`
`
`
`Paper No. __
`
`UNITED STATES PATENT AND TRADEMARK OFFICE
`
`———————
`
`BEFORE THE PATENT TRIAL AND APPEAL BOARD
`
`———————
`
`CISCO SYSTEMS, INC., DISH NETWORK, LLC,
`COMCAST CABLE COMMUNICATIONS, LLC,
`COX COMMUNICATIONS, INC.,
`TIME WARNER CABLE ENTERPRISES LLC,
`VERIZON SERVICES CORP., and ARRIS GROUP, INC.,
`Petitioner,1
`
`v.
`
`TQ DELTA, LLC,
`Patent Owner.
`
`———————
`
`IPR No. IPR2016-01021
`U.S. Patent No. 8,718,158 B2
`
`———————
`
`PETITIONER’S REPLY
`
`
`
`
`1 DISH Network, L.L.C., who filed a Petition in IPR2017-00255, and Comcast
`Cable Communications, L.L.C., Cox Communications, Inc., Time Warner Cable
`Enterprises L.L.C., Verizon Services Corp., and ARRIS Group, Inc., who filed a
`Petition in IPR2017-00417, have been joined in this proceeding.
`
`
`
`

`

`
`
`
`
`
`Petitioner’s Reply, IPR2016-01021
`
`TABLE OF CONTENTS
`Petitioner’s Exhibit List ............................................................................................. 4 
`I. 
`Summary ............................................................................................................ 7 
`II.  Claim Construction ............................................................................................ 7 
`A. 
`“scrambling…a plurality of carrier phases” .............................................. 7 
`B. 
`“transceiver” .............................................................................................. 7 
`III.  Combining the teachings of Shively and Stopler would have been obvious
`to a POSITA ....................................................................................................... 8 
`A.  DMT-based communications systems were known to be susceptible
`to challenges caused by signals with high PAR ........................................ 9 
`B.  Transmitting the same data on multiple carriers causes a spike in
`signal amplitude and increases PAR ....................................................... 10 
`C.  Shively transmits the same data on multiple carriers, which increases
`PAR ......................................................................................................... 11 
`D.  Phase scrambling was a known technique for reducing PAR ................. 12 
`E.  Combining Stopler’s phase scrambler with Shively’s transmitter is
`merely the use of a known technique to improve a similar device. ........ 14 
`F.  Market forces would have prompted a POSITA to combine Shively
`and Stopler ............................................................................................... 15 
`IV.  Stopler’s phase scrambler reduces PAR because it scrambles phases of
`individual QAM symbols ................................................................................ 16 
`A.  Dr. Short admitted that Stopler’s phase scrambler is applied to QAM
`symbols .................................................................................................... 16 
`B.  Stopler does not describe phase scrambling DMT symbols ................... 17 
`C.  Stopler contemplates systems with multiple pilot tones and multiple
`kinds of overhead channel symbols......................................................... 19 
`D.  TQ Delta’s arguments about the ’156 patent are moot because it
`changes the phase of individual QAM symbols ...................................... 20 
`V.  TQ Delta’s remaining arguments are without merit ........................................ 22 
`A.  Shively’s technique is not limited to 18,000 foot cables ........................ 22 
`B.  Dr. Short’s analysis of a hypothetical 18,000 foot cable is flawed ......... 26 
`1.  Likelihood of phase alignment for random data ............................. 27 
`
`
`
`2
`
`

`

`
`
`
`
`
`Petitioner’s Reply, IPR2016-01021
`
`2.  Likelihood of phase alignment using Shively’s bit-spreading
`technique ......................................................................................... 28 
`3.  Comparing the phase-alignment likelihoods for Shively’s
`technique and random data .............................................................. 29 
`C.  A rigorous analysis of using Shively’s technique shows that it
`significantly increases PAR and the likelihood of clipping .................... 31 
`D.  High PAR causes more problems than just clipping ............................... 36 
`E.  Stopler’s diagonalization technique is optional, not required ................. 36 
`F.  Ground #2: Combining Bremer with Shively and Stopler would
`have been obvious ................................................................................... 38 
`VI.  Conclusion ....................................................................................................... 39 
`VII.  Certificate of Word Count ............................................................................... 40 
`
`
`
`
`
`3
`
`

