Winslow Palmer, Appellant, v. John T. Mclamore, Milton J. Minneman, and Ted E. Dunn, Appellees, 220 F.2d 770 (C.C.P.A. 1955)Annotate this Case
Reginald V. Craddock, Great Neck, N. Y. (Henry S. Huff, Great Neck, N. Y., and John W. McLaren, Brooklyn, N. Y., of counsel), for appellant.
Floyd M. Harris, Princeton, N. J. (J. L. Whittaker, Princeton, N. J., of counsel), for appellees.
Before O'CONNELL, Acting Chief Judge, and JOHNSON, WORLEY, and COLE, Judges.
O'CONNELL, Acting Chief Judge.
This is an appeal from a decision of the Board of Patent Interferences of the United States Patent Office granting priority of invention of the subject matter of the interference to the junior party, McLamore, Minneman, and Dunn, hereinafter referred to as McLamore. Three counts, Nos. 1, 5, and 6, corresponding to claims 1, 5, and 6 of the McLamore patent, were before the board in the interference, the interference having been dissolved by the Primary Examiner as to all other counts. Only counts 1 and 5 are involved here, no appeal having been taken as to count 6.
The application of appellant Palmer, the senior party, No. 633,473, was filed December 7, 1945. The McLamore application, No. 745,028, was filed April 30, 1947. The real parties in interest are the Radio Corporation of America, assignee of McLamore; and Sperry Gyroscope Company, Inc., assignee of Palmer. As stated by the board:
"This interference relates to the Loran receivers, particularly to an arrangement utilizing sinusoidal waves as a time measuring standard for enabling the operator thereof to exhibit the separation of received pulses in time units upon a dial or dials."
Count 1, which presents the same question as count 5, relates to radio navigation systems, is illustrative and reads as follows:
"1. In a radio system wherein periodically recurring A pulses and B pulses are received from pairs of ground stations and wherein a deflecting wave of fixed timing and a like deflecting wave of adjustable timing are to be produced for deflecting the cathode ray of a cathode ray tube indicator, said A pulses having the same repetition period being different for each pair of ground stations, means including an oscillator followed by a chain of frequency dividers for producing a square wave having the same repetition period as that of the A and B pulses received from a particular pair of ground stations, the half cycle of said square wave that occurs during the occurrence of a B pulse being identified as the slave period, means for obtaining from said last means a continuous sine wave signal having a fixed phase with respect to said slave period for any of said ground station repetition rates, phase shifter through which said sine wave signal is passed to obtain a phase-shifted wave, means for converting said phase-shifted wave to short-duration timing pulses, means for selecting a desired one of said timing pulses, and means for producing said adjustable deflecting wave in response to the occurrence of said selected pulse whereby said adjustable deflecting wave may be shifted to a desired position along a time axis by selecting a desired timing pulse and by shifting the phase of the selected timing pulse by said phase shifter."
The interference relates specifically to certain improvements in "Loran" navigation radio receivers, a consideration of which is helpful in resolving the issues presented in this case.
As disclosed by the record and briefs, "Loran" is a system of "long range navigation" from which the name was derived and which was developed during World War II. It is now widely used for navigation purposes on ships and planes. Many different methods of determining one's position on a map or chart are used or are conceivable. In sight navigation, for instance, a line of bearing is taken to two or more known landmarks. These lines of bearing are transferred to a chart of the area, and the navigator's position is located at the intersection of the lines. Likewise, if the distances to two known landmarks were known, one might plot a circle from each landmark with the radius of the circle equal to the distance from the landmark. The navigator then is at a point where one circle cuts or touches the other circle.
The "Loran" system is basically one of distance measurement by which two curves are found, the navigator being located at the intersection of the curves. The distance measurement is based upon the fact that radio waves travel at a known speed — the speed of light — and if one could measure the time taken for a radio wave to arrive from a known transmitter, he would know his distance from that transmitter. However, the difficultly arises in trying to find the precise time at which the radio wave started from the transmitter. The difficulties involved in trying to synchronize timing apparatus with such precision is manifest.
