`
`
`In re Patent of: Hays et al.
`U.S. Patent No.: 5,659,891
`Issue Date:
`August 19, 1997
`Appl. Serial No.: 08/480,718
`Filing Date:
`June 7, 1995
`Title:
`
`Multicarrier Techniques in Bandlimited Channels
`IPR:
`
`IPR2016-00768
`
`
`
`
`DECLARATION OF DR. JAY P. KESAN
`
`1.
`
` My name is Dr. Jay P. Kesan. I understand that I am submitting a
`
`declaration for Mobile Telecommunications Technologies LLC (MTel”),
`
`offering technical opinions in connection with the above-referenced Inter
`
`Partes Review (IPR) proceeding pending in the United States Patent and
`
`Trademark Office for U.S. Patent No. 5,659,891 (the “’891 patent”), and
`
`prior art references relating to its subject matter. My current curriculum
`
`vitae is attached as Appendix A.
`
`2.
`
`
`
`I also provide selected background information here relevant to
`
`myself, my experience, and this proceeding.
`
`3.
`
`
`
`I am a Professor at the University of Illinois at Urbana-Champaign,
`
`where I am appointed in the College of Law, the Department of Electrical
`
`and Computer Engineering, the Coordinated Science Laboratory, and the
`
`Information Trust Institute. I have a Ph.D. in Electrical and Computer
`
`
`
`MTel., Exhibit 2003, Juniper v. MTel., Page 1, IPR2017-00640
`
`
`
`Engineering from the University of Texas at Austin and a J.D., summa
`
`cum laude from Georgetown University. I have also worked as a
`
`research scientist at the IBM T.J. Watson Research Center, and I am a
`
`named inventor on several United States patents. I have also served as a
`
`technical expert and legal expert in patent infringement lawsuits. I have
`
`been appointed to serve as a Special Master in patent disputes.
`
`Additionally, I have been appointed as a Thomas Edison Scholar at the
`
`United States Patent and Trademark Office (“USPTO”).
`
`4.
`
` My opinions in this report are based on my experience and expertise
`
`in the field relevant to the ’891 patent. To prepare this Report, I have
`
`reviewed and considered materials shown in Appendix B and referred to
`
`herein, principally including the ’891 patent and its file history, the
`
`Petrovic reference, and the extrinsic evidence cited.
`
`5.
`
`
`
`I anticipate using some of the above-referenced documents and
`
`information, or other information and material that may be produced
`
`during the course of this proceeding (such as by deposition testimony), as
`
`well as representative charts, graphs, schematics and diagrams,
`
`animations, and models that will be based on those documents,
`
`information, and material, to support and to explain my testimony before
`
`the Board regarding the validity of the ’891 patent.
`
`2
`
`MTel., Exhibit 2003, Juniper v. MTel., Page 2, IPR2017-00640
`
`
`
`6.
`
` This report is based on information currently available to me. To the
`
`extent that additional information becomes available (whether from
`
`documents that may be produced, from testimony that may be given or in
`
`depositions yet to be taken, or from any other source), I reserve the right
`
`to continue the investigation and study. I may thus expand or modify my
`
`opinions as that investigation and study continues. I may also
`
`supplement my opinions in response to such additional information that
`
`becomes available to me, any matters raised by and/or opinions provided
`
`by MTel’s experts, or in light of any relevant orders from the Board.
`
`7.
`
` Throughout this report, I cite to certain documents or testimony that
`
`support my opinions, including appendices C-K. These citations are not
`
`intended to be and are not exhaustive examples. Citation to documents
`
`or testimony is not intended to signify and does not signify that my
`
`expert opinions are limited by or based solely on the cited sources.
`
`8.
`
`
`
`I am an attorney, registered to practice before the United States Patent
`
`and Trademark Office, and a legal expert in United States Patent Law.
`
`9.
`
` A person of ordinary skill in the art at the time of the invention
`
`(PHOSITA) of the ’891 Patent would possess a bachelor’s degree in
`
`electrical engineering or its equivalent and about four years working in
`
`the field of wireless telecommunications networks and would possess
`
`3
`
`MTel., Exhibit 2003, Juniper v. MTel., Page 3, IPR2017-00640
`
`
`
`knowledge regarding frequency, amplitude, and masks as used in
`
`telecommunications, or equivalent education and work experience.
