This appeal calls for review of a decision of the Patent Office Board of Appeals, adhered to on reconsideration, wherein the board sustained certain of the examiner’s rejections of various claims in appellant’s application serial No. 699,563, filed November 29, 1957, for “Laminated Cores.” Two claims were allowed by the examiner, and the board reversed the examiner’s rejection of four more. The claims urged on this appeal, therefore, are 21, 22, 24-26, 28-30, 32-35, 37-43, 46 and 47. The board, in addition to affirming the examiner’s rejection of all these claims as unpatentable over the prior art, itself rejected claims 39, 46 and 47 for failure to comply with 35 U.S.C. § 112. That rejection was made in accordance with Rule 196(b).
The following background is helpful in understanding appellant’s claimed invention : Magnetic materials which are commonly used to form laminated cores in many electromagnetic inductive devices normally have a permeability curve (flux density plotted as a function of magnetizing force) which is not linear. For example, appellant’s specification shows a permeability curve for a ring-shaped lamination of a typical 50/50 alloy of iron and nickel. The curve, rather than approximating a straight line, has two distinct “knees.” In the region of the lower knee (at low values of magnetizing force) the flux density is less than it would be if the curve were a straight line, and in the region of the upper knee (at somewhat higher values of magnetizing force) the flux density is greater than the corresponding straight line value.
In many applications, such as in synchro resolvers and other precision inductive devices, this nonlinearity is undesirable because it introduces harmonics. Appellant’s goal, therefore, is to raise the lower knee and depress the upper knee of the permeability curve, so as to make the curve as linear as possible over the working range of the device. As stated in his specification:
The objects of this invention are accomplished by brinelling selected areas of the laminations which make up the core of the transformer, or the rotor or stator of the synchro. This brinelling is an operation whereby portions of the lamination are dimpled by being subjected to compressive forces which compress a local area in the annealed lamination and cause the metal to flow, and thereby subject the area of the lamination surrounding the compressed or dimpled portion to tension stress. The portion of the lamination which is under the highest tension has the highest permeability at lower magnetizing forces, and the portion of the lamination which is most severely deformed by the compressive force applied to it under the brinelling tools is physically much harder than before and magnetically less permeable in the higher ranges of magnetizing force. By controlling and selecting the relative compressive and tension areas of lamination and the amount of stretching under the brinelling tools, the desired result of making the permeability curve more linear is obtained.
Claim 21 is fairly illustrative of those on appeal and reads as follows:
21. A method of forming an electrical inductive device, comprising brinelling selected areas of an annealed ferromagnetic lamination in order to introduce tension stress into the bulk of the lamination to improve the magnetization curve of said device over its operating range and to suppress third harmonics, and to increase permeability at low values of magnetizing force and decrease it at somewhat higher values of magnetizing force so as to im*606part to said lamination a permeability curve which is more linear than a similar lamination which has not been so treated, and assembling a plurality of the treated laminations in an inductive device.
Claims 22 and 24 are drawn to a similar method, except that they call for “coining” and simply “stressing,” respectively, in place of brinelling. Claim 32 recites a “process of introducing tension stress into selected areas” of a lamination by brinelling. With minor variations, similar language is used in other claims; thus: “shot blasting” (claim 29), “sandblasting” (claim 30), “coining” (claim 33) and “compressively deforming” (claim 28). Claim 38 is similar to claim 28, except that it begins “A process of treating a ferromagnetic lamination”; in like manner, claim 40 is similar to claim 24. Claim 43 recites “in the method of improving the magnetization curve of a ferromagnetic lamination,” the step of compressing portions of the surface thereof.
Claims 34 and 35 define the invention in a somewhat different way, reciting a “new use” of the processes of brinelling and coining, respectively. Claims 25, 26, 41 and 42 claim an annealed ferromagnetic lamination per se, always with the limitation that the permeability curve be more linear than that of a similar unstressed lamination. Claim 37 recites “An improved electrical inductive device” having such laminations. Finally, claims 39, 46 and 47 recite “A method of improving the magnetization curve of an electrical inductive device,” and include specific limitations as to alloy and values on the magnetization curve.
The references finally relied on by the Patent Office are Garbarino patent No. 2,565,303 issued August 21, 1951, and Neurath patent No. 2,920,296 issued January 5, 1960. The examiner rejected all of the appealed claims as “unpatentable over” either Neurath or Garbarino. The board sustained the rejection of all claims on Garbarino but reversed the rejection based on Neurath as to all claims except 26 and 37. From the tenor of the examiner’s answer and the board’s decisions, it is evident that the rejection was one for obviousness under 35 U.S.C. § 103. We are thus required to determine whether the board was correct in holding that the subject matter defined by the appealed claims would have been obvious, as a whole, to a person having ordinary skill in this art at the time the invention was made (here the filing date).