`

`
`
`
`
`
`Petitioner’s Reply, IPR2016-01021
`
`Petitioner’s Exhibit List
`
`June 8, 2016
`
`Ex. 1001
`
`U.S. Patent No. 8,718,158 to Tzannes (“the ’158 patent”)
`
`Ex. 1002
`
`Prosecution File History of U.S. Pat. No. 8,718,158
`
`Ex. 1003
`
`Prosecution File History of U.S. Pat. No. 8,090,008
`
`Ex. 1004
`
`Prosecution File History of U.S. Pat. No. 7,769,104
`
`Ex. 1005
`
`Prosecution File History of U.S. Pat. No. 7,471,721
`
`Ex. 1006
`
`Prosecution File History of U.S. Pat. No. 7,292,627
`
`Ex. 1007
`
`Prosecution File History of U.S. Pat. No. 6,961,369
`
`Ex. 1008
`
`U.S. Provisional Application No. 60/164,134
`
`Ex. 1009
`
`Declaration of Dr. Jose Tellado under 37 C.F.R. § 1.68
`
`Ex. 1010
`
`Curriculum Vitae of Dr. Jose Tellado
`
`Ex. 1011
`
`U.S. Patent No. 6,144,696 to Shively et al. (“Shively”)
`
`Ex. 1012
`
`U.S. Patent No. 6,625,219 to Stopler (“Stopler”)
`
`Ex. 1013
`
`U.S. Patent No. 6,424,646 to Gerszberg et al. (“Gerszberg”)
`
`Ex. 1014
`
`Ex. 1015
`
`Harry Newton, NEWTON’S TELECOM DICTIONARY, 13th Ed. (1998)
`(selected pages)
`
`Kim Maxwell, “Asymmetric Digital Subscriber Line: Interim
`Technology for the Next Forty Years,” IEEE Communications
`Magazine (Oct. 1996).
`
`Ex. 1016 Walter Goralski, ADSL AND DSL TECHNOLOGIES (McGraw-Hill
`1998) (selected pages)
`
`
`
`
`4
`
`

`

`
`
`Ex. 1017
`
`Ex. 1018
`
`Ex. 1019
`
`
`
`Petitioner’s Reply, IPR2016-01021
`
`U.S. Patent No. 4,924,516 to Bremer et al. (“Bremer”)
`
`American National Standard for Telecommunications, Network and
`Customer Installation Interfaces—Asymmetric Digital Subscribers
`Line (ADSL) Metallic Interface (ANSI T1.413-1995)
`
`U.S. Patent No. 5,515,369 to Flammer, III et al. (“Flammer”)
`
`Ex. 1020
`
`Declaration of David Bader
`
`Ex. 1021
`
`Fig. 6 from Ex. 2009 (T. Regan, “ADSL Line Driver/Receiver
`Design Guide, Part 1” (February 2000)).
`
`Ex. 1022
`
`Robert T. Short, “Physical Layer,” in WIMEDIA UWB (2008).
`
`Ex. 1023
`
`Ex. 1024
`
`Ex. 1025
`
`Denis J. G. Mestdagh and Paul M. P. Spruyt, “A Method to Reduce
`the Probability of Clipping in DMT-Based Transceivers,” IEEE
`Transactions on Communications, Vol. 44, No. 10, (October 1996).
`
`Stefan H. Muller and Johannes B. Huber, “A Comparison of Peak
`Power Reduction Schemes for OFDM,” IEEE Global
`Telecommunications Conference (1997).
`
`Jose Tellado-Mourelo, “Peak to Average Power Reduction for
`Multicarrier Modulation,” A dissertation submitted to the
`Department of Electrical Engineering and the Committee on
`Graduate Studies of Stanford University (Sept. 1999)
`
`Ex. 1026
`
`Second Declaration of Dr. Jose Tellado under 37 C.F.R. § 1.68
`
`Ex. 1027
`
`Deposition Transcript of Dr. Robert T. Short
`
`Ex. 1028
`
`T. Starr, J. M. Cioffi, P. J. Silverman, UNDERSTANDING DIGITAL
`SUBSCRIBER LINE TECHNOLOGY (1999) (selected pages).
`
`Ex. 1029
`
`Abe, RESIDENTIAL BROADBAND (2000) (selected pages).
`
`Ex. 1030 Mohamed Zekri, et al., “DMT Signals with Low Peak-to-Average
`Power Ratio,” Proceedings of the IEEE International Symposium
`on Computers and Communications (held July 6-8, 1999).
`
`Ex. 1031
`
`Second Declaration of David Bader
`
`
`
`5
`
`