`
`10.
`
`I have considered Dr. Kakaes opinion regarding the level of skill of a
`
`PHOSITA (Ex. 1003 at ¶ 10), and my opinions expressed herein would
`
`not change, even under his definition.
`
`11. The ’891 Patent is directed to the field of telecommunications and to
`
`systems and methods for operating paging carriers.
`
`12. A brief background on carriers is helpful in understanding how
`
`the ’891 Patent is operating carriers.
`
`13. Most simply, in telecommunications an unmodulated carrier is, in
`
`general, a sinusoidal waveform. Drawing 1 below illustrates a carrier
`
`with a frequency of 1 Hz and an amplitude Ac.
`
`14. Drawing 1 depicts an ideal carrier in the time domain. However, in
`
`telecommunications, it is frequently useful to view carriers in the
`
`
`
`4
`
`MTel., Exhibit 2003, Juniper v. MTel., Page 4, IPR2017-00640
`
`
`
`frequency domain. In the frequency domain, the ideal carrier of Drawing
`
`1 has just a single frequency of 1 Hz. Drawing 2 below illustrates the
`
`carrier of Drawing 1 as shown in the frequency domain.
`
`15.
`
`In Drawing 2, the carrier is shown as an impulse with a single
`
`frequency. This is because the sinusoidal waveform of Drawing 1 is
`
`
`
`ideal.
`
`16. In the real word, it is not possible to transmit an ideal sinusoidal
`
`waveform even for an unmodulated carrier. Additional unwanted
`
`frequencies are generated. As a result, even in the frequency domain, a
`
`carrier has more than one frequency.
`
`17.
`
`In Drawing 2, the y-axis is not specified. In telecommunications, the
`
`frequencies of a carrier are often plotted in relation to their peaks
`
`intensities or their power levels. These types of plots can be referred to
`
`5
`
`MTel., Exhibit 2003, Juniper v. MTel., Page 5, IPR2017-00640
`
`
`
`as emission spectra. Drawing 3 below illustrates an emission spectrum
`
`for a real world carrier.
`
`
`
`18.
`
`In Drawing 3, the carrier’s attenuated power levels are plotted versus
`
`frequency. Drawing 3 shows that a real unmodulated carrier in an
`
`emission spectrum has a shape that is dependent on frequency.
`
`19. Radio frequency carriers are regulated in the United States by the
`
`FCC. The FCC specifies frequency channels or ranges for carriers and
`
`specifies a specific use for each channel. The FCC also specifies the
`
`maximum power levels at a given frequency for the carriers in each
`
`channel.
`
`20.
`
`In col. 1, ln. 12-14, the ’891 Patent describes that, at that time, the
`
`number of channels allocated by the FCC for mobile page use was
`
`limited. However, at that time, the demand for those channels was
`
`increasing rapidly. Ex. 1001 at 1:11-18. As a result, the ’891 Patent is
`
`6
`
`MTel., Exhibit 2003, Juniper v. MTel., Page 6, IPR2017-00640
`
`
`
`directed to the problem of the limited channels allocated for mobile
`
`paging at the time of the ’891 Patent.
`
`21. The ’891 Patent discusses two known solutions to this problem. The
`
`first solution is to increase the number of messages transmitted in a
`
`channel in a given period of time. Id. at 1:25-27. This can be stated most
`
`simply as increasing the message rate of the channel. The second
`
`solution is to increase the number of messages sent at the same time by
`
`placing multiple carriers in the same channel. Id. at 1:37-46. This can be
`
`stated most simply as increasing the message capacity of the channel.
`
`22. The ’891 Patent describes that a known method of increasing the
`
`message capacity of a channel is to use more than one carrier in the
`
`channel. Id.
`
`23. The ’891 Patent explains, however, that placing more than one carrier
`
`in the same channel has traditionally required “stringent protection levels
`
`between subchannels” of the multiple carriers. Id. at 2:1-6. The stringent
`
`protection levels described by the ’891 Patent are the additional
`
`limitations the FCC places on each type of channel. Id. at 1:57-67.
`
`These additional limitations are referred to as an emission mask for the
`
`channel.
`
`7
`
`MTel., Exhibit 2003, Juniper v. MTel., Page 7, IPR2017-00640
`
`
`
`24.
