`WASHINGTON, D.C.
`
`Before The Honorable Charles E. Bullock
`Chief Administrative Law Judge
`
`In the Matter of:
`
`Certain Static Random Access Memories And
`Products Containing Same
`
`Investigation No. 337-TA-792
`
`COMPLAINANT CYPRESS SEMICONDUCTOR CORPORATION’S RESPONSE TO
`RESPONDENTS’ PETITION FOR REVIEW OF THE REMAND INITIAL
`DETERMINATION ON VALIDITY AND ENFORCEABILITY
`
`1
`
`Qualcomm Incorporated
`EX1019
`Page 1 of 68
`
`
`
`TABLE OF CONTENTS
`
`Page
`
`I.
`
`II.
`
`III.
`
`INTRODUCTION ............................................................................................................. 1
`
`THE ’134 PATENT IS VALID ......................................................................................... 2
`A.
`Background............................................................................................................ 2
`B.
`Respondents “Undisputed Limitations” Argument is Wrong................................ 5
`C.
`The Stephens Device’s Burst May Be Interrupted ................................................ 8
`D.
`Stephens Cannot Meet the Limitations of Dependent Claims 2, 14 or 15............. 9
`E.
`Respondents Also Rely on “Undisputed Limitations” with Takasugi................. 10
`F.
`Respondents’ Speculation About Other Possible Embodiments Does Not
`Meet Their Burden............................................................................................... 11
`Takasugi Fails to Anticipate Dependent Claims 2, 14, or 15 by Clear and
`Convincing Evidence........................................................................................... 12
`
`G.
`
`THE ’937 PATENT IS VALID ....................................................................................... 13
`A.
`Background.......................................................................................................... 13
`B.
`Respondents Cannot Rely on Purported “Undisputed Limitations”.................... 16
`C.
`The Pawlowski Patent Does Not Use Complementary Clock Edges .................. 16
`D.
`The Pawlowski Patent Does Not Have Separate Input and Output Buses .......... 19
`E.
`The Supposed Combinations of Pawlowski plus either Kumanoya or
`Sakaue Fail to Cure the Deficiencies for Claim 12, and Are Technically
`Improbable ........................................................................................................... 20
`The Suzuki Reference Is Not Enabled................................................................. 22
`Respondents’ Proposed Combination of Suzuki plus either Kumanoya or
`Sakaue for Claim 12 Fails to Render the Claim Obvious, and Is
`Technically Improbable ....................................................................................... 25
`
`F.
`G.
`
`IV.
`
`THE ’477 PATENT IS VALID AND ENFORCEABLE................................................ 26
`
`A.
`B.
`
`C.
`D.
`
`E.
`
`Background.......................................................................................................... 26
`Hronik Goes To A Different Technology And Does Not Anticipate Claim
`8............................................................................................................................ 30
`Hronik And Jiang In Combination Do Not Make Claim 9 Obvious ................... 32
`Respondents Misrepresent the Jiang Patent; It Does Not
`Anticipate Claim 8 ............................................................................................... 33
`There Was No Inequitable Conduct..................................................................... 35
`1.
`The
` Was Not Material ......................................... 35
`2.
`There Was No Intent To Deceive ............................................................ 40
`
`-i-
`
`Page 2 of 68
`
`
`
`TABLE OF CONTENTS
`(continued)
`
`Page
`
`V.
`
`THE ’805 PATENT IS VALID ....................................................................................... 44
`A.
`Background.......................................................................................................... 44
`B.
`The Reexamination Proceeding is Irrelevant to This Case and the Request
`For Judicial Notice Is Untimely........................................................................... 47
`The ALJ Correctly Found That the Diagrams in Respondents’ Prior Art
`Are Silent on Relative Proportions ...................................................................... 49
`“Substantially Oblong” Does Not Subsume “Rectangular” ................................ 52
`Ishida ’041 Discloses the Exact Opposite N-region/P-region Arrangement ....... 57
`
`D.
`E.
`
`C.
`
`VI.