`

`
`
`Ex. 1032
`
`Ex. 1033
`
`
`
`Petitioner’s Reply, IPR2016-01021
`
`Peter S. Chow, et al., “A Practical Discrete Transceiver Loading
`Algorithm for Data Transmission over Spectrally Shaped
`Channels”, IEEE Transactions on Communications, Vol. 43, No.
`2/3/4 (1995).
`
`Kamran Sistanizadeh, et al., “Multi-Tone Transmission for
`Asymmetric Digital Subscriber Lines (ADSL)”, Communications,
`1993. ICC '93 Geneva. Technical Program, Conference Record,
`IEEE International Conference (held May 23-26, 1993)
`
`Ex. 1034
`
`
`ADSL transmitter simulation program by Dr. Tellado
`
`
`
`
`6
`
`

`

`
`
`I.
`
`Summary
`
`
`
`Petitioner’s Reply, IPR2016-01021
`
`Patent Owner TQ Delta raises various arguments, but they are all severely
`
`flawed, whether by a misinterpretation of the prior art or a misunderstanding of the
`
`technology involved. The obviousness rationale for combining Shively and
`
`Stopler derives almost entirely from a POSITA’s straightforward consideration of
`
`undisputed facts. The only substantive limitation in dispute—“scrambling”—is
`
`plainly taught by Stopler, which even uses the same “scrambling” terminology to
`
`describe the concept. Accordingly, the Board should issue a Final Written
`
`Decision holding all claims of the ’158 patent unpatentable for obviousness.
`
`II.
`
`Claim Construction
`
`A.
`
`“scrambling…a plurality of carrier phases”
`
`TQ Delta proposes construing the phrase “scrambling…a plurality of carrier
`
`phases” to mean “adjusting the phases of a plurality of carriers in a single
`
`multicarrier symbol by pseudo-randomly varying amounts.” Resp., pp.14-19.
`
`The “scrambling” phrase does not need construction, since the prior art
`
`relied upon—Stopler—uses the same “phase scrambling” terminology to describe
`
`pseudo-random phase changes. CSCO-1012, 12:24-31. Accordingly, the Board
`
`should not adopt TQ Delta’s proposed construction.
`
`B.
`
` “transceiver”
`
`TQ Delta notes that a district court—which applies a different claim
`
`
`
`7
`
`

`

`
`
`interpretation standard—adopted a different claim construction for “transceiver.”
`
`Petitioner’s Reply, IPR2016-01021
`
`
`
`Resp., pp.13-14. But TQ Delta does not show any error in the Board’s currently-
`
`adopted construction. Accordingly, there is no reason for the Board to modify its
`
`construction.
`
`III. Combining the teachings of Shively and Stopler would have been
`obvious to a POSITA
`
`TQ Delta argues that it would not have been obvious to combine the DSL
`
`technologies of Shively and Stopler. Resp., pp.44-45. TQ Delta further alleges that
`
`Cisco’s obviousness rationale suffers from hindsight bias. Id., p.49. But the
`
`obviousness rationale is based on agreed-upon facts and common engineering
`
`sense. As discussed further below, there is no dispute that:
`
` multicarrier communication systems generate signals with a high peak-
`
`to-average power ratio (PAR), which is undesirable;
`
` repeating the same bits on multiple carriers, the technique taught in
`
`Shively, increases PAR;
`
` scrambling the phases of individual signal carriers was a known
`
`technique for reducing PAR; and
`
` Stopler teaches a phase scrambler.
`
`In light of these facts, a POSITA would have found it obvious to incorporate
`
`a phase scrambler—like that in Stopler—into Shively’s system to counteract the
`
`increase in PAR caused by Shively’s bit spreading technique. CSCO-1009, ¶¶67-
`
`
`
`8
`
`

`

`
`
`68. Rather than hindsight, the combination would have been a straightforward
`
`Petitioner’s Reply, IPR2016-01021
`
`
`
`application of ordinary engineering sense to basic facts known to a POSITA.
`
`CSCO-1009, ¶¶18, 67-68.
`
`A. DMT-based communications systems were known to be
`susceptible to challenges caused by signals with high PAR
`
`There is no dispute that multicarrier systems using discrete multitone
`
`(“DMT”) technology faced issues relating to signals with a high peak-to-average
`
`ratio (“PAR”). See TQ-2003, ¶¶20-22. Similarly, there is no dispute that signals
`
`with high PAR can induce clipping in the transmitter circuitry. See TQ-2003,
`
`¶¶24-25, 30. These facts are succinctly stated in the ANSI T1.413-1995 standard:
`
`A DMT time-domain signal has a high peak-to-average ratio
`(its amplitude distribution is almost Gaussian), and large values
`may be clipped by the digital-to-analog converter.
`
`CSCO-1018, §6.5 (p.36).
`
`Cisco’s expert, Dr. Tellado, testified—and TQ Delta has not disputed—that
`
`a POSITA would have known such background information:
`
`[A]n understanding of the ’158 patent requires… appreciation
`for the potential for such multicarrier signals to have a high
`peak-to-average ratio, causing clipping during transmission.
`Such knowledge would be within the level of skill in the art.
`
`CSCO-1009, ¶18.
`
`The parties also agree that engineers knew that one way to reduce the
`
`
`
`9
`
`

