and Dunmore, published the results of their work, and issued announcements that the work had been accomplished by the Bureau *463 of Standards and was available to the public. It is further alleged that the administrative policy of the Bureau is, and since its creation has been, that the output of the Bureau belongs to the people; that the only justification for a patent by a member of the staff is to prevent private individuals from patenting, and thus restricting the free use by the public of the results of the Bureau's work, and that "all such patents, therefore, should naturally be dedicated for free unrestricted use by the public."

From these alleged facts, defendant draws the conclusion, and alleges in its "plea in bar," that the patents were inadvertently and by mistake and without authority of law granted to the respective inventors; that the inventions and patents are in equity the property of the United States; that, by reason of the dedication of the inventions to the public, the defendant has the right to practice them; and that, if the United States has not dedicated the inventions to the public, the United States has no authority by law to grant to itself a patent monopoly, and that the public is entitled in equity to practice the invention of any Letters Patent held by the United States. In the "plea in abatement," defendant relies upon the foregoing alleged facts to establish that the United States is an indispensable party to these suits.

The patents were not granted to the patentees by mistake. On the contrary, they were issued in strict conformity with the patent statutes. It is not questioned that Lowell and Dunmore were the inventors of the subject-matter of the first and second patents, and Dunmore the inventor of the subject-matter of the third. Rev. St. § 4886 (35 USCA § 31), conferred upon them rights to obtain patents for their respective inventions. It is true, of course, that patents may be granted and issued to the assignee of an inventor, "but the assignment must first be entered of record in the Patent Office." Rev. St. § 4895 (35 USCA § 44). Whether the United States had or had not an equitable right to require, if it so desired, an assignment of the inventions or patents to it, there has been no assignment of either.

In fact, defendant's contention with respect to the ownership of the patents by the United States is predicated neither upon an assignment, nor even upon any express terms of the contracts of hiring. It rests solely upon the nature of the service in which Lowell and Dunmore were engaged at the time the inventions were made and the implied duty which an employee owes to his employer with respect to the service in which he is engaged. Neither the statute creating the Bureau of Standards, 31 Stat. 1449 (15 USCA § 271 et seq.), nor any amendment thereof, has altered the patent statutes. Neither the Act of March 3, 1883, 22 Stat. 625 (35 USCA § 45), nor the Act of April 30, 1928, 45 Stat. 467 (35 USCA § 45), denies to an employee of the Bureau of Standards a right to take advantage of the patent statutes. See Squier v. American Telephone & Telegraph Co., 7 F.(2d) 831 (C. C. A. 2). Whether this should be so is a matter for the Congress, not the courts, to decide.

Papers describing the several inventions were published by the inventors and by the Bureau of Standards. But I find nothing in such publications constituting either expressly or by reasonable implication a dedication of the inventions, or any of them, to the public. From these findings, and without regard to what any further findings, if made, might be, three conclusions follow: (a) That upon the grant of the patents the respective patentees became vested with the legal title thereto; (b) that if, by reason of a relation of employer and employee existing between the United States and the patentees at the time the inventions were made, and by reason of the nature of the service to which the patentees were assigned and the particular circumstances under which the inventions were made, the United States has a right to require an assignment of the patents, such right, personal to the employer, is exercisable at its option alone; and (c) that, whatever the optional rights or equities between the government as employer and the patentees as employees were, such rights or equities cannot constitute a defense to infringement by a third party. Yablick v. Protecto Safety Appliance Corporation, 21 F.(2d) 885, 889 (C. C. A. 3).

Gayler v. Wilder, 10 How. 477, 493, 13 L. Ed. 504, Hendrie v. Sayles, 98 U.S. 546, 25 L. Ed. 176, Consolidated Electric Light Co. v. Edison Electric Light Co. (C. C.) 25 F. 719, and United States Stamping Co. v. Jewett (C. C.) 7 F. 869, relied upon by defendant to establish that the legal title to the patents is in the United States, turned upon assignments recorded or suitable for recording, and not upon implications arising from an employer-employee relation. The cases do not, I think, support the contention in aid of which they were cited.

I do not, of course, question the rules of law enunciated in Solomons v. United States, 137 U.S. 342, 346, 11 S. Ct. 88, 34 L. Ed. 667; Standard Parts Co. v. Peck, 264 U. S. *464 52, 59, 60, 44 S. Ct. 239, 68 L. Ed. 560, 32 A. L. R. 1033; United States v. Houghton (D. C.) 20 F.(2d) 434; Id., 23 F.(2d) 386, 390 (C. C. A. 4), certiorari denied 277 U.S. 592, 48 S. Ct. 528, 72 L. Ed. 1004; Goodyear Tire & Rubber Co. v. Miller, 22 F.(2d) 353, 356 (C. C. A. 9); Wireless Specialty Apparatus Co. v. Mica Condenser Co., 239 Mass. 158, 162, 163, 131 N.E. 307, 16 A. L. R. 1170, and British Reinforced Concrete Co. v. Lind, 34 R. P. C. 101, which have to do with inventions made by employees, but they are of no avail to the defendant. The principle laid down in the Yablick Case, and in the other cases there cited, that the government alone can avail itself of the employment of an inventor when the invention was made, is, I think, a broad one, and as applicable to employment in the Bureau of Standards as elsewhere. If the government has had the right to demand and obtain an assignment of the patents in suit, its exercise until now of its option not to do so has not served to confer any rights upon the defendant. On the contrary, the failure of the government to acquire title has completely insulated the defendant from the rights, if any, that it might have had under government ownership of the patents. As against the plaintiffs, defendant has shown under its plea in bar no rights either legal or equitable.

Since an infringement suit in equity may be maintained by the owner of the legal title to a patent, Independent Wireless Tel. Co. v. Radio Corp., 269 U.S. 459, 466, 46 S. Ct. 166, 70 L. Ed. 357, and as it is clear that, if damages and profits are recovered by the legal owners for infringement during such ownership, recovery may not again be had from the infringer for such infringing acts by any subsequent assignee, the United States is not an indispensable party to these suits.

Decrees in conformity herewith may be presented.