This appeal is from a decision of the Patent Office Board of Appeals which affirmed the examiner's rejection of claims 1-5, inclusive, of appellant's application serial No. 536,086 for a method of forming hollow, frusto-conical articles from flat metal stock. The claims were rejected under 35 U.S.C. § 103 as being obvious in view of the following references:

  Kelley et al.   1,901,035   Mar. 14, 1933
  Lindgren       Re. 19,790   Dec. 17, 1935
  British Patent    436,768   Oct. 17, 1935

Claim 1 is typical of the claims on appeal and reads as follows:

"1. The method of making a hollow generally conical-shaped article having a tapered side comprising the steps: working a blank having an annular portion of tapering cross [section]1 by mounting said blank on a spindle having a conical-shaped portion and while the spindle and blank are rotating, axially displacing said annular portion by making a roller move in a path non-parallel to the surface of the conical-shaped portion of the spindle so that the thickness of the side being formed taken at any point thereon in a direction generally parallel to the rotational axis of the spindle is equal to the corresponding thickness of the original blank."

Appellant's claims 1-3 are similar and for present purposes can be considered as one. Claim 4 recites the steps of making a cup-shaped blank with a tapered cross-section and claim 5 requires use of a cup-shaped blank with tapered cross-section.

The sole issue for determination is whether the board was correct under 35 U.S.C. § 103 in holding that appellant's claimed invention was obvious at the time it was made to one of ordinary skill in the art in view of the disclosures in the cited references.

Appellant's invention relates to a method of forming a hollow, frusto-conical article with a tapered wall from a flat or cup-shaped blank by spinning the blank and axially displacing the layers or strata of metal with a roller which moves relative to the face of a rotating spindle.

Appellant's spinning method is similar in general to that used in the metal spinning arts, as shown by the Lindgren and British patents. A round metal blank is first clamped on the top of a spindle having a frusto-conical shape. As the spindle and blank are rotated, the roller pushes the metal along the tapered sides of the spindle to form a hollow frusto-cone and the metal is deformed to the shape of the spindle by axial displacement. To understand that operation it is necessary to visualize the starting blank as being composed of an infinite number of layers or strata lying in a plane parallel to the axis of the spindle. The roller is so directed during the spinning operation that the metal strata of the blank are axially displaced relative to each other, i. e., they slide relative to each other and are not compressed or elongated in length. Thus at any point on the surface of the spindle, appellant contends, the thickness of the frusto-cone wall, measured in a direction parallel to the spindle axis, is equal to the thickness of the original blank, measured in the same direction at an equal radial distance from the rotational axis. In forming a hollow frusto-cone by a spinning method utilizing the principles of axial displacement, as disclosed by the prior patents, the inner surface of the wall of the hollow frusto-cone is defined by the spindle surface, and the outer surface of the cone wall is defined by the path of the roller which moves in a path parallel to the outer surface of the spindle.

Lindgren and the British reference, supra, both disclose the general method described. Appellant, however, predicates novelty and unobviousness of his method on the asserted fact that the methods of such references are unable to form by spinning a metal frusto-cone with a tapered or non-uniform wall thickness following the principle of axial displacement while he is able to do so. In appellant's method, he first rolls or otherwise forms a tapered or cup-shaped blank from a flat piece of sheet metal. This tapered or cup-shaped blank is then spun on a frusto-conical spindle. To form a frusto-cone with a tapered wall, appellant's roller is positioned to move along a path non-parallel to the spindle surface, so that it comes closer to the spindle as it moves along the spindle surface. Appellant asserts that this method results in axial displacement of the metal blank since the non-parallel roller path is so designed that the thickness of the tapered wall, measured in a direction parallel to the spindle axis, is the same as the thickness of the tapered or cupshaped blank, measured in the same direction at the same radial distance from the spindle axis.

Appellant states his position in his brief:

"The change over the references is a simple one, the change being in the use of the tapered blank and in a non-parallel roller path being chosen so that the thickness of the side being formed, taken at any point on the side in a direction parallel to the axis of the spindle, is equal to the corresponding thickness of the tapered blank."

The sole issue thus is whether, in using a blank having a tapered cross section and a roller path not parallel to the conical surface of the spindle, appellant has done anything more than would have been obvious to a worker of ordinary skill in the art at the time appellant's invention was made, having before him the method of forming cones taught by Lindgren and the British patent, and faced with the problem of adapting that method to forming a cone with a tapered wall.

The board held that the use of a blank with a tapered cross section and a roller path not parallel to the conical surface of the spindle is a mere extension of the idea of axial displacement taught by the Lindgren and British references and was, therefore, but an obvious adaptation of that idea to the specific shape desired in the finished article and that the steps of forming a tapered blank from flat metal stock is fully shown by the Kelley et al. reference.