`

`
`
`likelihood of clipping was to design transmitters to handle a greater dynamic range.
`
`Petitioner’s Reply, IPR2016-01021
`
`
`
`TQ-2003, ¶26. But it was also known that such transmitters were more expensive,
`
`less efficient, consumed more electricity, and generated more heat. TQ-2003, ¶26;
`
`CSCO-1027, 45:21-46:6. Thus, engineers would have sought out other techniques
`
`to reduce PAR. CSCO-1027, 46:23-47:3.
`
`B.
`
`Transmitting the same data on multiple carriers causes a spike in
`signal amplitude and increases PAR
`
`There is no dispute that when many carriers of a multicarrier signal have the
`
`same phase, the result is a signal with “large spikes in amplitude and, therefore, a
`
`high PAR.” TQ-2003, ¶22. TQ Delta acknowledged that high PAR occurs “where
`
`the same bit or bits is/are purposely sent in a redundant manner on multiple
`
`carriers.” Resp., pp. 6-7; TQ-2003, ¶22; CSCO-1027, 97:13-15. As TQ Delta’s
`
`declarant Dr. Short illustrated in the figure below, when the signal carriers carry
`
`the same bits, they have the same phases, which add coherently to create a
`
`transmission signal with a large spike in amplitude. TQ-2003, ¶22. TQ Delta
`
`admitted that such signal has a “high PAR.” Resp., p.7.
`
`
`
`10
`
`

`

`
`
`
`
`
`Petitioner’s Reply, IPR2016-01021
`
`TQ-2003, ¶18.
`
`
`
`C.
`
`Shively transmits the same data on multiple carriers, which
`increases PAR
`
`In Shively’s system the same bits are purposely sent in a redundant manner
`
`on multiple carriers, which TQ Delta admits leads to high PAR.2 CSCO-1011,
`
`11:17-18; Resp., p.10; CSCO-1027, 9:7-10. Because the same data (Shively’s “k-
`
`bit symbol”) is modulated on multiple carriers, the carriers will phase align and
`
`will add coherently to create a transmission signal with a spike in amplitude
`
`
`2 Elsewhere TQ Delta argues that Shively does not suffer from increased PAR
`
`(Resp., pp.47-49), but that argument is based on an erroneous analysis of a
`
`hypothetical 18,000-foot ADSL line. TQ Delta and its expert repeatedly concede
`
`that Shively transmits the same data on multiple carriers, which increases PAR.
`
`Resp., pp.10; TQ-2003, ¶22; CSCO-1027, 96:16-20, 97:10-15, 100:4-7.
`
`
`
`11
`
`

`

`
`
`(power). TQ-2003, ¶ 22; CSCO-1027, 97:21-23. This spike would not have
`
`Petitioner’s Reply, IPR2016-01021
`
`
`
`occurred without Shively’s technique, since the carriers would have been deemed
`
`unusable, and no data would have been transmitted on those carriers. CSCO-1026,
`
`¶6. Thus, Shively’s technique creates new amplitude spikes in the multicarrier
`
`signal and causes an increase in PAR. Id.; Resp., pp.6-7.
`
`D.
`
`Phase scrambling was a known technique for reducing PAR
`
`A POSITA would have known that one way to reduce PAR is to scramble
`
`phases of individual carriers. As Dr. Tellado stated, “phase scrambling was
`
`probably the most popular way” to reduce PAR. TQ-2002, 100:9:13; CSCO-1027,
`
`9:11-13 (“just by rotating the symbols, then you can reduce the peak-to-average
`
`ratio.”).
`
`Prior art technical publications confirm that phase scrambling was a known
`
`way to reduce PAR. A 1996 article notes the use of “discrete multitone (DMT)
`
`modulation technique … for applications [such as] asymmetric digital subscriber
`
`line (ADSL).” CSCO-1023, p.1234; CSCO-1031, ¶2. The article describes prior
`
`efforts “to reduce the peak-to average power ratio of the DMT signal,” and
`
`proposes a technique employing a “random phasor transformation.” CSCO-1023,
`
`p.1234-35. Dr. Short agreed that such phasor transformations include phase
`
`scrambling. CSCO-1027, 77:18-20.
`
`Another 1997 article notes that “it is highly desirable to reduce the PAR” of
`
`
`
`12
`
`