`
`CONCLUSION................................................................................................................ 59
`
`-ii-
`
`Page 3 of 68
`
`
`
`TABLE OF AUTHORITIES
`
`CASES
`1st Media, LLC v. Elec. Arts, Inc.,
`694 F.3d 1367 (Fed. Cir. 2012)....................................................................................39, 40, 43
`
`Page(s)
`
`Applied Med. Res. Corp. v. United States Surgical Corp.,
`147 F.3d 1374 (Fed. Cir. 1998)................................................................................................19
`
`Automed Techs., Inc. v. Microfil, LLC,
`No. 2006-1620, 2007 U.S. APP. LEXIS 16956 (Fed. Cir. July 16, 2007) ........................52, 53
`
`Aventis Pharma S.A. v. Hospira, Inc.,
`675 F.3d 1324 (Fed. Cir. 2012) (“Aventis”).............................................................................39
`
`Cambridge Prods., Ltd. v. Penn Nutrients, Inc.,
`962 F.2d 1048 (Fed. Cir. 1992)......................................................................................6, 30, 35
`
`Certain Automated Media Library Devices,
`Inv. No. 337-TA-746, Comm’n Op. ..........................................................................................7
`
`Finnigan Corp. v. U.S. Int’l Trade Comm’n,
`180 F.3d 1354 (Fed. Cir. 1999)................................................................................................59
`
`Hartness Int’l, Inc. v. Simplimatic Eng’g Co.,
`819 F.2d 1100 (Fed. Cir. 1987)................................................................................................57
`
`Hockerson-Halderstadt, Inc. v. Avia Grp. Int’l,
`222 F.3d 951 (Fed. Cir. 2000)......................................................................................49, 50, 54
`
`Hybritech Inc. v. Monoclonal Antibodies, Inc.,
`802 F.2d 1367 (Fed. Cir. 1986)..................................................................................................6
`
`In the Matter of Certain Liquid Crystal Display Devices and Products Containing the
`Same,
`Investigation No. 337-TA-631, 2010 ITC LEXIS 2910 (U.S. I.T.C. 2010)............................50
`
`In re Rosuvastatin Calcium Patent Litig.,
`703 F.3d 511 (Fed. Cir. 2012)............................................................................................37, 39
`
`In re Ruskin,
`347 F.2d 843 (C.C.P.A. 1965) .................................................................................................19
`
`In re Wright,
`569 F.2d 1124 (C.C.P.A. 1977) ...............................................................................................50
`
`iii
`
`Page 4 of 68
`
`
`
`TABLE OF AUTHORITIES
`(continued)
`
`Kalman v. Kimberly-Clark Corp.,
`713 F.2d 760 (Fed. Cir.1983), cert. denied, 465 U.S. 1026 (1984) .........................................19
`
`Kao Corp. v. Unilever U.S., Inc.,
`441 F.3d 963 (Fed. Cir. 2006)..................................................................................................59
`
`Kinetic Concepts, Inc. v. Smith & Nephew, Inc.,
`688 F.3d 1342 (Fed. Cir. 2012)..........................................................................................20, 21
`
`Krippelz v. Ford Motor Co.,
`667 F.3d 1261 (Fed. Cir. 2012)..........................................................................................49, 50
`
`Minn. Mining & Mfg. Co. v. Chemque, Inc.,
`303 F.3d 1294 (Fed. Cir. 2002)..................................................................................................5
`
`Minn. Mining & Mfg. Co. v. Johnson & Johnson Orthopedics, Inc.,
`976 F.2d 1559 (Fed. Cir. 1992)................................................................................................16
`
`Nystrom v. TREX Co.,
`424 F.3d 1136 (Fed. Cir. 2005)................................................................................................50
`
`PowerOasis, Inc. v. T-Mobile USA, Inc.,
`522 F.3d 1299 (Fed. Cir. 2008)..................................................................................................6
`
`Ralston Purina Co. v. Far-Mar Co.,
`772 F.2d 1570 (Fed. Cir. 1985)..............................................................................................5, 6
`
`Research Plastics, Inc. v. Fed. Packaging Corp.,
`421 F.3d 1290 (Fed. Cir. 2005)................................................................................................53
`
`Scanner Techs. Corp. v. ICOS Vision Systems Corp.,
`528 F.3d 1365 (Fed. Cir. 2008)..................................................................................................6
`
`Structural Rubber Prods. Co. v. Park Rubber Co.,
`749 F.2d 707 (Fed. Cir.1984)...................................................................................................19
`
`Therasense, Inc. v. Becton, Dickinson & Co.,
`649 F.3d 1276 (Fed. Cir. 2011)........................................................................................ passim
`
`OTHER AUTHORITIES
`
`37 C.F.R. § 1.301...........................................................................................................................47
`
`iv
`
`Page 5 of 68
`
`
`
`TABLE OF FREQUENTLY USED ABBREVIATIONS
`
`ABBREVIATION
`
`MEANING
`
`ID
`CPHB
`CIB
`CRB
`CRRP
`
`RIB
`RRB
`
`Remand Pet.
