delivered the opinion of the court:
This appeal is from the decision of the Patent Office Board of Appeals affirming the examiner’s rejection of claims 9, 11, 12, 14, 17, 29, and 30, all remaining claims in appellant’s application serial No. 363,561, filed April 29,1964, for a “Vibratory Motor.” We reverse.
THE INVENTION
The invention relates to the kind of vibratory motor used in hand-held massagers, sanders, clippers, and the like. Appellant claims a motor which employs conventional alternating “house” current to impart vibratory motion to the spring-mounted armature of an E-shaped electromagnet. This vibrating armature may be mechanically comiected to various workpieces, which will then reciprocate as the armature vibrates. Figure 8 of the specification, said to show structure useful as a hair clipper motor, in which 72 is the spring-mounted armature and 21A is the electromagnet having coil 71, is illustrative:
*1193The nub of the invention is the provision on the armature of one or more “wraparound” extensions such as 77 and 78 1 which overlap the adjacent arms 73 and/or 75 of the E-shaped core at all times during the armature’s vibration. The wraparound armature extension and the outer arm of the core define, in the words of claim 9,
* * * a flux gap * * * as small as possible so long as no mechanical interference is created between the armature and core during operation of the motor.
All the other claims contain the same or similar language. The purpose of this construction, we are told, is to reduce the reluctance in the electromagnetic circuit, thereby causing the motor to heat less in operation with resultant increases in the life expectancy of the motor and in the consumer’s comfort when the motor is employed in hand-held devices.
During prosecution, appellant submitted four Buie 132 affidavits. In the first of these, the inventor himself averred that his invention was novel and “unexpectedly resulted in the generation of substantially less heat” than vibratory motors in which the armature and core were not formed in the manner claimed. In the next two, employees of Dor-meyer Industries, Chicago, Illinois, owned by the assignee of the instant application, described the results of an experiment in which vibratory motors like those depicted in Fig. 8 were operated as shown and with the outer core arms 73 and 75 removed at the working gap face on the base 76. Briefly, these affidavits indicate that vibratory motors in which the outer arms had been removed drew more current and heated up to a higher temperature than otherwise identical vibratory motors in which the outer arms had not been removed. The fourth, submitted with a Bequest for Keconsideration before the board, attempts to distinguish the references and is not relevant here.
THE REJECTION
The references are:
Hartwell - 1,259,396 Mar. 12,1918
McElroy et al- 2,343,237 Mar. 7, 1944
Wallace - 2,451,789 Oct. 19, 1948
Lewandowski- 2,451,840 Oct. 19, 1948
Caracciolo - 3,194,231 July 13, 1965
The examiner rejected claims 2 (since cancelled), 9, and 11 on McElroy on Caracciolo in view of Wallace and Hartwell and claims 12,14,17, 29, and 30 on McElroy or Caracciolo in view of Wallace and Lewan-dowski. The board affirmed the rejection of all claims, referring additionally to section 5-49 of the McGraw-Hill Book Company’s Standard Handbooh for Electrical Engineers (9th ed. 1957).
*1194McElroy and Caracciolo disclose electromagnetic, vibratory motors powered by alternating current and having armatures spring-mounted on E-shaped cores. Appellant concedes that these features of his claimed invention are old and has relied for patentability solely on the minimization of the flux gap permitted by his wrap-around armature extensions; accordingly, it is unnecessary to consider these references further.
Wallace discloses a vibratory armature electromagnetic motor (see Fig. 2 below) powered by interrupted direct current and having a spring-mounted armature A which parallels the core of the electromagnet,
* * * is pivoted at one end b on one of the pole pieces (P2) and is provided at the opposite end with an extension a overlying the top of the opposite pole piece (Pi) in such manner as to provide a relatively small air gap between the pole piece Pi and the armature extension a over the relatively large pivotal angle of movement of the armature.
Armature A is shown in its outward position. Through linkage L it it reciprocates piston D of a small pump.
Appellant objected to the examiner’s reliance on Wallace on the ground, inter alia, that it disclosed a motor powered by interrupted direct current in contrast to his ac-powered device, and Hartwell was thereupon cited solely as a teaching of the adaptability of an electromagnet such as that used by Wallace to either alternating or direct current operation.
*1195Lewandowski discloses an electromagnetic motor in which the armature parallels the core of the electromagnet for most of its length, but has extensions “wrapping around” the electromagnet at either end. It does not, however, suggest in any way that these wraparound extensions should be kept as close as possible to the core.
The cited portion of the Standard Handbook for Electrical Engineers, referred to by the board, states:
In cases where it is desired to Obtain a greater force at the initial .position at the expense of a smaller force in the final position, portions of the armature or core are so arranged that the distance between the coacting parts shall be relatively small at the initial position of the armature, and then a portion of the magnetic flux is increasingly shunted at an angle to the direction of travel of the armature as the latter moves toward the core end or stop.
