The plaintiff, assignee of Dann patent, No. 1,374,553, applied for February 18, 1915, and granted April 12, 1921, has brought this action against the defendant for an injunction and accounting based upon alleged infringement of said patent. The defendant has alleged the invalidity of the patent, in that previous patents cover the invention claimed, and also because the features of the patent were described in printed publications in the United States and foreign countries more than two years before the application therefor was filed. It also denies that its construction infringes plaintiff’s patent.-
The patent in suit is for an improvement in electrical transformer tanks. A transformer is a device used to transform electric energy from one voltage to another. Its form or method of operation is not involved in this action. In the performance of its functions, heat is created in its core and windings, which must be dissipated if the transformer is not to be impaired or destroyed. For ,the purpose of radiating the heat incident to its operation, it has been the practice to immerse the transformer in a tank filled with a light oil. In the operation of a very small transformer, the oil and a tank of an ordinary type were sufficient to dissipate the heat created; but, as the device increased in size and power, it became necessary to increase the cooling power by some means other than the mere increase of the size of the tank. *594This was rendered necessary, in part, by the fact that the volume of any body increases with the cube of one dimension, while the surface increases with the square of the same dimension, or, in other words, because the surface of any body increases more slowly than does its volume. This necessity was met by two methods. By the first, the oil in the tank, in which the transformer was immersed, was artificially cooled by means of a stream of water passed through coils surrounding it. By the second method, the tank was constructed in such a manner as to offer a large amount of surface to the air. Transformers cooled by the first method are known as artificially cooled transformers, while those using the second are called self-cooled transformers. It is with the second class, self-cooled transformers, that our present inquiry is concerned.
By the history of self-cooling transformers prior to plaintiff’s patent it-appears that the first step in advance was taken when the transformer was placed in a smooth iron tank filled with oil. The possibilities of this type of tank having been exhausted, the transformer was increased in size and consequent power, by being placed in a corrugated iron tank. The indentations in this type increased the cooling surface presented by the plain iron tank. The improvement so created was increased by deepening the corrugations until some tanks were indented 15 inches or more. This type having reached its limit, the next step forward consisted in adding wings' or boxes to the main-body of the transformer. These parts, as the corrugations in its predecessor, were added for the sole purpose of increasing the cooling surface. They were- integral parts of the main'tank body. - An illustration of the type is found in figure 13, on page 452, of a publication of the Pittsfield-Sehenectady Midyear Convention of the American Institute of Electrical ■ Engineers. See Defendant’s Exhibit No. A — 1. -
- The next step in advance, in point. of ultimate merit, although perhaps preceding the last-mentioned construction in point of ■time, is found in a construction of the Stanley Electric Company, which, in 1898, made a tank that had external radiating tubes. See Defendant’s Exhibit No. A-12.. These tubes were detachably connected by screw joints with the top and bottom of the maip ■part of the ■ transformer tank,' and extend radially with, respect to the tank. This type has'been greatly advaneed-sinee its first use by increase in the number, and improvement in the shape of the radiator pipes. An example, although not the latest form, of this transformer is found in Harter patent, No. 1,155,656. In the Harter construction the flattened external tubes- are welded to the top and bottom of the main tank body.' In operation the Stanley device is the same as that of plaintiff. The oil in the main tank, heated by the transformer, rose to the top of the tank, and entered the radiator tubes. Becoming cool, it re-entered the main tank at its bottom to take the place of the oil heated by the transformer.
The next step in transformer tank construction was the addition, first, of a single row, and later of a double row, of rectangular radiators, with their greater lengths laid- tangentially to the tank. The operating principle of these tanks was the same as that of the Stanley construction, supra. The oil entered the radiators by means of conduits riveted to the wall of the tank. These conduits, with the pipes entering at the bottom of the tank, supported the radiators. An example of this transformer tank is found in Defendant’s Exhibit No. A-13. This transformer had a capacity of 3,000 k. v. a. It was in operation for several years when the plaintiff’s form of cooling tank was introduced.
In the latter part of the year 1914 the plaintiff company built and installed the transformer tank which is the subject-matter of the Dann patent.
Plaintiff has relied upon claims 7, 10, 11, 12,14, and 15 of the Dann patent. Of these claims we quote claims 10 and 15, as together they comprehend all the material features of Dann’s alleged invention: .