The difficulty is solved by having two transmitters, separated at some distance from each other. One station, called the master station, sends one radio pulse, and the other, called the slave station, sends out a pulse which originates a known interval of time after the first. The navigator seeking to find his location measures the difference in time between the arrival of the master pulse and the arrival of the slave pulse. By comparing this difference with the known difference between the pulses at origin, he may find out how much the slave pulse has gained on the master pulse, or how much it has lost. Once he has found this interval, he can by simple mathematics calculate how much closer he is to one station than to the other.
Rather than measuring distance directly, a difference in distance is measured. However, the effect is the same; once the navigator knows how much further he is from one point than another, he can plot a curve, hyperbolic in form, on which every point is an equal distance further from point A than from point B. Then by taking a different pair of stations, the navigator through the same process may obtain another hyperbolic curve, representing the difference in distance he is from the second pair of stations. The navigator will be located at the intersection of these two curves. However, in actual practice, the navigator measures only the time difference between pulses for each pair of stations, and then refers to charts upon which the curves corresponding to the total time differences have been plotted, thus eliminating much calculation at the point of navigation.
It is clear from the foregoing discussion that the essential operation in Loran navigation is the precise measurement of the elapsed time between the reception of the master pulse and the slave pulse. Since radio waves travel at the speed of light, it is equally clear that the time differences will be extremely minute. It is said that these time differences are measured accurately to one microsecond — a millionth part of a second. The subject matter of the present interference relates particularly to means for accurately measuring this time difference, and for allowing the measurement to be read directly from dials without recourse to correction tables or multiple readings.
The system used for this precise measurement is electronic, and involves principally a visual presentation of the pulses upon the screen of a cathode ray tube. The cathode ray is cause to make horizontal sweeps, the frequency of these sweeps being adjustable within certain limits by the operator. The signals from the master and slave shore stations are then impressed upon this cathode ray, causing a vertical displacement, a "blip." It will be seen, that if there are exactly the same number of cathode ray sweeps per second as there are pulses per second, then the pulse will be displayed at exactly the same spot on the cathode ray screen; that is, the "blip" will appear to stand still. On the other hand, if the sweep rate and the pulse rate are not exactly synchronized, the pulses will not appear at the same spot at each sweep, but will appear to "move along" the trace.
It is this fact that makes possible the identification of the particular pair of master and slave stations desired. Each pair operates at a different repetition rate, but upon the same radio wave. For example, one pair of stations will send out a master pulse and a slave pulse every 40,000 microseconds, 25 of each kind of pulse a second. Another pair of stations might send out a pair of pulses every 39,900 microseconds, slightly more than 25 pulses a second. Thus it is clear that if a cathode ray makes a complete horizontal sweep every 40,000 microseconds, the pulses from the pair of stations with that repetition rate will appear at exactly the same spot on the trace on the cathode ray tube. On the other hand, the pulses with a repetition rate of 39,900 microseconds will appear at a slightly different spot on each sweep; that is, the pulses from this second pair of stations will appear to "move along" the trace.
In taking Loran readings, the operator sets the sweep rate of his receiver to equal precisely that of one pair of stations, and then by electronic means to be discussed later he measures the time difference between the master station pulse and the slave station pulse thus identified. Next he changes the sweep rate of the receiver to coincide with that of another pair of stations, and again measures the time difference between pulses. From the information thus obtained, the operator can find on special Loran charts two intersecting lines on both of which he must be located; hence he must be located at the intersection of the lines, and his problem is solved.
The particular problem to which both of the parties to the interference directed themselves is that of providing a means whereby the Loran operator could determine the time difference by reading figures directly from dials mechanically connected to electronic markers to be placed under the pulses from each "blip," where the time difference between the markers was adjustable and known. It appears that in the prior art it was necessary for the operator in effect to measure the distance between the pulses by counting the number of evenly spaced lines appearing on the face of the cathode ray tube between the two pulses. The possibilities of error inherent in this method are manifest.