The board drew from this a presumption that Wallace used a relatively small air gap to serve “the useful function of increasing the initial pull on the armature” as shown in two of the figures in the Handbook.
OPINION
The board stated that:
It is obvious, even without the affidavits, that reducing the length of the air gaps in a magnetic circuit would decrease the reluctance thereof, and, with all other conditions being the same would reduce the heating produced. This follows from elementary magnetic circuit consideration. Thus, it seems to us that one skilled in the art would design the magnetic circuit of McElroy et al. to have the 'lowest reluctance consistent with the use to which the device is to be put. This would include placing the arm 61 as close to the core leg as practicable.
Unfortunately, the board did not set forth the “elementary magnetic circuit consideration” which it thought suggested appellant’s expedient.
The brief for the solicitor is equally unhelpful. The sum total of the analysis of this problem contained therein is the following paragraph:
With regard to the improved heat relationship, shortening the space across which the armature has to be drawn will quite expectedly improve the heat problem of the motor. It is elementary 'that decreasing the work load of any motor will reduce motor heating. Shorten the space across which am armature is to travel, and .the electromagnet will be cooler because it had to d'o less work. That is fundamental to all machines regardless of their type, and as such, appellant has done no more than to state the obvious.
However, we do not understand the removal of the outer core arm in the experiments described in appellant’s affidavits to have had any effect on “the space across which the armature * * * [was] drawn.” That parameter seems to be set by the physical characteristics of the armature and armature mounting, the armature traveling from its rest position to contact with the core •( shown in Fig. 8, supra, as having a resilient bumper 79 at the point of contact) each time the device is actuated.
*1196Wallace is clearly the reference most relevant to the critical issue of the obviousness of minimizing the amount of air in the magnetic circuit in order to reduce the operating temperature of the device, although Wallace’s device differs considerably in detail from the embodi-much of appellant’s invention depicted in Fig. 8.2 Specifically, Wallace states that the armature extension a in the device depicted in Fig. 2 overlaps the pole piece “in such manner as to provide a relatively small air gap between the pole piece Px and the armature extension a over the relatively large pivotal angle of movement of the armature.” However, Wallace does not indicate why this design feature is desirable, and, while we agree that his disclosure makes the use of a similarly minimized air gap in appellant’s structure prima facie obvious, we do not think that Wallace anticipates the feature upon which appellant relies for patentability. Accordingly, we find apposite appellant’s attempts to rebut the inference of obviousness to be drawn from the close structural similarity of Wallace’s device to his, and, as appears below, we find those attempts to be dispositive of this case.
According to appellant’s affidavits, two devices of the type disclosed in Fig. 8 were tested on ordinary house current and found to operate at 148° F. and 140° F. while drawing .175 amperes and .171 amperes, respectively. Two other such devices, identical to the first except that the air gap in the magnetic circuit had been increased by removing the outer core arms 73 and 75, were also tested on ordinary house current, and they were found to operate at 185° F. and 175° F. while drawing .225 amperes and .213 amperes, respectively.3 This difference is obviously advantageous in a hand-held device and is, we think, enough to rebut prima facie obviousness if the result exemplified by appellant’s experiment would not have been fairly predictable to a person of ordinary skill in the design of electromagnetic motors.
*1197As remarked above, tire Patent Office has not been of much help in aiding us to determine whether it would have been fairly 'predictable to such a person that decreasing the flux gap in the magnetic circuit of an electromagnetic motor of the type shown in Fig. 8 would result in cooler operation, and in fairness we add that the appellant’s brief is no better than the solicitor’s. Thus, we have been left to make an unescorted venture into relatively sophisticated electromagnetic theory. Therefore, we will set out our reasoning in some detail in order to make clear the considerations which have led us to reverse the rejection. Under the circumstances, we cannot guarantee technical accuracy; we can only guarantee that we have done our best with a skimpy record.
As we understand the operation of appellant’s devices, there are three sources of heat — magnetic, electrical, and mechanical — the relative importance of which depends upon factors not recited in his claims and which could therefore vary widely in different devices within the scope of his claims. However, based on this record and our admittedly limited knowledge of electromagnetic theory, it seems to us that a person of ordinary skill in the design of electromagnetic motors would have expected all three sources to increase their steady-state heat output with decreases in magnetic circuit reluctance.