“10. The combination with a device to be cooled, a fluid in which the device is immersed, and a tank containing the device and the fluid, of a plurality of separately detachable flat and elongated radiators secured to the exterior of the tank and having their longer axes respectively extending longitudinally and radially with respect to the tank. * * *
“15. The combination with a device to be cooled, a fluid in which said device is immersed, and a tank containing said device and said fluid-‘and having flanged openings near the top and bottom, of a'plurality.of radiator units' detachably .secured"to said tank and having flanged members at the top and bottom which are adapted to register with, and be secured to, the flanges surrounding said tank openings, whereby the entire weight of each ra,diator..unit:is supported by .-the means which admit of fluid communication between *595said unit and said tank, the vertical dimension of the radiating surface of each of said units being substantially .the same as the distance between the upper and lower openings in said tank.”
The patentee’s problem, which he claims to have solved by the exercise of invention, was to evolve a self-cooling transformer of considerably greater capacity than any of its predecessors in the art that could be conveniently shipped and easily installed. As the main tank inelosing the transformer had about reached its maximum for convenient shipping, the one way open to him was to obtain a greater radiating space in connection with the transformer tank; any increase in the power of the transformer itself being merely a matter of construction. The method adopted by the patentee, at least as viewed after accomplishment, seems exceedingly simple. He provided a number of flanged openings at the top and bottom of an otherwise ordinary containing tank, and to the main tank he conected rectangular radiators by means of strong conduits, which had flanges thereon that registered at the one end of the top and bottom of each radiator, with corresponding flanged openings at the top and bottom of the containing tank. Putting it more briefly, the patentee, with a flanged joint conduit connection in the place of a welded or screw connection, substituted a number of rectangular radiators, each with a large cooling surface, for the simple external single pipe radiators of the 1898 Stanley Electric Company’s construction herein-before mentioned. See defendant’s Exhibit No. A-12. The operating principle of the two devices is identical. The height of these rectangular radiators was approximately the height of the containing tank, and the radial axis of each was greater than its side tangential to the body of the tank.
We have examined the various features of the patent in suit for which novelty is claimed. After a comparison of them with the prior art as shown, we have reached the conclusion that each of said features, considered by itself, was not new when application was made for the patent. These features are, briefly stated, the radial arrangement of the radiators, their height the same as the height of the containing tank, the entire support of the radiators by the connections with the tank, and the detaehability of the radiators. Before considering these features individually, we call attention to the fact that the patent claims no invention in so far as the form of the radiators is concerned, as a casual reading of the specifications of the patent might lead to a wrong impression in this respect.
First, we may consider the claim of novelty in the radial arrangement of the radiators. Such arrangement is so obviously the best for a plurality of external radiators, in the interest of economy of space, that we cannot conceive that any invention can exist in. it. As has been pointed out by counsel, it is but the adoption of the same custom observed by a librarian in placing his books upon the shelf. It is directly anticipated in the transformer art. The Stanley and Harter (No. 1,155,-656) constructions show a series of single pipe radiators attached at the top and bottom of the containing tank. The Stanley tubes were attached to the tank by screw joints, while the Harter tubes were welded to it, but they were attached radially. These tubes were not as effective radiators as that shown as the preferred type in the Dann patent specifications, but in function they were radiators just as much as were those described in the Dann application or the improved form now used by both plaintiff and defendant. It will be noted that the form of the radiator is not the subject of any claim. In Defendant’s Exhibit No. A-3, an article in the General Electric Review of August 15, 1915, and in the following Exhibit, No. A — 4, advertisement in the Electric World of March 22,1913, is a representation of a transformer tank described as “figure 2. A 300 k. v. a. transformer built in 1904. This type of tank has become standard for units of 100 to 1000 k. v. a.” The radiator boxes shown are integral parts of the cooling tank, it is true, but they extend radially from the central part of the tank. Separate them from the tank, connect them by a flanged conduit, and we have the Dann construction. The Evans patent, No. 787,617 (1905), disclosed a boiler, or tank, in which water is heated and circulates through radiators attached edgewise. Incidentally, the radiators were detachably connected. We have been convinced by the testimony offered that nothing of novelty was disclosed by the radial arrangement of radiators of the Dann patent.