If the repetition rate for each pair of stations were the same, there would have been no difficulty in providing direct reading apparatus. However, since the rates are different, the prior art methods made a reading for one pair of stations represent a different time interval than that of an identical numerical reading from another pair of stations. In that case it would be necessary to resort to correction tables to obtain the correct absolute time difference.
In so far as we are concerned in this interference, we may limit our discussion to what is known as the "expanded sweep." That is, the cathode ray trace does not represent the entire repetition period, but only the portions of it upon which the two pulses occur. Loran transmissions are so designed that the master pulse occurs during the first half of the repetition period and the slave pulse occurs during the second half of the repetition period. In the expanded sweep, the cathode ray is caused to make one sweep during that portion of the first half of the repetition period in which the master pulse occurs. It is caused to make another sweep in that portion of the second half of the period in which the slave pulse appears. The second sweep is adjustable so that two pulses can be made to coincide upon the face of the cathode ray tube. It is apparent that the time difference between the start of the first sweep and the start of the second sweep will be equal to the time difference between the master and slave pulses.
In the instant interference, appellee McLamore, from whose patent the counts were copied, moved to dissolve the interference on the ground that Palmer could not make the counts. The Primary Examiner denied this motion. The Board of Patent Interferences reversed the examiner, upon original argument and upon rehearing, and Palmer has appealed from that decision. Thus the sole question before us is the sufficiency of the Palmer application to support the counts.
The only portion of the counts over which there is controversy is that which reads as follows in count 1:
"* * * means for obtaining from said last means a continuous sine wave signal having a fixed phase with respect to said slave period for any of said ground station repetition rates * * *."
Both parties illustrate a continuous sine wave signal obtained as required by the count. The only question is whether this sine wave has a "fixed phase" with respect to the slave period, defined as the second half of the repetition period, for any ground station repetition rate.
The Primary Examiner held that this relationship was sufficiently present in Palmer's disclosure for Palmer to make the count, stating:
"This terminology is also considered readable upon Palmer Figure 7, wherein oscillator 201, divider 202, and filter 216 applies a sine wave signal to the phase shifter 219. It will be noted that since oscillator 201 is connected to both square wave generator 211 and the input to phase shifter 219 there will be a fixed phase relationship with respect to the entire repetition period, which includes that portion of the repetition period which the count defines as being the `slave period.' It is considered important to note that the sine wave has a fixed phase with respect to the entire repetition period, and, therefore, with respect to a part thereof defined as the `slave period.'" [Emphasis quoted.]
The Board of Patent Interferences found that Palmer did not in terms disclose the "fixed phase" relation, and that for him to make the count, such relation must be inherent or necessary in his system. The board studied his disclosure, and then concluded
"We are of the opinion that there is insufficient evidence of record to determine with any degree of satisfaction what part if any is played by the sine wave phase relation to the beginning of the slave period in the accuracy or directness of reading of the Palmer device, and therefore there is no sound basis to say that it is a necessary feature of the device. We see nothing in any of the explanations offered by either party from which such a conclusion may be properly drawn, although it must be admitted that the disclosure does not preclude such a possibility as an actuality."
The hypertechnical and complex nature of the subject matter of this interference makes it difficult to define the issues without reference to the schematic drawings and specifications of the parties. It is deemed inexpedient to reproduce these documents here, however, in view of the fact that there are more than 350 elements depicted by McLamore and more than 200 illustrated by Palmer. Insofar as possible we shall endeavor, therefore, to describe the apparatus in broad terms, although we are well aware of the dangers inherent in oversimplification.
As mentioned hereinbefore this interference relates particularly to the problem of producing a pair of "expanded sweep" pulses the recurrence rate of which may be varied to exactly coincide with a multitude of Loran station repetition rates, and to the problem of measuring the time interval between such pulses in units which are independent of a change in repetition rate.