`
`47 C.F.R. §22.99 of the current FCC regulations defines an emission
`
`mask as “[t]he design limits imposed, as a condition or certification, on
`
`the mean power of emissions as a function of frequency both within the
`
`authorized bandwidth and in the adjacent spectrum.”
`
`25. The ’891 Patent describes that FCC emission masks are directed to the
`
`near-far interference problem that occurs between carriers. Ex. 1001 at
`
`4:12-13.
`
`26. The near-far interference problem can be illustrated by considering
`
`two adjacent carriers that are very close in frequency. Drawing 4 below
`
`illustrates two adjacent carriers that are very close in frequency.
`
`
`
`27.
`
`In Drawing 4, Carrier 1 has a center frequency of 1 Hz and Carrier 2
`
`has a center frequency of 1.4 Hz. Near-far interference occurs when a
`
`receiver is much closer (near) to the transmission source of, for example,
`
`Carrier 1 and much farther (far) from the transmission source of Carrier 2.
`
`8
`
`MTel., Exhibit 2003, Juniper v. MTel., Page 8, IPR2017-00640
`
`
`
`Drawing 5 below illustrates the near-far interference experienced by a
`
`receiver that receives Carrier 1 and Carrier 2.
`
`
`
`28.
`
`In Drawing 5, the Receiver receives Carrier 1 from Transmitter 1,
`
`which is close. The Receiver receives Carrier 2 from Transmitter 2,
`
`which is far. Due to the inverse square law of electromagnetic power
`
`transmission, the Receiver receives much more power from Transmitter 1
`
`than from Transmitter 2. See https://en.wikipedia.org/wiki/Near-
`
`far_problem as of June 13, 2016. As a result, at the Receiver, Carrier 2
`
`cannot be distinguished from a portion of Carrier 1.
`
`29. The ’891 Patent explains that the near-far problem of Drawing 5, for
`
`example, can be eliminated by placing emission mask limitations on the
`
`two carriers. Drawing 6 below illustrates how placing emission mask
`
`9
`
`MTel., Exhibit 2003, Juniper v. MTel., Page 9, IPR2017-00640
`
`
`
`limitations on the two carriers eliminates the near-far problem for the two
`
`carriers of Drawing 4.
`
`
`
`30.
`
`In Drawing 6, Mask 1 requires that Carrier 1 is narrowed so that
`
`additional frequencies are not generated at a level that will interfere with
`
`Carrier 2. Similarly, Carrier 2 is narrowed so that additional frequencies
`
`are not generated at a level that will interfere with Carrier 1. Drawing 7
`
`below illustrates how these masks can eliminate the near-far interference
`
`problem.
`
`
`
`10
`
`MTel., Exhibit 2003, Juniper v. MTel., Page 10, IPR2017-00640
`
`
`
`31.
`
`In Drawing 7, the Receiver again receives Carrier 1 from Transmitter
`
`1, which is close. And, the Receiver receives Carrier 2 from Transmitter
`
`2, which is far. Carrier 1 and Carrier 2 are also still received at the
`
`Receiver at different power levels due to the inverse square law.
`
`However, Carrier 1 no longer interferes with Carrier 2, because Carrier 1
`
`was required to attenuate its signal more at the frequencies that would
`
`interfere with Carrier 2.
`
`32. Although the “stringent protection levels” afforded by the subchannel
`
`masks of Drawing 6 eliminate the near-far problem as shown in Drawing
`
`7, the ’891 Patent teaches away from this solution by listing its
`
`drawbacks. Ex. 1001 at 2:1-12. The chief drawback listed is a
`
`symmetric condition required by this approach. This symmetric
`
`condition is that “[t]he carriers are symmetrically located within the
`
`channel such that they are evenly spaced relative to each other and to the
`
`band edges of the primary mask defining the primary channel.” Id. at
`
`2:6-9. Drawing 8 below illustrates the symmetric condition.
`
`11
`
`MTel., Exhibit 2003, Juniper v. MTel., Page 11, IPR2017-00640
`
`
`
`
`
`33.
`
`In Drawing 8, distance Dc defines the center point between Carrier 1
`
`and Carrier 2, or more precisely, half the distance between the center
`
`frequencies of adjacent carriers. Distance Dm is the spacing between
`
`Carrier 1 and the band edge of the Primary Mask of the primary channel.