`JX-1
`JX-2
`JX-3
`JX-4
`GSI
`Cisco
`Avnet
`Cypress / Complainant
`RX
`RDX
`RPX
`RX-354C
`RX-476
`RX-471
`RX-475
`RX-487C
`RX-493C
`RX-495C
`CX
`CDX
`CX-385.1C
`CX-386C
`CX-428C
`CX-429C
`CX-435.1C
`CX-436C
`
`Initial Determination
`Complainant’s Pre-Hearing Brief
`Complainant’s Initial Post-Trial Brief
`Complainant’s Post-Trial Reply Brief
`Cypress Semiconductor Corporation’s Response to Respondents’
`Contingent Petition for Commission Review
`Respondents’ Initial Corrected Post-Trial Brief
`Respondents’ Post-Trial Reply Brief
`Respondents’ Petition For Review Of The Remand Initial
`Determination
`U.S. Patent No. 6,534,805 (the “’805 patent”)
`U.S. Patent No. 6,651,134 (the “’134 patent”)
`U.S. Patent No. 7,142,477 (the “’477 patent”)
`U.S. Patent No. 6,262,937 (the “’937 patent”)
`Respondent GSI Technology, Inc.
`Respondent Cisco Systems, Inc.
`Respondent Avnet, Inc.
`Complainant Cypress Semiconductor Corporation
`Respondents’ Exhibit
`Respondents’ Demonstrative Exhibit
`Respondents’ Physical Exhibit
`Witness Statement of Robert J. Murphy
`U.S. Patent No. 5,386,385
`U.S. Patent No. 5,717,653
`U.S. Patent No. 5,268,865
`Rebuttal Witness Statement of Robert J. Murphy
`Supplemental Direct Witness Statement of Robert J. Murphy
`Supplemental Rebuttal Witness Statement of Robert J. Murphy
`Complainant’s Exhibit
`Complainant’s Demonstrative Exhibit
`Witness Statement of Joseph McAlexander
`Witness Statement of Seth T. Kaplan, Ph.D.
`Rebuttal Witness Statement of Joseph McAlexander
`Rebuttal Witness Statement of Seth T. Kaplan, Ph.D.
`Supplemental Witness Statement of Joseph McAlexander
`Supplemental Rebuttal Witness Statement of Joseph McAlexander
`
`v
`
`Page 6 of 68
`
`
`
`I.
`
`INTRODUCTION
`
`Respondents’ Petition for Review of the Remand Initial Determination on Validity and
`
`Enforceability (“Remand Pet.”) deliberately fails to confront Respondents’ burden of showing
`
`invalidity by clear and convincing evidence. They shirk that burden by ignoring the substantial
`
`differences between their prior art references and Cypress’s asserted patents, by claiming without
`
`support that there are “undisputed limitations,” and finally by misrepresenting what even their
`
`own expert admitted at the hearing.
`
`On the ’134 “burst mode” patent, their prior art simply doesn’t have the key
`
`predetermined and non-interruptible elements. No amount of argument can change that.
`
`On the ’937 two-operations-per-cycle patents, their prior art patents can do only one
`
`operation per cycle; one of their patents is, by Respondents own admission, not enabled; and the
`
`suggested combinations of prior art do not solve these flaws.
`
`On the ’477 overlapping read-write patent, Respondents’ prior art does not overlap. And
`
`when they argue the patent ought not be enforceable, they ignore both the standard of proof set
`
`out by the Federal Circuit, and the fact that the PTO has recently confirmed the validity of the
`
`patent over Respondents’ supposed invalidating reference.