In making comparative measurement which is to have any meaning, not only must the units of measurement be the same, but the end points of the factors to be measured must have the same relation, or be correctible to the same relation, to the measuring stick. Let us take for a specific example the measurement of the length of two pieces of paper. The zero point of the ruler or yardstick is aligned with one end of the first piece of paper and a notation is made of the place on the ruler where the other end touches. In measuring the second piece of paper, the zero point, or some reference point, such as the one inch mark, which is a known distance from the zero point, must be aligned with one end of the second piece of paper, otherwise the two measurements will not be comparable.
Speaking in terms of the interference count, the ruler must be in "fixed phase" with one end of each paper; one of the gradations of the ruler must be aligned with an end of each piece of paper to be measured. Similarly, one might measure from an equivalent spot on each piece of paper, e. g., the mid-points or from any reference point whose distance from the end point is known, e. g., a line on each paper which is known to be 5 inches from the end. In all cases, the accuracy of the measurements depends on how nearly a line on the ruler was aligned ("in phase") with the reference points on each paper.
Inasmuch as the only question before us is whether the counts read upon the Palmer disclosure, we do not need to enter into a detailed discussion of the McLamore patent. We will content ourselves with stating that the party McLamore chose the mid-point of the repetition period as the place he would start all measurements, placing the master pulse at the beginning of the repetition period for all measurements. Then McLamore chose as his "ruler," one in which each of the gradations was exactly half the difference between adjacent Loran repetition periods. Thus, when the receiver is changed from one repetition rate to another, the mid-point is always aligned in "fixed phase" with one of the gradations of the ruler, since as the repetition period is shortened by two units, the mid-point is shifted exactly one unit to the left. More specifically, we are told that Loran repetition periods differ by exact multiples of 100 microseconds. McLamore therefore chose as his measuring unit a sine wave having a repetition period of exactly 50 microseconds. Thus the mid-point of the repetition period, or beginning of the "slave period," always would be in fixed phase with the sine wave for any repetition period.
We note here that the measurements obtained by the McLamore apparatus would not give directly the total time between the reception of the master pulse and the slave pulse for different repetition periods. McLamore does not state exactly how this adjustment for a total time reading is made, but the amount of the adjustment would be a known quantity, and several methods are conceivable for making it. We need not, however, concern ourselves here with that aspect of the problem.
In the Palmer apparatus, a somewhat different solution is adopted. In Palmer the reference point, from which all measurements are started, is not in terms pointed out in the specification, and, due to the peculiar nature of the apparatus, at first blush would appear to be the first expanded sweep pulse. The measuring sine wave is clearly and necessarily in "fixed phase" with the sweep pulse, because that pulse is derived directly from the sine wave.
Thus in the Palmer apparatus certain elements are set into operation at the start of the repetition period. After a given delay the device is ready to generate an expanded sweep. The first sine wave pulse which occurs after the set is in this condition will trigger the expanded sweep. Consequently, by the necessary construction of the Palmer apparatus, the first expanded sweep will be in fixed phase with the measuring sine wave. The same is true with respect to the second expanded sweep, when zero phase shift has been applied to the sine wave. The start of the slave period sets into operation certain elements, which after a given delay are in a condition to generate an expanded sweep. The next sine wave pulse triggers the second expanded sweep apparatus. When no phase shift has been applied to the measuring since wave, it is obvious that the first expanded sweep and the second expanded sweep must be in fixed phase with the sine wave and with each other. However, is the first sweep pulse the real zero point or "time reference point" as contended by appellee? We think not, at least not in the same sense as we have used that term in McLamore, for reasons to be discussed later.