`
`The symmetric condition of the ’891 Patent, therefore, occurs when Dc =
`
`Dm.
`
`34. The ’891 Patent teaches away from the symmetric condition by
`
`describing that it “often necessitates the need for sophisticated receiver
`
`and transmitter schemes.” Ex. 1001 at 2:11-12.
`
`35.
`
`Instead of using subchannels that require the symmetric condition,
`
`the ’891 Patent describes and claims transmitting multiple carriers from
`
`the same location that are in the same channel in order to increase the
`
`message capacity of the channel. Id. at 3:44-46.
`
`12
`
`MTel., Exhibit 2003, Juniper v. MTel., Page 12, IPR2017-00640
`
`
`
`36. The co-location of the transmission by the invention of the ’891
`
`Patent is shown in Figures 1 and 2. Figure 1 is reproduced below.
`
`
`
`37. Figure 1 of the ’891 Patent shows that two carriers are transmitted
`
`from the same location by one antenna. From Figure 1 and the stated
`
`purpose of the ’891 Patent, a PHOSITA would also conclude that the two
`
`carriers are transmitted at the same time. For example, as described
`
`above, the purpose of the ’891 is to increase the message capacity of the
`
`channel. If the two carriers of Figure 1 are not transmitting at the same
`
`time, there is no improvement in message capacity of the channel. As a
`
`result, there is no need for multiple carriers.
`
`38. The ’891 Patent describes that co-location does not give rise to the
`
`near-far problem. Ex. 1001 at 4:12-15. Since all carriers are transmitted
`
`from the same location, at any receiver, all carriers are attenuated the
`
`same amount. In other words, there are no longer near and far distances.
`
`13
`
`MTel., Exhibit 2003, Juniper v. MTel., Page 13, IPR2017-00640
`
`
`
`39. Because there is no near-far problem with co-location, carriers can be
`
`spaced closer together than when subchannels are used. Id. This means
`
`that the symmetric condition is no longer optimal. Id. at 4:15-17.
`
`40. The ’891 Patent defines a new—more narrow—condition for co-
`
`located carriers. This condition provides that “the frequency spacings
`
`between adjacent carriers, while symmetric to each other, can be smaller
`
`than the frequency spacings between the band edges of the mask and the
`
`nearest respective carrier.” Id. at 4:17-20. In other words, the distance
`
`between adjacent carriers can be smaller than the distance between an
`
`outer carrier and the band edge. The ’891 Patent refers to this condition
`
`as an asymmetry. Id. at 4:24-34. Drawing 9 below illustrates this
`
`asymmetric condition.
`
`41.
`
`In Drawing 9, distance Dc is half of the frequency spacing between
`
`Carrier 1 and the next adjacent carrier, which is now Carrier 3. Distance
`
`
`
`14
`
`MTel., Exhibit 2003, Juniper v. MTel., Page 14, IPR2017-00640
`
`
`
`Dm is the spacing between Carrier 1 and the nearest band edge of the
`
`Primary Mask of the channel. The asymmetric condition of the ’891
`
`Patent, requires that Dm > Dc.
`
`42.
`
`In comparison with the symmetric condition (Drawing 8), the
`
`asymmetric condition of Drawing 9 allows closer spacing of adjacent
`
`carriers. Ex. 1001 at 4:14-15. This closer spacing allows Carrier 3 to be
`
`added between Carrier 1 and Carrier 2. The additional Carrier 3 means
`
`additional data is sent in the same channel at the same time. Thus, the
`
`addition of Carrier 3 increases the message capacity of the channel,
`
`which is the purpose of the ’891 Patent.
`
`43.
`
`In addition, the asymmetric condition allows carriers to overlap
`
`without concern of the near-far problem. Id. at 4:24-30. In Drawing 9,
`
`Carrier 1 and Carrier 3 overlap, and Carrier 3 and Carrier 2 overlap.
`
`44.
`
`Independent claims 1, 3, and 5 of the ‘891 Patent all recite (1)
`
`transmitting multiple carriers from the same location, and (2) and
`
`transmitting the multiple carriers according to the asymmetric condition.