The first of these heat sources is magnetic loss, the combination of hysteresis loss and eddy-current loss. In a given device of this type, subjected to cyclic magnetization and demagnetization by means of periodic excitations, these losses are a function of maximum flux density and maximum magnetic potential gradient. Beducing the reluctance in the magnetic circuit should increase the magnetic flux in the circuit, which should in turn increase both the flux density and the magnetic potential gradient in the unchanged portions of the magnetic circuit, leading to an increase in core heating. See section 3-6, “Energy Loss in Ferromagnetic Cores,” in Chapter 3, “Alternating-Current Excitation of Ferromagnetic Structures,” in Electromechanical Energy Conversion by Vembu Gourishankar (International Textbook Co., Scranton, Pa., copyright 1965).
The second heat source is resistance loss in the electrical circuit used to excite the electromagnet. Although the voltage applied to the circuit in appellant’s experiments was the usual sinusoidal, 60 cycle, 115 volt ac of the domestic supply, the induced electromotive force caused by changes in the flux linkage must have caused the waveform of the exciting current to vary considerably from the sinusoidal. Nevertheless, one would expect the increased flux resulting from the decreased reluctance to result in an increase in the root-mean-square value of the current flowing in the core, which would in turn result in an increase in coil heating. See section 3-4, “Waveforms of Exciting Current in a Ferromagnetic System with Sinusoidal Flux” in Gouri-shankar, op. cit. supra, particularly Figs. 3-6 and 3-8.
*1198The third source of heat is the flexing of the spring, resulting in mechanical heating. The number of times per second the spring is flexed is, of course, set by the frequency of the exciting current, which was constant throughout appellant’s experiments. However, the mechanical heat generated by flexing the spring would seem to be a function, not only of the rate at which the spring is flexed, but the angular velocity which the spring attains intermediate its points of instantaneous rest. Here again, reducing the reluctance in the magnetic circuit might be expected to increase the flux and therefore the force with which the armature is attracted to the electromagnet on each cycle, resulting in higher angular velocity flexing of the spring and thus in a higher, not lower, operating temperature.
The above analysis suggests that appellant’s expedient, whatever its other merits which may have led Wallace to adopt the same expedient in a slightly different device,4 should have resulted in a vibratory motor the steady-state operating temperature of which was higher than the steady-state operating temperature of an otherwise similar device not employing his expedient. However, appellant’s affidavits, the accuracy of which has not been challenged by the Patent Office, indicate that his invention in fact causes such devices to operate at a lower temperature than they otherwise would. Accepting, as we must, the unchallenged assertions as fact, it seems on the present record that this result was not reasonably predictable.
Of course, we do not mean to imply that appellant has found an exception to Maxwell’s equations which is going to necessitate their revision. No doubt, if appellant’s affidavits are accurate, those skilled in the art will eventually be able to explain why appellant’s results should have been predictable before appellant undertook his experiments, all based on electromagnetic theory well known long ago. What we do mean is that, based on this sketchy record and with our admittedly limited knowledge of electromagnetic theory, appellant’s results seem to us unexpected at the time his invention was made. Since the Patent Office has not shown why they were to be expected, and since, if unexpected, they are sufficient to establish patentability, we reverse.
The numeral 78 at the left In Fig. 8 Is a duplication made in error; the numeral 78 referred to here is the one at the right. The other one indicates the end of the armature where the moving clipper element would be connected.
The armature A in Wallace’s Fig. 2 parallels the core of the electromagnet, whereas the armature 72 in appellant’s Fig. 8 is perpendicular to the core. Correspondingly, Wallace’s armature extension a overlaps, not an outer arm of the electromagnet’s core, as do appellant’s wraparound extensions 77 and 78, but the pole piece at one end of the core. Additionally, we note appellant’s argument that Wallace, at most, teaches the desirability of minimizing air gaps in electromagnetic circuits energized by interrupted direct current. However, we think that the exact placement of the air gap or gaps in the circuit is of little importance to the issue of the obviousness of minimizing the amount of air in the circuit in order to reduce the operating temperature of the device, and we agree with the solicitor that the conversion of Wallace’s device to ac operation would be obvious to any sophomore electrical engineering student.
In his final rejection, the examiner attacked the relevancy of the experiments described in these affidavits on the ground, inter alia, that they compared vibratory motors with and without the outer two of the three arms of the E-shaped core rather than vibratory motors with all three arms but varying gaps between the armature and the outer core arms. The board, however, accepted the affidavits as proving that reduction in the length of air gaps in the magnetic circuit in appellant’s motor resulted in a cooler motor, but said that it was obvious that this would be the result. Like the board, we accept the affidavits as proving the fact that air gap reduction in appellant’s motor resulted in a cooler motor, but, unlike the board, we are not persuaded that that result was obvious before appellant’s invention.
Presumably, decreasing the air gap, thus decreasing the magnetic potential drop therein, would permit redesign of other components of the device (as, for instance, reducing the number of turns in the exciting coil) while still obtaining the desired wort output.