Opinion
Patent Appeal No. 6921.
June 20, 1963.
Fidler, Beardsley Bradley, Chicago, Ill. (Raymond E. Fidler and Charles W. Bradley, Jr., Chicago, Ill., of counsel), for appellant.
John V. Regan, Princeton, N.J. (A. Russinoff, Princeton, N.J., of counsel), for appellee.
Before WORLEY, Chief Judge, and RICH, MARTIN, SMITH and ALMOND, Jr., Judges.
This appeal is from the decision of the Board of Patent Interferences which awarded priority of invention in Interference No. 89,337 to the junior party Lo, patentee named in U.S. Patent 2,818,555. Appellants Chedaker et al. are senior by virtue of the earlier filing date of their application in which claims 8 and 12 of the Lo patent, corresponding to the counts in issue, were copied and added by amendment. The board found no support in the Chedaker et al. application for the copied claims.
Issued December 31, 1957 on application Serial No. 524,664, filed July 27, 1955.
Serial No. 248,716, filed September 28, 1951.
The counts relate to a magnetic device and associated windings useful in the storage and control of electrical signals representing information. The magnetic material employed has the characteristic of a "rectangular hysteresis loop." This term is derived from the graphical presentation obtained when magnetizing force is plotted against the magnetic flux density produced in the material. The resulting curve, called the hysteresis curve or loop, takes on a roughly rectangular shape.
In practical effect, such a hysteresis curve or loop indicates that the material has high magnetic remanence or retentivity. This means that if the material is magnetized in one direction by an applied magnetizing force, which may be created by current flow through a winding associated therewith, it will remain magnetized in that direction even after the magnetizing force is removed, i.e., it "remembers" the flux direction. Similarly, when a magnetizing force is applied in the opposite direction and then removed, the material will remain magnetized in the opposite direction. There is an abrupt transition between flux in one magnetic state and flux in the opposite state. In each state, the remanent flux is nearly equal to the saturation flux of material characterized as having a square hysteresis loop.
Since there are two stable "states," or "senses," depending on the direction of the applied magnetic force, with an abrupt change between these states when the direction of the magnetizing force is reversed, the material is capable of "remembering" which of the two forces was last applied.
There are two counts in issue. Count 1 is illustrative and reads:
"1. A magnetic device comprising a unitary core of magnetic material characterized by having a substantially rectangular hysteresis loop, said core having a plurality of apertures therein, means for producing a magnetic flux completely around said core in one sense, and a winding wound through a first and a second of said apertures exclusively so that, when energized by a pulse of either polarity, the magnetizing force generated thereby will produce a flux change from said one sense to the sense opposite the one sense in at least a portion of said core adjacent a third of said apertures."
The language of the counts gives rise to two issues of interpretation. The first is whether the phrase "characterized by having a substantially rectangular hysteresis loop" modifies the word "core" or the word "material." The Chedaker et al. device uses material having such a characteristic but not a core so characterized over-all.
The second issue is whether the Chedaker et al. disclosure shows a winding in which, "when energized by a pulse of either polarity, the magnetizing force generated thereby will produce a flux change from said one sense to the sense opposite the one sense * * *." Appellant contends that there is such a disclosure; appellee, the contrary.
As to the first issue, the record shows that each of the parties employs magnetic material which is characterized as having a rectangular hysteresis loop. When a core is made from such material, however, it may or may not have a rectangular hysteresis loop, depending on the geometry or configuration of the core.
The core disclosed in appellee's patent is a disc of magnetic material having a number of apertures therein. Each aperture has a function in the operation of the device. In the center of the disc is the blocking aperture, through which a winding is passed. When a uni-directional current pulse of sufficient amplitude is applied to this winding, the entire core is saturated with flux in one direction, placing it in what appellee describes as a blocked condition.
A pair of smaller setting apertures are symmetrically located in the disc on either side of the larger diameter blocking aperture. A winding extends through those apertures from a setting-pulse source. The windings are so arranged that a pulse of either a positive or negative polarity from the source unblocks the core by changing the flux direction in a portion of the core.
A third essential aperture in appellee's device is designated an output aperture. Passing through this aperture is an input winding adapted to receive a signal pulse and a separate output winding which responds to the signal pulse by producing an output which is indicative of whether the core is blocked or has been unblocked by reason of a setting pulse which was applied to the winding passing through the setting apertures.