Still less does the testimony support the claim of novelty in the height of the radiators. Reverting to the Stanley and Harter types of tank, an examination will show the radiator pipes extending from top to bottom of the tank. In fact,-the principle of operation of all transformer tanks, having external radiators, requires exit of the oil at its hottest at the top of the tank and re-entry of *596it after cooling at the bottom, to cool the transformer. There is no particular advantage in having a rectangular form of radiator have its fop level with that of the tank other than to save ground space. In so far as the radiation of heat is concerned, if one conduit receives the hot oil at the top of the tank and another brings it back to the bottom, it is immaterial whether the radiator is the entire height of the tank or only half the height, if in the latter instance its length is increased to give an amount of radiating surface equal to that of the former. This, of course, assumes that the radiators have the same radial relation to the tank. It follows that the arrangement of the dimensions of the radiator is a mere mechanical matter, and not a result of invention. On page 452 of an article by J. J. Frank and H. O. Stephens in a publication of the American Institute of Electrical Engineers, dated February 15, 1911, Defendant’s Exhibit No. A-l is shown a corrugated steel tank having external radiating boxes extending from top to bottom of the tank. Similar constructions are shown in German publications offered in evidence.
A third novel feature of its patent claimed by plaintiff is the method of attachment by which the connections between the tank proper and the radiators serve both as conduits and brackets. The patent shows flanged openings riveted or welded (riveted is recommended) to the tank. The inlets and outlets of the radiators have flanges which register with the openings in the tank, and, when these flanges are clamped together, the joints are able to carry the radiators without the use of brackets. If the radiators were to be made detachable, it is plain a joint of some kind was necessary. The patentee had in mind the shipment of the containing tank and the radiators separately. In making the connection, by means of a flanged joint, he made use of a method of pipe joinder where a strong joint was required, and this is all he did. The flanged joint was long known in the handling of hot oil. In adopting it rather than a screw joint in making the connection desired, the patentee merely followed the best practice. An examination of the transformer tank shown in Defendant’s Exhibit No. A-13, dating about 1913, will show rectangular radiators attached tangentially to the tank. They are connected to the tank by means of conduits riveted to the tank and screwed to the radiators. It is plain that these radiators could be detached, and that the conduits served in part as brackets. The patentee has called attention to the bracket used to evidence the discovery in his method of joinder, but we find ourselves unable to agree with him. A conduit with a flanged joint is certainly no stronger than a conduit without any joint. Assuming the necessity of a bracket in the transformer tank shown in Defendant’s Exhibit No. A-13 (which is denied), and that the patentee’s connection worked satisfactorily without a bracket, the patentee has simply replaced the bracket of the Exhibit No. A-13 construction with a heavier conduit. A reading of the patent specifications will make it quite plain, we think, that patentee, in making his connection, had a merely mechanical problem for solution. His preferred form shows flanges riveted rather than welded to a bridge plate and the tank to insure greater strength to bear the radiators, and his choice of a heavy pipe and strong joint was a selection of mechanical means of the same kind. The selection was doubtless a wise one, but it was not inventive. The use of flanged joints was well known in the radiator art. The plaintiff’s witness, E. H. Schwarz, had admitted the general use of it. See Evans patent, No. 787,617, Exhibit No. N-8; Tudor patent, No. 559,551, Exhibit No. N-6.
Another element of novelty claimed by the plaintiff is the detaehability of the radiators. . Counsel for defendant has argued that what patentee claims as detaehability in the patent was really attachability. The patentee’s tank was manufactured and shipped without joinder of the tank proper and the radiators. When the tank was installed, the radiators were attached, and thereafter there was little or no need to detach them. The patentee’s actual accomplishment was the separation of the parts of the tank in shipment. This enabled him to put larger radiators in the tank and a larger transformer within it, but the parts, separated in shipment, performed the same functions that they had theretofore performed when assembled. True, a larger transformer and greater radiating space gave more power, but mere increase in the size of parts of a device of well-known functions has never been held to be patentable. See Coffield v. Sunny Line Appliance, Inc. (C. C. A. 6 Ct.) 297 F. 609, and cases cited; Gates Iron Works v. Overland Gold Mining Co., 147 F. 700, 78 C. C. A. 88 (8 Ct.).
As stated, we are of the opinion that all parts of the plaintiff’s construction were old. However, it is sometimes possible that a new combination of old devices may be patentable.
*597In Specialty Manufacturing Co. v. Fenton Metallic Manufacturing Co., 174 U. S. 492, 498, 19 S. Ct. 641, 643 (43 L. Ed. 1058), the rule was stated:
“Where a combination of old devices produces a new result, such combination is doubtless patentable, but where the' combination is not only of old elements, but of old results, and no new function is evolved from such combination, it falls within the rulings of this court in Hailes v. Van Wormer, 20 Wall. 353, 368 [22 L. Ed. 241]; Reckendorfer v. Faber, 92 U. S. 347, 356 [23 L. Ed. 719]; Phillips v. Detroit, 111 U. S. 604 [4 S. Ct. 580, 28 L. Ed. 532]; Brinkerhoff v. Aloe, 146 U. S. 515, 517 [13 S. Ct. 221, 36 L. Ed. 1068]; Palmer v. Corning, 156 U. S. 342, 345 [15 S. Ct. 381, 39 L. Ed. 445]; Richards v. Chase Elevator Co., 158 U. S. 299 [15 S. Ct. 831, 39 L. Ed. 991].”