`

`
`
`a multicarrier signal. CSCO-1024, p.1; CSCO-1031, ¶3. To do so, the article
`
`Petitioner’s Reply, IPR2016-01021
`
`
`
`describes a transmitter that “constructs its transmit signal with low PAR by
`
`coordinated addition of appropriately phase rotated signal parts” where the signal
`
`parts are subcarriers within a multicarrier vector. CSCO-1024, p.1. This technique
`
`was applied to DMT systems, where the DMT tones (carriers) had a “phase
`
`rotation applied to each tone.” CSCO-1028, p.238; CSCO-1031, ¶7.
`
`Thus, prior to November 9, 1999 (the earliest claimed priority date of the
`
`’158 Patent), multiple engineers working with DMT-based systems had written
`
`about the use of phase scrambling to reduce PAR. These articles support Dr.
`
`Tellado’s opinion that a POSITA would have been familiar with phase scrambling
`
`as a known technique for reducing PAR. CSCO-1004, ¶¶60, 67; CSCO-1026, ¶54.
`
`TQ Delta presents no evidence to the contrary. While, TQ Delta’s declarant
`
`Dr. Short agreed that randomizing phases of individual carriers is a way to reduce
`
`PAR, he pointedly declined to provide any opinion regarding the fact that phase
`
`scrambling was known in the prior art. CSCO-1027, 50:3-8, 51:4-11.
`
`Thus, all of the evidence of record supports the unchallenged conclusion that
`
`a POSITA in the prior art timeframe would have been familiar with phase
`
`scrambling as a technique for reducing PAR in DMT systems.
`
`
`
`13
`
`

`

`
`
`
`
`
`Petitioner’s Reply, IPR2016-01021
`
`E. Combining Stopler’s phase scrambler with Shively’s transmitter
`is merely the use of a known technique to improve a similar
`device.
`
`TQ Delta argues that Cisco failed to sufficiently explain how the Shively
`
`and Stopler combination is the use of a known technique to improve a similar
`
`device in the same way. Resp., pp.45-47. Instead, TQ Delta argues that Cisco
`
`used hindsight bias to make the combination. Resp., pp.49-50. These assertions
`
`are unfounded.
`
`First, the Petition unambiguously identified the “known technique” as
`
`Stopler’s use of a phase scrambler to randomize the phases of subcarriers, which
`
`improves (reduces) PAR. Pet., p.14.
`
`Second, the Petition explains that “[c]ombining Stopler’s phase scrambler
`
`into Shively’s transmitter would have been a relatively simple and obvious solution
`
`to reduce Shively’s PAR.” Pet., p.15. Thus, in the combination, the phase
`
`scrambler improves Shively’s transmitter (by reducing PAR) in the same way that
`
`the phase scrambler improves Stopler’s transmitter.
`
`Finally, the Petition explains that Shively and Stopler describe similar
`
`devices, specifically, “multicarrier communications apparatuses, such as modems.”
`
`Pet., p.17.
`
`Thus, the combination is merely the use of a known technique (phase
`
`scrambling) to improve a similar device (multicarrier modem) in the same, known
`
`
`
`14
`
`

`

`
`
`way (to reduce PAR). CSCO-1009, ¶¶ 62-70.
`
`
`
`Petitioner’s Reply, IPR2016-01021
`
`F. Market forces would have prompted a POSITA to combine
`Shively and Stopler
`
`TQ Delta argues that there were no “market forces” to prompt the
`
`combination of Shively’s and Stopler’s techniques.” Resp., pp.55-57.
`
`But the most basic of market forces—cost—was recognized at the time as
`
`influencing the development of multicarrier technology. Multiple engineers wrote
`
`of the importance of minimizing cost:
`
`Cost – This is the dominant factor when dealing with consumer
`markets.
`
`CSCO-1029, p.70; CSCO-1031, ¶8.
`
`Success of a service (and its underlying technology) depends
`greatly on its price and its relation to available alternatives.
`Service price, in turn, depends greatly on the cost of equipment
`and labor costs for operation….
`
`CSCO-1028, p.17.
`
`As Dr. Short admitted, an increase in PAR is associated with more
`
`expensive communication equipment. CSCO-1027, 45:15-46:12. The drive to
`
`reduce equipment costs would have motivated a POSITA to include Stopler’s
`
`phase scrambler in Shively’s transmitter to reduce PAR. CSCO-1026, ¶54. Thus,
`
`combining Shively and Stopler would have been obvious for this additional reason.
`
`
`
`15
`
`