While more apertures may be present, the minimum number disclosed by the Lo patent is four, to wit: a blocking aperture, two setting apertures and an output aperture. A feature of practical importance in appellee's device is that a setting pulse of either polarity in the winding through the setting apertures produces a reversal of the flux direction in the path between the output aperture and the blocking aperture. While a positive setting pulse reverses a different portion of the blocked core than that reversed by a negative setting pulse, the reversed portions overlap so that each pulse reverses the portion of the core between the output aperture and the blocking aperture. Since the flow of flux around the output aperture is changed, the induced voltage in the output winding indicates either a blocked or an unblocked response, whether the unblocked response is a result of a positive or a negative setting pulse.
The core disclosed in the Lo patent is so constructed that the entire device exhibits the characteristic of a rectangular hysteresis loop so that the blocked condition may control all parts of the core. With some geometric configurations, the flux density might not be saturated in all portions of the core, so that some portions might indicate no strong flux direction (i.e., these portions might "forget" the last applied magnetic force), even though the composition of the core is a rectangular hysteresis loop material.
In appellants' device, as disclosed in the Chedaker et al. application, it is not desired to have each portion of the core saturated with flux. Only one portion of the multi-apertured core in Chedaker et al. must be maintained at high flux density. In the remainder of the core, air gaps are introduced in the paths where high remanence is not required to concentrate the remanence flux in the critical paths. Alternatively, Chedaker et al. state in the specification that "it may be advantageous to make only the core section DGHF [the section providing the memory function] from a high remanence, high coercivity material and make the remainder of the core from a high permeability, low coercivity material."
Appellants argue that the language of the counts merely requires that the starting material exhibit the rectangular hysteresis loop characteristic. It is a basic rule of grammar, they contend, that limiting clauses modify the closest preceding antecedent, here "material," and if "core" were intended to be modified, appropriate language could have been used. Moreover, appellants point out that the normal meaning in the field is that the term "rectangular hysteresis loop" generally is used to modify "material," since only toroids form rectangular loop cores. Bar magnets, for example, exhibit a non-rectangular hysteresis loop even when made of rectangular hysteresis loop material. In addition, the loop characteristics vary at different points of the core in many configurations, according to appellants. A number of exhibits were introduced into the record by appellants in an attempt to show that "material" is the word modified by "rectangular hysteresis loop" in the count. They conclude:
"In light of this evidence, it will be clear that magnetic material is customarily described in terms of its hysteresis loop, and magnetic cores are customarily described in terms of the hysteresis loop of the material they contain, by the manufacturers of magnetic cores in the description of their cores, and by those ordering cores including counsel for party Lo, by recognized standard texts, by an article coauthored by party Lo, by the Bureau of Standards standard handbook on Precision Measurements, and by numerous patentees in the description of magnetic core devices. Also, patents have been granted to party Lo and others on methods and apparatus for measuring the intrinsic B-H hysteresis loop of magnetic material. B-H hysteresis loops are commonly employed by physicists in determining internal stresses, hysteresis loss, coercive force and remanent flux density, and by engineers in designing magnetic circuits."
Appellee, on the other hand, argues that a core is always impliedly referred to because "rectangular hysteresis loop" is an intrinsic property which cannot be divorced from core geometry. He states it thus:
"This record makes it plain that when persons in the art speak of a material as having a rectangular hysteresis loop, they understand this is an intrinsic, theoretical property (B-H) which is deduced from a closed core and which is realized in a practical core only when the core provides continuous flux paths entirely in the material."
The Board of Patent Interferences found that the word "core" rather than the word "material" was modified in the counts since the term "rectangular hysteresis loop" is merely "a theoretical matter" in relation to material. The board stated:
"* * * An invention is practical rather than theoretical, and definitions of an invention should be given practical interpretation rather than theoretical where there is doubt."
It appears that the board found the language of the counts ambiguous and chose the practical interpretation rather than the theoretical interpretation. Appellants assert that this was error; that if ambiguity is found, it is to be resolved by reference to the specification of the Lo patent where the counts originated; that the Lo patent makes mention several times of rectangular hysteresis loop materials, but "There is not a single reference anywhere in the Lo specification to a rectangular hysteresis loop core."