Palmer's figure 7 is the only figure which specifically relates to Loran receivers. We have carefully studied the Palmer specification, and we are convinced that in the device illustrated in figure 7 and described in the specification not only is the measuring sine wave in fixed phase with the first expanded sweep pulse and the second expanded sweep pulse (absent any applied phase shift), but that it is also in fixed phase with the slave period, as required by the counts. As pointed out by the Primary Examiner, the square wave generator which starts the repetition period and the slave period, is triggered by a pulse derived from the same source as the sine wave. To speak in terms of the specification, the square wave generator 211 is triggered in response to a pulse from frequency divider 203, which is coincident with a pulse from frequency divider 202. However, the pulse from frequency divider 202 is the measuring sine wave pulse. Hence, the square wave generator, which produces the "slave period," is triggered by a pulse that is simultaneous with a pulse from the measuring sine wave. From this we logically conclude that the sine wave signal in Palmer's figure 7 would have a fixed phase with respect to the slave period, as required by the counts.
The Board of Patent Interferences dismissed this reasoning by postulating unknown phase shifts in the frequency dividing chain resulting from a change from one Loran station to another. It adopted this postulate in spite of Palmer's statement in the specification that the transient delay circuit 207 "eliminates timing irregularity introduced by the frequency dividing circuits." In addition to these explicit statements by Palmer, it seems to us that a fair reading of the description of figure 7 would lead one to assume that there were no undescribed phase shifts of any significance.
In any event, the word "fixed" is a word of relative meaning. It can have many shades of meaning, from absolutely unchangeable to relatively unchangeable. Thus a scientific law is "fixed" and unchangeable; a fixture in property law is "fixed" though capable of being moved; and a man's habits may be "fixed" though subject to minor deviations. It has been said that words in an interference count are to be given the broadest interpretation possible. However, we do not need to avail ourselves of this principle here. We may look to appellee's disclosure and give to the word "fixed" the same meaning it must have there. We find that "fixed" there is subject to any phase shifts that may occur in the frequency dividing chain, which we assume are minor. Thus the word "fixed" as applied to the McLamore apparatus can only mean fixed within the limits of any variation introduced by the frequency dividing chain. Looking to Palmer's figure 7, we see that the sine wave is clearly in fixed phase with the slave period, in that sense of the word "fixed," since the slave period is derived from the sine wave by an almost identical series of frequency dividers. We conclude that the counts read in terms upon Palmer's figure 7. The main contention of appellee McLamore is, however, that even if the count can be read in terms on the Palmer specification, such disclosure is merely incidental, and that incidental disclosure is insufficient to support the counts. In support of this last conclusion McLamore cites Gray Telephone Pay Station Co. v. Baird Mfg. Co., 7 Cir., 174 F. 417, and Brill v. Third Ave. R. Co., C.C., 103 F. 289, as holding that claims are not anticipated merely because they read in terms on prior art, and cites Minton v. Thomas, 48 F.2d 425, 18 C.C.P.A., Patents, 1153, as indicating that the same rule should apply to interference cases.
In our view of the case, it is unnecessary to determine whether appellee's statement of the law is correct, for it is our opinion that there is sufficient disclosure in the Palmer application for us to conclude that he is entitled to the invention defined by the counts. We have carefully studied the Palmer specification, and are convinced that his inventive concept basically included the particular embodiment defined by the counts, whether or not he fully recognized the significance of the particular relationship in controversy here at the time he filed his application.
We are struck by the basic similarity, after all the technical verbiage is taken away, between the operation of the McLamore device and the Palmer device. If Palmer's figure 7 were constructed without selectors 212 and 214, then the positive-going wave from the square wave generator would trigger the first expanded sweep, and we would have essentially the same apparatus as that of McLamore. In such a case, the fixed index marker, or time reference pulse, whichever it may be called, would be generated by the square wave in both. The delayed pulse or variable index marker timing circuit would be started by the second half of the square wave in both. Delay is introduced by somewhat different means in both, but the particular means are not essential. Both, however, use a pulse derived from a phase-shifted sine wave to finally generate the expanded sweep wave occurring in the slave period.