`

`
`
`IV.
`
`
`
`Petitioner’s Reply, IPR2016-01021
`
`Stopler’s phase scrambler reduces PAR because it scrambles phases of
`individual QAM symbols
`
`TQ Delta argues that Stopler’s phase scrambler does not reduce PAR
`
`because the phase scrambler is applied to DMT symbols, not to individual QAM
`
`symbols. See Resp., pp.14-18, 57-59.3 But TQ Delta’s argument is based on an
`
`illogical reading of Stopler and undermined by the testimony of its expert.
`
`A. Dr. Short admitted that Stopler’s phase scrambler is applied to
`QAM symbols
`
`Stopler states that “the phase scrambler is applied to all symbols.” CSCO-
`
`1012, 12:26-27. Dr. Short agreed that in this statement, Stopler refers to phase
`
`scrambling QAM symbols:
`
`A. …“However, to simplify implementation, the phase
`scrambler is applied to all symbols,” not just the overhead
`symbols, so he is implying that he's rotating the entire bank of
`symbols.…
`Q.
`In that sentence, however, to simplify what you read --
`A. Yes.
`Q.
`-- it says it applies to all symbols, right?
`A.
`Correct.
`And those symbols are QAM symbols?
`Q.
`Correct.
`A.
`
`
`3 A DMT symbol comprises multiple signal carriers, and signal carriers are also
`
`referred to as QAM symbols.
`
`
`
`16
`
`

`

`
`
`CSCO-1027, 59:9-12, 60:15-22 (emphasis added).
`
`
`
`Petitioner’s Reply, IPR2016-01021
`
`More broadly, Stopler uses the word “symbol” multiple times in describing
`
`the phase scrambler, and Dr. Short agreed that these refer to QAM symbols:
`
`CSCO-1027,
`54:17-55:3.
`
`CSCO-1027,
`55:19-24.
`
`CSCO-1027,
`59:9-12 &
`60:15-22.
`
`
`
`CSCO-1012, 12:20-28 (annotated).
`
`Furthermore, there is no dispute that the input to Stopler’s phase scrambler is
`
`a sequence of QAM symbols. CSCO-1027, 58:6-8. Thus, there is no dispute that
`
`Stopler’s phase scrambler takes in QAM symbols and applies phase scrambling to
`
`QAM symbols.
`
`B.
`
`Stopler does not describe phase scrambling DMT symbols
`
`TQ Delta’s argument—that a POSITA would have understood Stopler as
`
`teaching the phase scrambling of DMT symbols—is illogical and has no basis in
`
`Stopler’s text. Resp., p.34. Indeed, Dr. Short admitted that Stopler does not
`
`describe applying the phase scrambler to a DMT symbol:
`
`Q. Well, you would agree with me that [Stopler] doesn't
`
`
`
`17
`
`

`

`
`
`
`
`
`Petitioner’s Reply, IPR2016-01021
`
`expressly teach applying the phase scrambler to the DMT as a
`whole?
`A.
`I would agree with that.
`
`CSCO-1027, 60:11-14.
`
`Stopler states that the purpose of the phase scrambler is to “randomize the
`
`overhead channel symbols.” CSCO-1012, 12:24. Since overhead channel symbols
`
`are QAM symbols (see CSCO-1027, 55:19-24), Stopler’s intent is to randomize the
`
`phases of QAM symbols. Randomizing DMT symbols—as TQ Delta argues—
`
`would not achieve Stopler’s stated purpose for the phase scrambler. CSCO-1026,
`
`¶57.
`
`The illogic of TQ Delta’s argument is further demonstrated by the broader
`
`context of the phase scrambler in Stopler’s system. CSCO-1026, ¶¶55-58. Stopler
`
`contemplates that the phase scrambler could be used with either a DMT or CDMA
`
`modulator. CSCO-1012, 12:55-57; see also Resp., pp.32-33. Since a CDMA
`
`modulator does not employ DMT symbols, there is no reason for Stopler’s phase
`
`scrambler to operate on DMT symbols. CSCO-1026, ¶58. In contrast, both DMT
`
`and CDMA modulators employ QAM symbols. Id. Thus the straightforward
`
`reading of Stopler—as applying the phase scrambler to individual QAM
`
`symbols—is the only possible reading that is logically and technically coherent.
`
`Id.
`
`
`
`18
`
`