`
`Finally, with the ’805 memory cell layout design patent, Respondents’ prior art lacks any
`
`information about relative proportions or the scale of the allegedly invalidating figures.
`
`Respondents make the creative argument that black letter law should not apply to memory cell
`
`patents, but of course they have no authority for that argument.
`
`It is improbable that every one of the patents asserted by Cypress is invalid, but
`
`Respondents argue stridently that this must be so. They would have a more credible Petition if
`
`they had acknowledged and addressed their clear and convincing burden, and admitted that some
`
`of the patents lack certain elements. Instead, they persist in misrepresenting the record and
`
`1
`
`Page 7 of 68
`
`
`
`grossly ignoring the page limits set by the ALJ and Commission.
`
`When the curtain is pulled back, Respondents have not shown invalidity of any of the
`
`patents by clear and convincing evidence.
`
`II.
`
`THE ’134 PATENT IS VALID
`
`A.
`
`Background
`
`The ’134 patent describes a memory chip with a predetermined and non-interruptible
`
`“burst” mode. “Burst” memory moves data stored at multiple (typically consecutive) addresses
`
`into or out of the memory chip by using an initial address, and then letting the chip generate the
`
`succeeding addresses itself. For example, if the memory chip is asked for address “1,” it can
`
`send addresses 1, 2, 3, and 4 in a burst without making the CPU waste time asking for 2, 3, and 4
`
`individually.1
`
`This is valuable because it frees up the CPU to do other tasks, and in the demanding high-
`
`speed systems where these SRAMs are used—such as Internet switches and mobile phone base
`
`stations—saving even small amounts of time on each memory access will add up to a significant
`
`performance improvement. The ’134 patent also claims the novel memory and associated access
`
`control circuitry as an integrated circuit, i.e., a single chip, and as a chip where its control and
`
`access wires are freed up once the command is given, allowing the processor to queue the next
`
`command while the chip is still reading or writing in response to the earlier command.
`
`Respondents try to gloss over core claim limitations:
`
`1. First, the burst length (the number of chunks of data that the chip will send out)
`
`must be pre-determined, i.e., must be fixed before the burst begins.
`
`1 In the jargon, address 1 is usually referred to as A (the starting address), address 2 as
`A+1, 3 as A+2, and so on.
`
`2
`
`Page 8 of 68
`
`
`
`2. Second, it must be non-interruptible—no signal sent on the wires to the chip can
`
`be used to halt the burst.
`
`3. Third, the device must be an integrated circuit—both the control circuits and the
`
`memory cells must be contained in a single chip.
`
`4. Fourth, once the burst begins, the control wires going to the chip must be freed up
`
`to allow queuing of a new request to the chip.
`
`These are explicit requirements of the patent’s claims. In the Remand Petition,
`
`Respondents have dropped some of their previous prior art references and all of their proposed
`
`combinations of prior art; they now focus only on the Stephens and Takasugi patents. But
`
`neither can satisfy all the claim limitations, hence Respondents cannot show invalidity even by a
`
`preponderance of the evidence—much less by clear and convincing evidence.
`
`The chief arguments are:
`
`(cid:120)
`
`Respondents argue vigorously that there are “undisputed limitations” (Remand Pet.
`
`at 11). Yet this is a chief basis for the ALJ’s scathing criticism of Respondents’
`
`purported “proof” of invalidity, and Respondents are simply wrong that any exist.
`
`These “undisputed limitations” are merely an attempt to shift to Cypress the
`
`Respondents’ burden on invalidity
`
`(cid:120)
`
`The Stephens patent is silent about whether it can be interrupted during a burst.
`
`Respondents ignore that, and boldly argue that Stephens’ burst cannot be halted.
`
`But their citation to “no clock delays” in the specification goes to a wholly different
`
`issue. It is undisputed that the Stephens patent says nothing about whether the
`
`burst is non-interruptible.
`
`3
`
`Page 9 of 68
`
`
`
`(cid:120)
`
`(cid:120)
`
`Respondents argue that dependent claim 2 of the ’134 patent (claiming a fixed burst
`
`length) is met by Stephens. This is nonsensical. Stephens specifically states that
`
`its burst length may be changed.