If some of the claims of the patentee before the Patent Office had been made out in the instant case, it might be properly held that he had disclosed such a combination. It was there found, in ex parte hearing, that, “where * * * those skilled in the transformer art were led by precedent to think of transformer casings as essentially integral and special problems, Such as the difficulty of making oil tight joints and joints of sufficient strength to be self-supporting, had to be overcome, and the history of the art shows that the art remained stationary, as regards the size and capacity of transformers, until the advent of appellant’s (patentee’s) construction when the capacity was increased 67 per cent, and above in a single step, it is reasonable to conclude that the departure from the conventional practice was not an obvious step to take.”
The foregoing quotation is from the opinion of the Examiners-in-Chief of the Patent Bureau accompanying their reversal of the action of the Primary Examiner, who had held the claims not patentable. The allegations of fact contained, therein have not been all established in the present case. We have already discussed the special problems mentioned, and have held them to have been anticipated in the art and known in common practice. The testimony does not disclose any long halt in the increase in the size of transformers. It is true that no transformer of greater capacity than 3,000 k. v. a. had been built prior to the application for the patent in suit in 1915, and that, after that time, the plaintiff built a transformer of 5,-000 k. v. a.; but it had not been established that users were demanding greater capacity in transformers and the manufacturers were unable to meet the demand. As a matter of fact, the transformers had been steadily growing in size and capacity, and the 3,000 k. v. a. tank was put out about 1913. The General Electric Company, in an advertisement in the Electric World of March 22, 1913, said:
“ * * * We are now building single phase self-cooled transformers in 3,000 k. v. a. units and are prepared to build larger sizes. * * * A new impetus is thus given to the large isolated substation and the problem of cooling large units in localities not adapted to water systems is becoming simplified.”
The testimony. further tends to contradict the declaration in the opinion of the Chief Examiners to the effect that makers and purchasers were all obsessed by the idea that transformer casings were essentially integral. Top and bottom covers had long been separated from the main tanks, as well as a variety of smaller fittings. In 1907 the Allis-Chalmers Company, of Cincinnati, built and sold to the Pittsburgh Railways Company three transformer tanks, which were too large for convenient shipping without separation of parts. The manufacturer therefore shipped each tank in two sections of about equal size, and, when installed for use, the upper part was connected with the lower by means of a flange joinder. See Defendant’s Exhibits I (photograph) and H (drawing), and testimony of Louis C. Nichols. If we assume that the transformer art had been at a standstill for a number of years, and that by the patentee’s arrangement alone manufacturers had been immediately enabled to pass the old barrier and make transformers nearly twice the capacity of their previous best efforts, we are easily led to the belief that patentee’s arrangement, though the elements were old, contained something of novelty and value that was patentable. A somewhat careful study of the prior art, and the condition of the transformer art when the application for patent was made, as shown by the testimony, has led us to revise our first impressions in this respect. No barrier actually existed to prevent increase in the size of self-cooled transformers, and no such combination of old elements was made as led to an unforeseen result. So far as plaintiff’s device is concerned, its combination, although of greater power by reason of increased size, in no respect meets the demands of Hailes v. Van Wormer, 87 U. S. (20 Wall.) 353, 22 L. Ed. 241; Pickering v. McCullough, 104 U. S. 310, 26 L. Ed. *598749; Richards v. Chase Elevator Co., 158 U. S. 299, 15 S. Ct. 831, 39 L. Ed. 991, and numerous other cases. Merely improving a machine in the production of well-known results is not sufficient to make patentable a rearrangement or aggregation of old elements. Anton v. Grier Bros. Co., 185 F. 796, 108 C. C. A. 173 (3 Ct.)
We are of opinion that plaintiff’s patent is invalid for lack of invention, and that the bill must be dismissed.
Having reached the conclusion just stated, we have not felt it necessary to give any extended attention to the defense of noninfringement. Our examination of defendant’s device has impressed us with the belief that it is covered by the claims of the patent under consideration herein; but, such claims having been held invalid, it is unnecessary to so decide at this time.
Let a decree be drawn in accordance with the findings of the foregoing opinion.