We are inclined to agree with appellants on this branch of the case. It is settled law that ambiguity should be resolved by reference to the specification where the count originated. Neumair v. Malocsay, 77 F.2d 622, 22 CCPA 1349; Peterson v. Schatzel, 129 F.2d 556, 29 CCPA 1124. We find nothing in the Lo patent specification to indicate that the core made from the rectangular hysteresis loop material must exhibit such a characteristic in every portion of the core. It is critical that some portions of the core be so characterized, but not each portion or the core as a whole. There is no limitation in the counts that the core be shaped as a disc, and presumably any protrusion from a core would be likely to have a reduced flux density and a more linear loop characteristic. It appears to us that the portion of appellants' core that provides the memory function would necessarily exhibit a rectangular hysteresis loop even though air gaps are introduced in other portions of the core where such a characteristic is not desired or required.
We agree with the board that "rectangular hysteresis loop" might be termed a theoretical determination which must be tied to the practical device. The loop characteristic is a property of the material, but it is dependent on the particular shape in which the material is used. We are of the opinion that the word "material" is the one modified in the counts, but that, by necessary implication, the critical portions of the core must exhibit such a characteristic. In other words, if appellants showed no paths or portions of the core exhibiting such a characteristic, they could not support the counts. Such is not the case here, however. The experts representing each side testified that at least part of the core disclosed in appellants' specification (the path providing the memory function) exhibits a "substantially rectangular hysteresis loop." After considering all of the voluminous evidence of record, we are of the opinion that appellants can support the language "a unitary core of magnetic material characterized by having a substantially rectangular hysteresis loop" of the counts.
Finding that appellants can support the first disputed portion of the counts is not conclusive of the case, however. Appellants must also show support for the second disputed clause concerning the switching of the magnetic flux from "one sense to the sense opposite the one sense" as recited in the counts.
The consideration of this branch of the case requires a more detailed analysis of the devices disclosed by the respective parties.
The Chedaker et al. core is similar to that disclosed in the Lo patent in that it is made from material characterized by a rectangular hysteresis loop; it has plural apertures and associated windings; it can be used for "remembering" two responses (e.g., yes or no); an "inquisitor" pulse is used to "ask" the device which response it "remembers"; and the "answer" is determined by the voltages induced in output windings through certain apertures.
In the operation of appellants' device, the information is impressed on a portion of the core by one of two input windings through a primary aperture. One winding is used for a positive pulse and the other for a negative pulse. These pulses produce magnetic flux in opposite directions around the primary aperture in the core. To determine which pulse was last applied, an inquisitor pulse is sent through a winding through two secondary apertures. The inquisitor pulse induces flux around the two secondary apertures each of which has a common flux path with the primary aperture. The flux around one of the secondary apertures is in the same direction as that around the primary aperture and the other is in the opposite direction because the inquisitor winding is between the secondary apertures. Therefore, the flux induced by the inquisitor flows in the same direction as the primary flux in one leg and flows counter to and reverses the primary flux in the other leg, depending on the direction of the primary flux. Output windings are around each of the two common legs, and different voltages are induced in each one depending on whether the primary flux is in the same or the opposite direction. These two output voltages are compared to determine which was the last information pulse applied, i.e., which input is "remembered." In practical effect, since the primary flux is near saturation, the additive effect of the inquisitor flux in the one leg is not great, and the reversing effect in the other leg is only temporary. The reversal of the flux exists only while the inquisitor pulse is on, after which the flux direction returns to that of the last applied force because of the strength of the residual magnetism of the "remembered" information. The inquisitor pulse produces no permanent reversal of the primary flux of the core. Appellants' specification states:
"The repetition rate of the inquisitor pulses of the above described embodiment wherein the various components have the values as stated above may be as high as 50,000 pulses per second."
Appellants in their brief concede that the application is in error in stating that the inquisitor pulse reverses all of the flux in the leg where it counters the direction of the remanent flux. Nevertheless, they argue, some of the flux is reversed in order to indicate the flux change reflected in the different induced voltages in the output windings. The counts, it is asserted, merely call for flux reversal in a portion of the core, which is shown by appellants' device. Moreover, increasing the strength of the inquisitor pulse would result in flux reversal in a greater portion of the core, according to appellants.
The testimony of the experts is pertinent here. Appellants take the position that permission to take testimony on the second disputed clause of the counts was neither sought nor granted, that there was timely objection by appellants, and that, as a result, the testimony concerning this limitation is incompetent. Appellants rely on Creed et al. v. Potts, 96 F.2d 317, 25 CCPA 1084, and Weintraub v. Hewitt et al., 38 App.D.C. 82, 1912 C.D. 466, as support for the proposition that appellee could not take testimony concerning the flux reversal of the respective devices.