Garbarino is directed to reducing the noise caused by vibration at butt gaps in laminated magnetic cores. As Garbarino puts it:
An analysis which I have made shows that the magnetic flux crossing from one laminar layer to another to by-pass a butt gap has its greatest density close to the butt gap. In other words, most of the flux in any given laminar layer waits until it reaches the very edge of the butt gap before it crosses in a perpendicular plane to the adjacent laminar layers, and very little of the flux passes over to the adjacent laminar layers before reaching the edge of the butt gap. This flux passage at the edge of the butt gaps causes mechanical forces which can be shown to be proportional to the square of flux density. It is this concentration of flux at the edge of the butt gaps which is an important factor in causing the laminar vibrations which raise the noise level of the magnetic core.
Garbarino’s proposed solution to the noise problem is
* * * to provide a high reluctance region in the immediate vicinity of butt gaps between magnetic laminations in order that most of the magnetic flux approaching a gap in any given laminar layer will cross to adjacent laminar layers before reaching the immediate vicinity of the butt gap, and thereby reduce the vibratory noise level of the magnetic core.
Garbarino increases the reluctance (decreases the permeability) at the butt gap by bevelling the edges of the laminations, *607cutting, grinding, coldworking or scoring.
We cannot agree with the Patent Office that the differences between Garbarino and the appealed claims are such that the claimed subject matter as a whole would have been obvious. Unquestionably, Garbarino teaches that the permeability of annealed ferromagnetic laminations can be decreased by introducing strain. But there is absolutely no teaching that selective stressing will result in a more linear permeability curve. Appellant’s specification concedes that it is well known that “the permeability curve of many magnetic materials can be altered by stress.” Indeed, the relevant teachings of Garbarino amount to no more than confirmation of such fact. But appellant’s specification goes on to state:
The object of this invention is not particularly to increase [or, we might add, decrease] the permeability of laminated cores but is rather to increase permeability at low values of magnetizing force and decrease it at somewhat higher values of magnetizing force in such way as to make the permeability curve much more linear over the working range of the inductive device. * * * [Emphasis added.]
Apparently the solicitor also had some misgivings regarding the adequacy of the Garbarino reference, for in his brief he states:
Appellant * * * discloses no criterion for selection of areas to be compressed. The Examiner * * * noted that the instant claims are drawn in terms so broad that the “selected areas” could include either the entire area of the lamination or an area small as a pin point. There is nothing in the specification to suggest that either so large or so small an area could achieve appellant’s desired result of a more linear permeability curve. In any event, the term “selected areas” in these claims does not distinguish from the areas selected by Garbarino in his disclosure.
What the solicitor seems to be suggesting is that the specification is defective in showing how to use the claimed process or how to make the claimed article, in precisely those particulars wherein the claimed invention differs from Garbarino, viz. the concept of stressing to make the permeability curve more linear vis-a-vis the concept of stressing to increase or decrease permeability in gross. We note, however, that except for claims 39, 46 and 47, there is no outstanding rejection under 35 U.S.C. § 112. We have no choice, therefore, but to treat the specification and remaining claims as in complete compliance with the requirements of that section. And when this is done, it is clear that the claimed subject matter would not have been obvious from the teachings of the Garbarino patent.
The Neurath patent is concerned with minimizing noise in transformer cores, where the noise is due to magnetostriction. Magnetostriction is a dimensional change which many magnetic materials undergo when subjected to a magnetizing force. Neurath recognized that magnetostriction is greatly increased when magnetic core materials are subjected to. compressive stresses during the assembly-process. His solution was to coat the. core materials with a suitable varnish, or other substance before annealing so. as to introduce tensile stresses in the material. Such pre-tensioning counteracts, the compressive stresses encountered later in the assembly process and results, in a material having low magnetostriction.
Like Garbarino, Neurath contains no. recognition that the linearity of permeability curves can be improved by selective stressing of magnetic materials. We think the board erred in holding that the subject matter defined by claims 26 and 37 would have been obvious in view of the Neurath patent.
There remains the problem of claims 39, 46 and 47, which were re*608jected by the board for noncompliance with section 112. After receiving the board’s rejection of those claims, appellant offered an amendment in an attempt, presumably, to cure the trouble. The record before us is unclear as to whether such amendment, or any other, was ever entered or, if it was, what its effect may have been; it seems likely, however, that the version of claims 39, 46 and 47 which presently appears in the record was not the version on appeal before the board. In view of this confused state of affairs, we must remand the case to the board for further consideration of claims 39, 46 and 47 in connection with the section 112 rejection.
The rejection of all the appealed claims under section 103 is reversed, and the case is remanded with respect to claims 39, 46 and 47 for further proceedings consistent with this opinion.
Reversed and remanded.