`

`
`
`
`
`
`Petitioner’s Reply, IPR2016-01021
`
`C.
`
`Stopler contemplates systems with multiple pilot tones and
`multiple kinds of overhead channel symbols
`
`TQ Delta argues that Stopler should be understood as phase scrambling
`
`DMT symbols because the phase scrambler’s purpose is to randomize the overhead
`
`channel symbols, and each DMT symbol in the ANSI T1.413-1995 standard has
`
`only one pilot tone. Resp., p.40; see CSCO-1018, p.46. This argument fails for
`
`multiple reasons, including that it ignores the plain text of Stopler.
`
`First, Stopler does not limit overhead channel symbols to a single pilot tone
`
`per DMT symbol. Indeed, Stopler refers elsewhere to multiple “pilot tones.”
`
`CSCO-1012, 12:51-52. Furthermore, Stopler’s techniques are not limited to ANSI
`
`T1.413-1995 or even DMT, but are applicable to any “particular signal modulation
`
`desired.” CSCO-1012, 12:56-57. As Dr. Short admitted, other multicarrier
`
`technologies can use more than one pilot tone. CSCO-1027, 61:15-18.
`
`Second, pilot tones are just an example of overhead symbols. See CSCO-
`
`1012, 9:62 (“overhead signals, such as pilot tones”) & 12:51-52 (“overhead bits
`
`(e.g., pilot tones)”) (emphases added). Many other kinds of overhead channel
`
`symbols are also possible. For example, the TIE/EIA-95 Standard referenced in
`
`Stopler describes multiple overhead channels that include one pilot channel, one
`
`sync channel and up to 9 paging channels. TQ-2005, pp.728-729; CSCO-1012,
`
`12:61-63. Thus, the entire logic of TQ Delta’s argument is based on the incorrect
`
`premise that Stopler assumed that the only overhead would be a single ANSI
`
`
`
`19
`
`

`

`
`
`T1.413-1995 pilot tone.
`
`
`
`Petitioner’s Reply, IPR2016-01021
`
`A POSITA would have understood that data in the overhead channel
`
`symbols will probably not be random, but is likely to be highly structured. CSCO-
`
`1009, p.24. In the ANSI T1.413-1995 standard, for example, the bits encoding the
`
`pilot tone are held constant at zero. CSCO-1018, p.64. Such non-random,
`
`structured data increases the likelihood for phases of carriers to align, thereby
`
`increasing PAR. CSCO-1009, ¶59. To break up the structured data, Stopler
`
`employs a phase scrambler that scrambled phases of overhead channel symbols,
`
`and thereby reduces PAR. CSCO-1009, ¶60; CSCO-1026, ¶¶57-58.
`
`D. TQ Delta’s arguments about the ’156 patent are moot because it
`changes the phase of individual QAM symbols
`
`TQ Delta argues that Stopler’s phase scrambler operates on DMT symbols
`
`because it is implementing a noise-immunity technique described in
`
`U.S.6,370,156. Resp., pp.38-40. This argument fails for several reasons.
`
`First, TQ Delta does not offer any explanation for how Stopler, the inventor,
`
`would have learned of the noise-immunity technique in the ’156 patent, which was
`
`not published before it was granted in 2002. Thus, at the time of the Stopler
`
`patent’s filing in 1999, the inventor Stopler could not have known of the patent
`
`application that resulted in the ’156 patent.
`
`Second, the ’156 patent does not scramble DMT symbols. Rather, the ’156
`
`patent describes sequencing the pilot carrier—and only the pilot carrier—through
`
`
`
`20
`
`

`

`
`
`four different phase values. See TQ-2004, 7:5-6, 7:40-45. Dr. Short agreed that the
`
`Petitioner’s Reply, IPR2016-01021
`
`
`
`’156 Patent rotates only the phase of the pilot carrier (tone), not the phase of an
`
`entire DMT symbol. CSCO-1027, 67:4-9.
`
`Thus, not even the ’156 patent describes changing the phase of an entire
`
`DMT symbol. Even if Stopler used the ’156 patent’s technique—and the evidence
`
`discussed below suggests that it does not—that fact would still not support TQ
`
`Delta’s assertion that Stopler’s phase scrambler is applied to DMT symbols.
`
`Third, there are significant technical differences between Stopler’s phase
`
`scrambler and the technique described in the ’156 patent. The ’156 patent changes
`
`the phase of only the pilot carrier. TQ-2004, 7:40-45; CSCO-1027, 67:4-9.
`
`Stopler’s phase scrambler, however, is applied not just to the pilot tone, but to “all
`
`symbols.” CSCO-1012, 12:26-28.
`
`Stopler and the ’156 patent also differ in their phase rotation sequences. The
`
`’156 patent changes its pilot carrier using a repeating sequence of four rotations.
`
`TQ-2004, 7:40-45; see also CSCO-1027, 76:24-77:5 (explaining that the ’156
`
`patent does not require random phase changes). Stopler, on the other hand, uses a
`
`“phase scrambling sequence… generated by a pseudo-random generator,” making
`
`the phase changes effectively random. CSCO-1012, 12:28-29.
`
`Thus, there is virtually no similarity between Stopler’s phase scrambler and
`
`the ’156 patent’s technique, and therefore no reason to believe that Stopler’s
`
`
`
`21
`
`