Appellant does not predicate any patentable significance on the steps relating to forming a tapered blank from flat stock.

Appellant's contention is that the references do not suggest how the methods as therein disclosed can be used for forming cones with a tapered wall without relying on appellant's disclosure. Specifically, appellant contends that Lindgren's discussion of axial or strata displacement where the blank is flat and the roller moves parallel to the spindle surface and is spaced therefrom at a distance t₂=t₁ sine k is not applicable to the process of forming a frusto-cone with a tapered wall.

In determining the issue here we are required by 35 U.S.C. § 103 to do so from the vantage point of one having ordinary skill in the metal spinning art and then to determine whether or not the claimed invention would have been obvious to such a person at the time the invention was made. This requires us to view the prior art without reading into that art the teachings of appellant's invention. In re Murray and Peterson, 268 F.2d 226, 46 CCPA 905.

Neither appellant's application nor the record before us give us anything of a factual nature upon which to ascertain the problem faced by the prior art in spinning a frusto-cone with a tapered wall. Assertions are made by appellant's counsel that the problem was to prevent swaging in the metal as it was spun and was solved by appellant through use of 1) a tapered blank and 2) a non-parallel roller path. These assertions have not been disputed by the examiner or the board, nor has any reference been cited which shows the formation by a metal spinning method of any article having a tapered wall. In judging what requires uncommon ingenuity, the best standard is what common ingenuity has failed to contrive. Western States Machine Co. v. S. S. Hepworth Co., 2 Cir., 147 F.2d 345, 347. In the absence of any contrary showing in the record before us, we must assume that no one before appellant had ascertained, as he did, that a frusto-conical article with a tapered wall could be axially spun by use of 1) a tapered blank and 2) a roller path non-parallel to the surface of the forming spindle.

Once appellant's solution to the problem of making a tapered wall frusto-cone is disclosed, it is easy to see how the prior references can be modified and manipulated to produce this type of cone. The change admittedly is simple and by hindsight seems obvious. However, the simplicity of new inventions is oftentimes the very thing that is not obvious before they are made. This court, in In re Osplack, 195 F.2d 921, 39 CCPA 932, stated:

"We think this case is one of that category of inventions which, when viewed after disclosure and explanation by an applicant, seem simple and such as should have been obvious to those in the field. Yet this does not necessarily negative invention or patentability. Goodyear Tire & Rubber Co., Inc. [et al.] v. Ray-O-Vac Co., 321 U.S. 275, 64 S. Ct. 593, 88 L. Ed. 721; In re DeLancey, 159 F.2d 737, 34 C.C.P.A. (Patents) 849. Indeed, simplicity may even be some evidence of invention. Baldwin-Southwark Corporation v. Tinius Olsen Testing Mach. Co. [et al.], 3 Cir., 88 F.2d 910." The fact that the invention seems simple after it is made is not determinative of the question of obviousness. If this were the rule, many of the most beneficial patents would be stricken down. If those skilled in the mechanical arts are working in a given field and have failed to discover a certain new and useful improvement, the one who first makes the discovery frequently has done more than make an obvious improvement which would have suggested itself to a mechanic skilled in the art, and such an invention is entitled to the grant of a patent thereon. Expanded Metal Co. v. Bradford, 214 U.S. 366, 29 S. Ct. 652, 53 L. Ed. 1034.

The solicitor, in answer to appellant's assertion that the Lindgren formula t₂=t₁ sin K is not applicable to spinning an article with a tapered wall, has submitted a drawing based upon Lindgren's figure 6 to show that the formula can be made applicable to appellant's process. While we agree with the board and solicitor that the formula disclosed by Lindgren could be used by appellant to define the relation between the roller path distance and the blank width in a direction parallel to the axis of rotation at any point on the surface of the spindle, we do not think that this conclusion is determinative of the issue of obviousness. Lindgren (and the British reference) discloses spinning processes where the roller path is parallel to the spindle surface and the blank thickness is constant. The formula disclosed by Lindgren is useful in ascertaining the roller path in one calculation only where the blank thickness, t₁, remains constant. Appellant's non-parallel roller path cannot be found unless the formula of Lindgren is applied several times along the radial width of the tapered blank, which tapered blank is not shown by Lindgren or the British reference.

Lindgren shows in figure 4 a product which has a cone wall of constant thickness with a lower or outer flange of thickness greater than the cone wall. It is clear that Lindgren did not contemplate use of a blank of variable thickness to obtain such a product, since he states:

"If the desired thickness of the outer flange is less than of the original material but greater than that of the conical mantle, I produce, by spinning, a semi-manufactured disk in the shape of two conical mantles, one below the other, the top angle of the lower mantle being made greater than that of the upper mantle." [Emphasis added.]