`
`As described above, these two limitations are solutions to the problem of
`
`increasing the message capacity of the channel. As a result, the language
`
`of these claims should be considered in relation to solving the problem of
`
`increasing the message capacity of the channel.
`
`15
`
`MTel., Exhibit 2003, Juniper v. MTel., Page 15, IPR2017-00640
`
`
`
`45.
`
`In regard to the limitation of (1) transmitting multiple carriers from
`
`the same location, a PHOSITA would understand this limitation to mean
`
`transmitting multiple carriers from the same location at the same time.
`
`As described in above paragraph 36 and in reference to Figure 1 of
`
`the ’891 Patent, if the two carriers of Figure 1 are not transmitting at the
`
`same time, there is no improvement in message capacity of the channel.
`
`As a result, there is no need for co-location.
`
`46.
`
`In regard to the limitation of (2) transmitting the multiple carriers
`
`according to the asymmetric condition, a PHOSITA would understand
`
`this limitation to mean transmitting the multiple carriers so that the
`
`asymmetric condition is met. The asymmetric condition of claim 1, for
`
`example, is that “the frequency difference between the center frequency
`
`of the outer most of said carriers and the band edge of the mask defining
`
`said channel is more than half the frequency difference between the
`
`center frequencies of each adjacent carrier.” This is more simply
`
`described in Drawing 9 as Dm > (½)Dbetween adjacent carrier center frequencies, which
`
`is equivalent to Dm > Dc, where Dc is half the frequency difference
`
`between the center frequencies of adjacent carriers.
`
`47. The specification of the ’891 Patent provides sufficient guidance for a
`
`PHOSITA to identify the band edge of the mask defining said channel
`
`16
`
`MTel., Exhibit 2003, Juniper v. MTel., Page 16, IPR2017-00640
`
`
`
`referred to in the independent claims. The simplified mask of Drawing 9
`
`has only one band edge, so there is no problem determining the band
`
`edge in Drawing 9. Actual FCC emission masks, however, are more
`
`complex and rely on mathematical formulas to specifically identify the
`
`constraints imposed by the mask, and thereby, the band edge of the mask
`
`defining a given channel.
`
`48. The word “band” of “band edge” refers to a frequency band or range.
`
`A PHOSITA would, therefore, understand that the term “band edge of a
`
`mask” means all points along the edge of the mask that limits the
`
`frequency band.
`
`49.
`
`In 47 C.F.R. §90.210 of the current FCC regulations, for example, the
`
`FCC lists the emission masks for a number of different channels. The
`
`table from 47 C.F.R. §90.210 is shown below.
`
`50. Appendix C is an application note from Silicon Labs that includes
`
`figures of some current FCC emission masks. Appendix C is an
`
`
`
`17
`
`MTel., Exhibit 2003, Juniper v. MTel., Page 17, IPR2017-00640
`
`
`
`exemplary technical pamphlet that I have relied on to corroborate my
`
`personal knowledge that FCC emission masks have multiple points
`
`which together form the band edge. Figures 1-3 of Appendix C are
`
`shown below.
`
`
`
`
`
`18
`
`MTel., Exhibit 2003, Juniper v. MTel., Page 18, IPR2017-00640
`
`
`
`51. Figures 1-3 of Appendix C demonstrate that FCC emission masks can
`
`
`
`have multiple band edges. Figures 1-3 are recent FCC emission masks
`
`and are for a different frequency range than the frequency range
`
`described in the ’891 Patent. The prosecution history of the ’891 Patent,
`
`however, provides that FCC emission masks at the time of the ’891
`
`Patent had multiple band edges even for the frequency range described in
`
`the ’891 Patent. On June 7, 1995, the applicants filed an information
`
`disclosure statement (IDS) with an exemplary mask from the FCC Part
`
`22 regulations at the time. Ex. 1012 at 79-84. In the IDS they provided,
`
`the mask was drawn according to the interpretation of the FCC Part 22
`
`regulations at the time. Id. at 84. This mask is shown below.
`
`19
`
`MTel., Exhibit 2003, Juniper v. MTel., Page 19, IPR2017-00640
`
`
`
`
`
`Id.
`
`52.
`
`In determining the band edge of the asymmetric condition of the ’891
`
`Patent, a PHOSITA would first look to the purpose of the band edge in
`
`the asymmetric condition in the context of the ’891 Patent.
`
`53.