For purposes of exposition we have considered Palmer's device without selectors 212 and 214. We must consider what effect these selectors have upon the apparatus. We are convinced that by use of these selectors Palmer has merely increased the precision of his measurement. Palmer clearly states that this is one of the objects of his invention. In the McLamore apparatus the variable index marker is derived from the 20,000 cycle sine wave, and thus has the precision of that wave. The fixed index marker is derived from a 50 cycle pulse, and has only the accuracy of that pulse. Thus the measurement obtained from McLamore will only have the accuracy or precision associated with the lower frequency pulse. In Palmer, however, the lower frequency waves are used merely to set up the conditions for choosing the higher frequency pulse, and consequently it has the precision of the higher frequency wave. In Palmer, any phase shift which occurs in the frequency dividing chain is automatically compensated for by a counterbalancing shift in the first sweep pulse. Palmer's primary purpose, however, was the production of a direct reading apparatus and this he would have accomplished without the increased precision introduced by the utilization of the higher frequency sine wave to select both sweep circuits.
It is the peculiar character of this additional advantage secured by the Palmer apparatus which enables appellee to argue that the fixed phase relationship is not an essential feature of appellant's structure. McLamore suggests that if the high frequency sine wave of the count be shifted continuously outside of the frequency divider chain, then the sine wave would continually vary in phase with respect to the slave period, but that the accuracy of Palmer's measurement would not be affected. This seems to be clearly a modification that has no basis in either disclosure to justify it. We are convinced that in actual practice of the two inventions there would be no greater phase shift with respect to the slave period in Palmer than there would be in McLamore; the only difference is, that in Palmer any such phase shift would not affect the validity of the measurement. We do not feel that Palmer's skill in refining his invention should handicap him in an interference. Furthermore, it is not evident that McLamore's criticism would necessarily be true in all embodiments of the Palmer device, e. g., if the phase shifters were not mechanically connected so that the measuring sine wave could be shifted relative to the other waves.
It is also pertinent to note here that McLamore states that the time measured by Palmer is between the first expanded sweep pulse and the second expanded sweep pulse, and therefore that Palmer's time reference point is the first sweep pulse, while the McLamore time measurement is from the start of the slave period. We are of the opinion that this discussion and McLamore sheets 1 and 2 which seek to interpret the Palmer device are erroneous and misleading. These sheets and graphs seem to indicate that the time loss in shifting from one repetition rate to another occurs in Palmer between the beginning of each half of the square wave and the associated sweep pulse. The graphs show that the wave which "turns off" selectors 212 and 228 "loses counts" — that is, that the length of the wave out of 212 and 228 varies according to the number of counts lost. This is incorrect, for although the wave from 204 does lose counts, it does not lose them during the time selectors 212 and 228 are active, due to the action of transient delay circuit 207. The time loss occurs between the sweep pulse and the end of the associated half-period. To state the same thing in other language, the two sweep pulses occur a constant time period after the start of their respective half-periods (absent applied phase shift). Thus, the Palmer apparatus does not directly measure the total time interval between the two pulses for different repetition periods. At zero phase shift there will not be a constant time interval between the two pulses for every repetition period, but the time will vary by one half the difference of the repetition periods. This is the same characteristic we have noted in the McLamore apparatus.
We also note that Palmer's timing measurement is unchangeably linked, not to the first sweep pulse, but to the start of the slave period. It is the start of the slave period which sets Palmer's timing apparatus into operation, and Palmer gives clear indication that at zero phase shift there is a constant time between the start of the slave period and the selection of the pulse. The Palmer apparatus cannot measure any and all times between the first and second pulses. When zero phase shift is applied, the second sweep is at a spot determined by the start of the slave period.
We are of the opinion that the Palmer application contains sufficient disclosure to support the portion of the counts relating to a fixed phase relationship between the sine wave and the slave period, and since no other limitations are in controversy, it follows that the decision of the Board of Patent Interferences must be and hereby is reversed.
GARRETT, Chief Judge, did not participate in the hearing or decision of this case because of illness.
WORLEY, J., dissents.