`

`
`
`description of the phase scrambler should be read as using ideas from the ’156
`
`Petitioner’s Reply, IPR2016-01021
`
`
`
`patent. And even if it were, the ’156 patent does not rotate the phase of entire
`
`DMT symbols, so there is no reason to disturb the natural reading of Stopler as
`
`applying its phase scrambler to QAM symbols.
`
`V.
`
`TQ Delta’s remaining arguments are without merit
`
`TQ Delta raises a variety of additional arguments, none of which hold up to
`
`even moderate scrutiny.
`
`A.
`
`Shively’s technique is not limited to 18,000 foot cables
`
`In support of its allegation that Shively’s technique does not increase PAR,
`
`TQ Delta presents a supposed analysis of Shively’s PAR for an 18,000 foot cable.
`
`Resp., pp.27-29. While that analysis is deeply flawed—as discussed below—the
`
`entire premise for looking at an 18,000 foot example is illusory. Shively is not
`
`limited to cable lengths of 18,000 feet. Shively merely mentions this length in
`
`passing as an example of where “high signal attenuation… is usually observed.”
`
`CSCO-1011, 9:66-10:1. Notably, the T1.413-1995 standard defines multiple “test
`
`loops,” including loops of ten, 6000, 9000, and 12000 feet. CSCO-1018, p.118;
`
`TQ-2002, 57:3-4 (“modems have to work with different lengths of loops, not only
`
`18,000 feet”).
`
`TQ Delta’s attempt to pigeonhole Shively’s technique to 18,000-foot cables
`
`is also inconsistent with Shively’s disclosure. Shively describes its bit-spreading
`
`
`
`22
`
`

`

`
`
`technique as a way to use “impaired parts of the frequency band,” and to
`
`Petitioner’s Reply, IPR2016-01021
`
`
`
`“compensate for high attenuation and/or high noise” in such impaired subcarriers.
`
`CSCO-1011, 15:58-59, 15:50. Dr. Short admitted that other impairments—
`
`including crosstalk noise—occur on line lengths of less than 18,000 feet and
`
`agreed that Shively’s technique could be used on lines with noise-induced
`
`impairments. CSCO-1027, 24:8-25:10, 93:11-94:8.
`
`The ANSI T1.413-1995 standard describes crosstalk as potentially
`
`significant noise source in an ADSL system. There are many sources of crosstalk
`
`and relatively short lines can have significant crosstalk noise. CSCO-1026, ¶7.
`
`The ANSI standard provides multiple graphs showing the potential near-end cross
`
`talk (“NEXT”) noise levels caused by various kinds of adjacent communication
`
`lines. As illustrated below, the NEXT noise levels are above the -140 dBm/Hz
`
`background noise floor:
`
`
`
`23
`
`

`

`
`
`
`
`
`Petitioner’s Reply, IPR2016-01021
`
`background
`noise floor
`
`background
`noise floor
`
`
`
`
`
`
`
`24
`
`

`

`
`
`
`
`
`Petitioner’s Reply, IPR2016-01021
`
`background
`noise floor
`
`CSCO-1026, ¶¶9-11; CSCO-1018, pp.138, 140, 142.
`
`The ANSI T1.413-1995 standard also provides an example graph showing
`
`potential far-end crosstalk in an ADSL system:
`
`
`
`
`
`25
`
`

`

`
`
`
`
`
`Petitioner’s Reply, IPR2016-01021
`
`background
`noise floor
`
`
`
`CSCO-1026, ¶12; CSCO-1018, p.144.
`
`These graphs of crosstalk noise levels show that crosstalk noise can be a
`
`significant line impairment, with the crosstalk noise exceeding the -140 dB/Hz
`
`background noise level. CSCO-1026, ¶13. Because crosstalk can be significant on
`
`even short lines, Shively’s technique could be usefully applied to any line length.
`
`CSCO-1026, ¶14. Accordingly, there is no basis for TQ Delta’s attempt to limit
`
`Shively’s technique to only 18,000-foot cables.
`
`B. Dr. Short’s analysis of a hypothetical 18,000 foot cable is flawed
`
`Because Shively’s technique is not limited to 18,000 foot cables, the entire
`
`premise of Dr. Short’s supposed analys