`
`Respondents argue that dependent claim 14 of the ’134 patent (claiming the device
`
`as an integrated circuit) is met “inherently” merely because there were numerous
`
`integrated circuits in the market when the Stephens device was conceived. That is
`
`not clear and convincing evidence.
`
`(cid:120)
`
`Stephens says nothing about whether the address and control wires running to the
`
`chip are freed up when sending a burst (dependent claim 15). As with claim 14,
`
`Respondents argue that certain of those wires might be freed up. It is telling that
`
`(cid:120)
`
`(cid:120)
`
`they cannot cite anything in the patent for that proposition.
`
`With the Takasugi reference, Respondents once again argue that Cypress agrees
`
`there are “undisputed limitations.” That’s demonstrably incorrect, and again
`
`ignores Respondents’ burden.
`
`Respondents next argue that while the Takasugi device’s burst can be interrupted,
`
`there might be other possible embodiments in which changing a control signal
`
`would not halt the burst. But they cite nothing in the patent for that claim, and their
`
`expert admitted precisely the opposite on cross-examination. Bald attorney
`
`argument is not clear and convincing evidence.
`
`(cid:120)
`
`Respondents suggest that Takasugi anticipates dependent claims 2 (a fixed burst
`
`length—clearly incorrect because the burst can be interrupted at any point); 14 (an
`
`integrated circuit—Respondents have no evidence from the patent for this); and 15
`
`(control wires freed up—impossible because the control wires are used throughout
`
`4
`
`Page 10 of 68
`
`
`
`
`
`the burst to set the dynamic burst length.) None of these arguments proves
`
`invalidity, and certainly not by clear and convincing evidence.
`
`B.
`
`Respondents “Undisputed Limitations” Argument is Wrong
`
`The Remand ID criticized Respondents for repeatedly failing to carry their burden on
`
`validity. They do not cure that deficiency here. As the ALJ correctly held:
`
`Instead of clearly setting forth how each and every limitation of claim 1 is met by
`[a prior art reference], Respondents appear to rely on the fact that certain
`limitations in claim 1 are “undisputed.”…This reliance is inapposite. Cypress
`stated that Stephens does not anticipate the ’134 patent “at least because it does
`not disclose an address generation method that is non-interruptible or where the
`number of addresses to be generated is predetermined with a fixed burst length.”
`(CIB at 42 (emphasis added).) Respondents seem to misinterpret this statement as
`Cypress stating that only one element is absent, which is incorrect. In fact,
`nowhere in the post-hearing briefing does Cypress ever state that any of the
`limitations of claim 1 are present in the prior art. (See generally CIB at 42-44.)
`As Cypress correctly notes, “it is not Cypress’s burden to dispute every claim
`element; rather, it is Respondents’ burden to show by clear and convincing
`evidence that each and every claim element is disclosed in the prior art reference.”
`(CRB at 15 (emphasis original); see also Linear Tech. Corp. v. Int’l Trade
`Comm’n, 566 F.3d 1049, 1066 (Fed. Cir. 2009) (internal citations omitted) (“The
`burden is on the party asserting invalidity to prove it with facts supported by clear
`and convincing evidence.)”
`
`RID at 9-10 (emphasis added).2
`
`Respondents were required to prove that all of the claim elements are met by one prior art
`
`reference to show anticipation. Minn. Mining & Mfg. Co. v. Chemque, Inc., 303 F.3d 1294, 1306
`
`(Fed. Cir. 2002). They did not, and cannot, do so. Instead, despite a litany of arguments about
`
`what they alleged in their pre-trial brief, at trial, and in their post-trial briefs, they ultimately
`
`concede that they rely on their “undisputed limitations.” That isn’t sufficient.3
`
`2 The ALJ found that Respondents used the same flawed strategy throughout their papers.
`See RID at 9-10; 12-13; 15; 18; 23-24; 28-29.