`
`In the asymmetric condition, the band edge is used to determine a
`
`frequency distance between the outer most carriers and the mask. In
`
`Drawing 9, this distance is Dm, for example. The purpose of the
`
`frequency distance, Dm, is implicitly defined in the ’891 Patent.
`
`54. Again, a PHOSITA would first look to the purpose of the frequency
`
`distance, Dm, in the context of the ’891 Patent.
`
`55. A PHOSITA would conclude from the specification of the ’891 Patent
`
`that the purpose of the frequency distance, Dm, is to prevent the outer
`
`most carriers from exceeding the mask limits when they are modulated.
`
`20
`
`MTel., Exhibit 2003, Juniper v. MTel., Page 20, IPR2017-00640
`
`
`
`In other words, the purpose of the frequency distance, Dm, is to prevent
`
`the outer most carriers from exceeding the band edge.
`
`56. This conclusion is made by first determining the type of modulation
`
`used in the ’891 Patent. Figures 5A, 6A, and 7A all depict modulated
`
`carriers of the ’891 Patent. All three figures include a maximum
`
`frequency deviation. A frequency deviation is known by a PHOSITA to
`
`be a parameter of frequency shift keying (FSK) modulation.
`
`57. Appendix D is an exemplary tutorial on digital modulation techniques
`
`provided by Electronic Design magazine. Appendix D is an exemplary
`
`technical periodical that I have relied on to corroborate my personal
`
`knowledge of modulation techniques. Appendix D describes three basic
`
`ways to modulate a sinusoidal carrier to transmit digital data. Appendix
`
`D at 1. These three ways are using amplitude shift keying (ASK), on-off
`
`keying (OOK), or frequency shift keying (FSK). These three modulation
`
`methods are graphically depicted in Figure 1 of Appendix D, which is
`
`shown below.
`
`21
`
`MTel., Exhibit 2003, Juniper v. MTel., Page 21, IPR2017-00640
`
`
`
`
`
`58. Appendix D describes that FSK modulation shifts the carrier between
`
`two different frequencies, fm and fs. fm is the mark or binary 1 frequency,
`
`and fs is the space or binary 0 frequency. The frequency deviation, ∆f, of
`
`FSK is calculated as ∆f = fs - fm.
`
`59. Because FSK modulation is the only digital modulation method to
`
`include a frequency deviation, Appendix D confirms that it is a
`
`modulation technique used in the ‘891 Patent.
`
`60. After determining that a modulation technique used in the ‘891 Patent
`
`is FSK, a PHOSITA would consider how FSK is related to the frequency
`
`distance, Dm, of the asymmetric condition.
`
`22
`
`MTel., Exhibit 2003, Juniper v. MTel., Page 22, IPR2017-00640
`
`
`
`61. A PHOSITA would realize that the frequency deviation, ∆f, of FSK
`
`modulation causes a carrier to get wider or spread out in terms of
`
`frequency. Appendix E is another tutorial on modulation provided by
`
`www.complextoreal.com as of June 12, 2016. Appendix E is an
`
`exemplary technical pamphlet that I have relied on to corroborate my
`
`personal knowledge of the frequency shift keying (FSK) modulation
`
`technique. Appendix E includes a detailed description of FSK
`
`modulation. Figure 9 of Appendix E is shown below.
`
`
`
`62. Figure 9 of Appendix E is a spectrum showing a carrier with a center
`
`frequency of 4 that has been modulated with FSK. Figure 9 shows that
`
`the carrier now has a frequency deviation, ∆f, of 2 on either side of the
`
`center frequency of 4. In other words, the shape of the carrier is now
`
`spread out between frequencies 2 and 6. As a result, Figure 9 of
`
`Appendix E confirms that the frequency deviation, ∆f, of FSK
`
`23
`
`MTel., Exhibit 2003, Juniper v. MTel., Page 23, IPR2017-00640
`
`
`
`modulation causes a carrier to get wider or spread out in terms of
`
`frequency.
`
`63. Because FSK modulation causes a modulated carrier to get wider and
`
`the frequency distance, Dm, of the asymmetric condition specifies a
`
`buffer distance between the band edge of a mask and the carrier, a
`
`PHOSITA would conclude that the purpose of the frequency distance, Dm,
`
`is to prevent the outer most carriers from exceeding the mask limits when
`
`they are spread out in frequency due to modulation.