It is also clear that the British patent did not contemplate use of a tapered blank to vary the wall thickness of the product. This patent states:

"The thickness of the [cone wall] material will of course vary and depends at each point on the angle of inclination." [Emphasis added.]

We are unable to find in the art relied upon by the board any teaching prior to appellant's invention that a metal spinning process could be used to provide the desired hollow conical-shaped article having a tapered side. Nor do we find any teaching in the art of the use of a tapered blank of variable thickness in order to form a spun article with a tapered or variable thickness wall.

Obviousness is a legal conclusion which we are required to draw from facts appearing in the record or of which judicial notice may be taken. Thus before we can conclude that any disclosed invention is "obvious" under the conditions specified in 35 U.S.C. § 103, we must evaluate facts from which to determine 1) what was shown in the prior art at the time the invention was made, and 2) the knowledge which a person of ordinary skill in the art possessed at the time the invention was made. Here, neither the record nor the facts of which we are able to take judicial notice supplies the factual data necessary to support the legal conclusion of obviousness of the invention at the time it was made. We are unwilling to substitute speculation and hindsight appraisal of the prior art for such factual data. For this reason we think there is a doubt as to the factual basis supporting the conclusion of the board of appeals that the invention would have been obvious to one of ordinary skill in the art of metal spinning. Under these circumstances, the doubt should be resolved in favor of the applicant. In re Devine, 261 F.2d 240, 241, 46 CCPA 725; In re Altmann and Bureau, 264 F.2d 894, 46 CCPA 818.

For the foregoing reasons, the decision of the Board of Appeals is reversed.

MARTIN, Judge (dissenting).

I disagree with the majority because I believe that one skilled in this art who wanted to manufacture a hollow generally conical-shaped article having a tapered side by means of the axial displacement procedure, would know that a tapered blank should be used and that the path of the roller should not be parallel to the spindle. I find nothing else in this application upon which patentability can be predicated.

KIRKPATRICK, Judge (dissenting).

I cannot agree that the appellant's contribution is anything beyond what could be expected of a person having ordinary skill in the art. The Lindgren and British references fully disclose the metal spinning technique used by the appellant to make a frusto-conical article from a flat disk. As described in the majority opinion, a round metal blank is first clamped on the top of a spindle having a frusto-conical shape. As the spindle and blank are rotated, a roller pushes the metal down along the sides of the spindle to form a hollow frusto-cone. If the finished article is to have tapered sides, it is hard for me to think of anything more obvious than to use a tapered disk to start with — the first of the two steps which the appellant says constitute the advance over the prior art made by his invention.1 Being provided with a tapered disk to work on, the necessity of moving the roller in a path not parallel to the sides of the spindle (the second step claimed by the appellant as novel) would seem to be equally obvious. In fact, if the taper of the disk is to be carried over into the finished article, it could not be done in any other way. In short, I cannot find anything unobvious about the features which the appellant asserts constitute an advance over the prior art.

The record supports the majority's finding that what the appellant did is new, and its utility has not been challenged. However, novelty and usefulness do not add up to patentability. The majority opinion speaks of the appellant's "solution to the problem of making a tapered wall frusto-cone." It is true that in a doubtful case the fact that an invention solves a problem in the art may sometimes be taken as evidence that it was not obvious, but that is only where there is evidence of an old and recognized need which the invention satisfies and which others have tried without success to supply. In the present case, if any such need existed, the record does not show it, nor does it appear that anyone ever tried or wanted to do what the appellant did. So far as the record shows, the "problem" which the appellant is said to have solved was merely that of accomplishing by a simple adaptation of known methods, something which he set out to do. The farmer who sets out to construct a fence around his ten-acre field has that kind of "problem," and he solves it when he completes the job. The fence and the way in which the farmer goes about building it will almost certainly differ from all other fences, but the fact that the desired result has been accomplished is no evidence that it was not perfectly obvious.

The majority, to fortify its conclusion, invokes the principle that a doubt as to patentability should be resolved in favor of the applicant. It is too late for me to question the existence of that rule, although I think it wrong to clothe an invention of admittedly doubtful validity with a strong presumption that the patent based on it is valid — a presumption which of itself, without more, puts the owner in a position to compel others to defend their use of it in a usually costly and always troublesome infringement suit. However, in the present case I feel that the invention before the court is so plainly an obvious one that there exists no doubt whatever which would call the rule referred to into play. Such doubt cannot be created by novelty and utility alone.

I would affirm the decision of the Board of Appeals.