`3 Respondents’ case citations don’t salvage this argument. In Ralston Purina Co. v. Far-Mar
`Co., 772 F.2d 1570, 1573-74 (Fed. Cir. 1985), the Federal Circuit reiterated that “Before
`rendering its judgment, the court must determine whether ‘all of the evidence establishes that the
`validity challenger so carried his burden as to have persuaded the decisionmaker that the patent
`
`5
`
`Page 11 of 68
`
`
`
`It was not Cypress’s burden to dispute every claim element; rather, it was Respondents’
`
`burden to show by clear and convincing evidence that each claim element is disclosed in the
`
`prior art reference. When an alleged infringer attacks the validity of an issued patent, the
`
`burden of persuasion is on the attacker to prove invalidity by clear and convincing evidence. See
`
`Hybritech Inc. v. Monoclonal Antibodies, Inc., 802 F.2d 1367, 1375 (Fed. Cir. 1986); Ralston
`
`Purina, 772 at 1573. Respondents have not done so. Their unsupported assertions that the other
`
`claim elements are met fall short. Cambridge Prods., Ltd. v. Penn Nutrients, Inc., 962 F.2d
`
`1048, 1051 (Fed. Cir. 1992) (finding defendant’s “bare assertion” insufficient to meet clear and
`
`convincing test).
`
`The ALJ correctly noted that even when Respondents attempted to show a claim element
`
`was present in a prior art reference,
`
`[M]uch of Respondents’ “proof” consists of conclusory statements with string cites to
`their expert’s testimony…This is, quite simply, nothing more than an improper attempt
`to circumvent the page limitations for post-hearing briefs. Merely citing the testimony
`and demonstrative exhibits of a party’s expert or portions of a prior art patent without any
`
`can no longer be accepted as valid.’” (Emphasis in original; citation omitted). Similarly, the
`court in PowerOasis, Inc. v. T-Mobile USA, Inc., 522 F.3d 1299, 1305 (Fed. Cir. 2008) noted that
`“[T]he party asserting invalidity, must still show by clear and convincing evidence that the
`asserted patent is invalid.” Moreover, in that case, the defendant admitted that its system
`practiced the prior art patent, and the sole question was the priority date of the patents at issue.
`Cypress makes no such admission.
`
`Finally, in Scanner Techs. Corp. v. ICOS Vision Systems Corp., 528 F.3d 1365, 1380 (Fed. Cir.
`2008), a prima facie case of invalidity was shown because “[I]n this case, the record shows that
`ICOS adduced certain facts, which were unrebutted by Scanner” (emphasis added). Even there,
`that court noted that, as always, “[I]t is ICOS’s burden to prove invalidity by clear and
`convincing evidence, and that that burden of proof never shifts to the patentee to prove
`validity” (emphasis added).
`
`If Respondents had proven by clear and convincing evidence that every claim limitation was
`present in a prior art reference, and if Cypress has not disputed any of that showing, Respondents
`might have been able to make a prima facie case of invalidity.
`
`But that isn’t what happened here. Rather, Respondents attempt to avoid confronting their
`burden by arguing that Cypress has not disputed that certain limitations are practiced. This is
`both wrong and a tactic disallowed by the very cases they cite.
`
`6
`
`Page 12 of 68
`
`
`
`explanation not only fails to constitute “a discussion” of the issue in the post-hearing
`brief as required by the Ground Rules, but is insufficient to prove that a patent is invalid
`by clear and convincing evidence.
`(RID at 10, emphasis added).4
`
`Respondents’ citation to the Commission’s holding in Certain Automated Media Library
`
`Devices, Inv. No. 337-TA-746, Comm’n Op. (Pub. Version) at 51-56 (Nov. 19, 2012) does not
`
`improve their position. There, the Commission noted that the evidence “was not contested
`
`before the ALJ” (emphasis added). The opposite is true here: Cypress vigorously contests
`
`Respondents’ claims that any of its patents’ claim limitations are present in the prior art.
`
`Respondents were required, as the ALJ’s Ground Rules made clear, to provide a discussion and
`
`explanation of their arguments, not to merely string cite to expert testimony and demonstratives.