`
`64.
`
`Indeed, the ’891 Patent confirms this by describing in reference to the
`
`modulated carriers of Figure 5A that “the carriers remained within the
`
`FCC mask while providing an acceptable error-rate versus signal strength
`
`performance.” Ex. 1001 at 4:61-63. In other words, the ’891 Patent
`
`explicitly points out that the carriers remained within the mask limits
`
`after applying the asymmetric condition and after modulation.
`
`65. As described above, the purpose of the frequency distance, Dm, is to
`
`determine what band edge to use in the calculation of the asymmetric
`
`condition. Since the purpose of the frequency distance, Dm, is to ensure
`
`that the modulated carrier does not exceed a mask limit, a PHOSITA
`
`would conclude that the band edge of the asymmetric condition is the
`
`24
`
`MTel., Exhibit 2003, Juniper v. MTel., Page 24, IPR2017-00640
`
`
`
`edge of the mask that is likely to be nearest in frequency to the outer
`
`most carrier when that carrier is modulated.
`
`66. Another way of describing the band edge, for FSK modulation, is the
`
`first limit of the mask to be exceeded as the frequency deviation of the
`
`carrier is increased.
`
`67. A PHOSITA would understand that the band edge nearest in
`
`frequency to the outer most carrier would be chosen for the asymmetric
`
`condition in order to minimize the frequency distance, Dm, of Drawing 9.
`
`In order to satisfy the spacing requirements of the claims, minimizing Dm.
`
`necessarily minimizes the frequency distance between carriers, Dc,
`
`according to the asymmetric condition (Dm > DC). This allows more
`
`carriers to be placed in the channel, which increases the message capacity
`
`of the channel, which is the stated goal of the ’891 Patent.
`
`68. Alternatively, using any other point on the mask in the measurement
`
`of Dm, results in a larger value of Dm, which in turn allows larger spacing
`
`between adjacent carriers. This would be contrary to the ’891 Patent,
`
`which teaches that “carrier spacing far closer than would ordinarily be
`
`allowed” is desirable. Ex. 1001 at 4:14-15.
`
`69. The point on the mask identified by Dr. Kakaes as the band edge of
`
`the mask is further in frequency from the outer most carrier. Ex. 1003 at
`
`25
`
`MTel., Exhibit 2003, Juniper v. MTel., Page 25, IPR2017-00640
`
`
`
`¶¶ 19-20. When the farthest frequency limit of the mask is used as the
`
`band edge for the asymmetric condition, the frequency distance, Dm, is
`
`maximized, allowing the frequency distance between carriers, Dc, to be
`
`larger. This would not allow more carriers to be placed in the channel
`
`and would not increase the message capacity of the channel. This is
`
`contrary to the teachings of the ’891 Patent.
`
`70. As shown above in the figures of FCC emission masks of above
`
`paragraph 50, the power attenuation requirements of masks are generally
`
`smallest near the center frequency of the outermost carrier, and increase
`
`as the distance from the center frequency of the outermost carrier
`
`increases. As a result, the band edges at higher power levels are nearest
`
`to the center frequency of the outer most carriers. In other words, masks
`
`generally take the shape shown in Figure 4 of the ’891 Patent. Ex. 1001
`
`at Figure 4.
`
`71. Two masks described in the ‘891 Patent and its prosecution history
`
`can be used to confirm that the band edge of the asymmetric condition is
`
`the edge of the mask that is nearest to the outer most carrier when that
`
`carrier is modulated. This is confirmed by hypothetically applying the
`
`two masks to the examples of the ’891 Patent application. Note that there
`
`is no explicit indication in the ’891 Patent that the examples of Figures
`
`26
`
`MTel., Exhibit 2003, Juniper v. MTel., Page 26, IPR2017-00640
`
`
`
`5A, 6A, or 7A were intended to work with the two masks described
`
`below and shown in Drawing 10 and Drawing 11.
`
`72. The first mask applied is the mask of Figure 4 of the ’891 Patent,
`
`which is reproduced below. Suppose, for example, the vertical mask
`
`edges at -10 kHz and 10 kHz of the mask of Figure 4 of the ’891 Patent
`
`are selected as the band edges of the asymmetric condition.