We are of the opinion that the testimony on this point may properly be considered. Appellee's motion to dissolve on the ground that Chedaker et al. have no right to make the counts specifically questioned whether there was support for the recitation relative to flux reversal as well as that regarding the rectangular hysteresis loop. Thus appellants had notice that Lo might be expected to argue both points at final hearing before the board in support of his motion.
The board placed no express limitations on the testimony to be taken, but said "testimony taken pursuant to the motion may be pertinent to matters which may be raised at final hearing." The two disputed phrases of the counts are closely intertwined in that flux reversal is necessarily related to a rectangular hysteresis loop characteristic. Under the facts of this case, we are of the opinion that testimony concerning both points is pertinent to the right to make the counts and could properly be taken by the permission granted by the board.
We do not find Creed et al. v. Potts, supra, or Weintraub v. Hewitt et al., supra, controlling here. In the former case, the court found an expressed limitation on the testimony to be taken from the language of the board in granting permission "in respect to the features specified in the motion and affidavits." We find no such limitation here on the testimony to be taken.
In the Weintraub case, the excluded testimony related to the matter of joint inventorship, a completely different issue from the right to make question for which permission was granted to take testimony. Here the testimony taken concerns the same issue for which leave was granted.
The testimony of the experts on behalf of appellee tends to show that the energizing pulse, called the "inquisitor pulse" by appellants, is incapable of producing a reversal of flux in the entire portion of the core where the flux direction resulting from the inquisitor pulse runs counter to the remanent flux of the "storage" portion of the core. Indeed, it appears that the air gaps were introduced into appellants' core for the purpose of reducing the effect of the inquisitor pulse on the remanent flux of the core. Only enough flux change is desired to indicate a voltage differential in the output windings to show the direction of the remanent flux in the output windings. Appellee's expert Dr. Briggs estimated that only 5 percent of the flux in the leg in question was reversed. Appellants do not challenge this estimate but refer to it in their brief. Appellee characterizes this flux change as "only a transient one, a second order disturbance." It appears to us that this characterization is correct, inasmuch as appellants' specification indicates that as many as 50,000 inquisitor pulses per second may be applied to indicate the direction of the remanent flux of the core.
In contrast, appellee's core requires a complete reversal of the flux in at least a portion of the core as an essential to its operation. The Lo device would be useless for its intended purpose if the setting pulse did not reverse a portion of the core to the direction opposite to the direction of the blocking pulse. If only a temporary disturbance or a 5 percent flux change resulted from the setting pulse in appellee's device, the device would only indicate the blocked response when energized, unless the setting pulse happened to be on at the time of inquiry. It is clear to us that the setting pulse in appellee's device must "produce a flux change from said one sense to the sense opposite the one sense" (emphasis added) of greater extent and of longer duration than that shown by appellants. It seems likely that the change produced in appellants' device is more like the change produced around the output aperture of appellee's device rather than that produced around the setting apertures where complete reversal is required. In appellee's specification, it is stated that "some small flux change is produced in the path about the output aperture 18 by the input pulses," which induces "unwanted voltage [that] may be considered a `noise' voltage and may be readily discriminated against in most applications." In appellants' device, it is the input (inquisitor) pulses which produce the 5 percent flux change which, in turn, induces a voltage in the output winding which, by comparison with the other output winding indicates remanent flux direction.
We agree with appellee's analysis of the testimony of experts Briggs and Tracy:
"* * * Both of these experts were careful to explain that, at most, there is a reversal of `magnetic moments' or movement of `magnetic dipoles' in microscopic areas of the core * * *. But the reversal of `magnetic moments' is, of course, only a random thing, scattered in different isolated regions, so that there is no identifiable portion of the leg DEF in which one can say that the flux changes from one sense to the sense opposite." (Appellee's emphasis.)
We conclude that appellants have not shown support for the limitation in each of the counts of "a winding wound through a first and a second of said apertures exclusively so that, when energized by a pulse of either polarity, the magnetizing force generated thereby will produce a flux change from said one sense to the sense opposite the one sense in at least a portion of said core adjacent a third of said apertures."
Accordingly, the decision of the board is affirmed.
Affirmed.