`
`
`
`73. Figure 5A of the ’891 Patent shows that the applicants of the ’891
`
`Patent selected edges at -10 kHz and 10 kHz as the band edges of the
`
`asymmetric condition. Figure 5A is reproduced below.
`
`27
`
`MTel., Exhibit 2003, Juniper v. MTel., Page 27, IPR2017-00640
`
`
`
`
`
`74. A PHOSITA would understand that Figure 5A of the ’891 Patent was
`
`meant to show that placing two carriers in a channel according to the
`
`asymmetric condition produced good results. In Figure 5A, the two
`
`carriers are centered at -4.590 kHz and 4.590 kHz. Half the frequency
`
`distance between the center frequencies of adjacent carriers, Dc, is,
`
`therefore, 4.590 kHz. If the band edges are at -10 kHz and 10 kHz, then
`
`the frequency distance between an outer carrier and a band edge, Dm, is
`
`5.410 kHz. As a result, Dm > Dc and the asymmetric condition is met.
`
`75. A PHOSITA would understand that using the mask of Figure 4 of
`
`the ’891 Patent, for the example, in Figure 5A of the ‘891 Patent would
`
`not produce the desired result if the carriers are operating at full power,
`
`even though the mask of Figure 4 includes edges at -10 kHz and 10 kHz.
`
`28
`
`MTel., Exhibit 2003, Juniper v. MTel., Page 28, IPR2017-00640
`
`
`
`This is because the edges at -10 kHz and 10 kHz are not the nearest edges.
`
`Drawing 10 below shows the mask of Figure 4 drawn on Figure 5A.
`
`76. Drawing 10 shows that not choosing the nearest edges creates a
`
`
`
`problem, because centering two carriers at -4.590 kHz and 4.590 kHz
`
`based on the vertical mask edges of the mask in Figure 4 at -10 kHz and
`
`10 kHz puts the modulated carriers outside of the mask limits. The two
`
`carriers exceed the diagonal mask edges. This would be a problem,
`
`because the ’891 Patent, in reference to Figure 5A, explicitly states that
`
`“carriers remained within the FCC mask limits.” Ex. 1001 at 4:61-62.
`
`They would not remain within the mask limits of Figure 4 of the ’891
`
`Patent.
`
`29
`
`MTel., Exhibit 2003, Juniper v. MTel., Page 29, IPR2017-00640
`
`
`
`77. A PHOSITA would also understand that using the mask of Figure 4 of
`
`the ’891 Patent, for the example, in Figure 5A of the ’891 Patent could
`
`produce the desired result if the carriers are not operating at full power.
`
`For example, if the power levels of the carriers are attenuated enough, the
`
`diagonal band edges of Figure 4 are no longer the nearest band edges,
`
`because the carriers are below them. Instead, the vertical edges at -10
`
`kHz and 10 kHz are now the nearest edges. Drawing 10b below shows
`
`the mask of Figure 4 drawn relative to the carriers of Figure 5A when the
`
`carriers are attenuated below the level of the diagonal band edges of
`
`Figure 4.
`
`78. Drawing 10b shows that if the FCC mask of Figure 4 is superimposed
`
`on Figure 5A and the power levels of the carriers are kept below the
`
`
`
`30
`
`MTel., Exhibit 2003, Juniper v. MTel., Page 30, IPR2017-00640
`
`
`
`diagonal lines of the FCC mask as shown in Fig 3B, the nearest band
`
`edges of the mask are the vertical lines of the mask.
`
`79.
`
`Indeed, there is evidence in the ’891 Patent that the carriers are
`
`restricted to lower power levels. Figures 3A and 3B of the ’891 Patent,
`
`reproduced below, show that power levels of the subchannels 30a and
`
`30b are limited to levels below where the diagonal lines of mask 31 come
`
`into play. In other words, carriers 32a and 32b of Figures 3B can never
`
`reach the diagonal lines of mask 31 when modulated, because their power
`
`levels are limited by the horizontal lines of subchannels 30a and 30b. If
`
`carriers 32a and 32b are not limited to these lower power levels, points
`
`on the diagonals of mask 31 would be the nearest band edge.
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`31
`
`MTel., Exhibit 2003, Juniper v. MTel., Page 31, IPR2017-00640
`
`
`
`