`
`Respondents’ intimation that Cypress admits certain claim elements are present is flat-out
`
`wrong. See CPHB, at 47-52 (explaining elements are missing in the prior art references); CIB, at
`
`42 and 44 (arguing, respectively, that the Stephens patent and Takasugi patents do not anticipate
`
`the ’134 patent at least because they do not disclose a number of the claim elements). Nowhere
`
`did Cypress state that any claim elements were present in the prior art.
`
`Those arguments about the supposed “admissions” and “undisputed limitations” are
`
`demonstrably inaccurate. With the Stephens patent (U.S. Patent No. 5,386,385; RX-476), for
`
`example, Cypress pointed out that the number of addresses to be created cannot be pre-
`
`determined because Stephens uses a “mode register.” CX-428C Q/A 57, 61. The “mode
`
`register” tells the device how many “chunks” of data to send, and that number may be changed at
`
`any time. By definition, this is not “fixed.” The Stephens part receives a “Read” or “Write”
`
`4 Respondents’ protestations that they are not trying to circumvent the page limits ring
`hollow. They used 99 of their allotted 100 pages in the Remand Petition, and then also
`submitted 189 pages of additional material regarding the ’805 patent (discussed in more detail
`below).
`
`7
`
`Page 13 of 68
`
`
`
`command at the same time that the mode register is checked. In other words, the burst length
`
`can vary up to and including the very moment that the circuit begins reading or writing. Id. at
`
`Q/A 61; CDX-593C, RX-476.5
`
`Similarly, with the Takasugi patent, a wire called “ CAS” controls how many chunks of
`
`data will be sent out. In Takasugi’s device, the burst continues as long as CAS stays low (i.e., as
`
`long as no voltage is sent down the CAS wire). As soon as CAS goes high (has voltage running
`
`down that wire to the chip), the chip outputs a final piece of data and stops the burst. Murphy,
`
`Tr. 583:11-25. That is not “predetermined.”
`
`C.
`
`The Stephens Device’s Burst May Be Interrupted
`
`Next, Respondents attempt to confuse the evidence about whether the Stephens part can
`
`meet the “non-interruptible” requirement. They argue at length that two passages in the
`
`specification (RX-476 at col. 5 ll.34-35; col. 8 ll.44-46) equate to non-interruptible. Remand Pet.
`
`at 14. But it is illuminating to focus on precisely what those passages say:
`o “In such a burst, each byte follows the last byte with no clock delays in
`between”;
`o “Counting continues in synchronism with the internal clock frequency
`until the burst operation is completed”
`
`First, had Stephens intended that his device be non-interruptible, he would have said so.
`
`He didn’t. Rather, he made it clear that if left alone, the part will output a chunk of data on each
`
`“tick” of the system’s clock, and that it will not, for example, delay the output (“no clock
`
`delays”) or send the data out at a non-clock frequency (“in synchronism with”). These are
`
`important facts to the designer of a larger system—the processor can expect to receive data from
`
`5 Respondents argue that Stephens nevertheless anticipates because during a burst
`operation, the burst length is fixed. But whether the burst length is fixed during the burst
`operation is irrelevant to the ’134 patent. The ’134 patent discloses a burst that is pre-
`determined—which by definition is one that is fixed before the burst begins.
`
`8
`
`Page 14 of 68
`
`
`
`the part on every beat of the clock, and precisely in sync with the system clock. But that says
`
`absolutely nothing about whether the burst can be stopped.
`
`Respondents’ expert first tries to argue that “
`
`
`
`
`
`e
`
`
`
`That is no more than speculation.
`
`The parties agree that Stephens is silent about whether a burst can be interrupted. This is
`
`a critical limitation of the independent claim in the’134 patent, and speculation about how a prior
`
`art device might be built does not meet the clear and convincing burden.
`
`D.
`
`Stephens Cannot Meet the Limitations of Dependent Claims 2, 14 or 15
`
`Claim 2 of the ’134 patent includes the independent claim’s requirements that the burst
`
`length be predetermined and non-interruptible, and adds the requirement that the device have a
`
`fixed burst length. Claim 14 adds that the device must be an integrated circuit, i.e., a single chip.
`
`Claim 15 adds that the device must free up its address and control wires while it is sending data
`
`so that another request may be queued. All of these are explicit requirements in the ’134 patent.
`
`None is present in Stephens.
`
`First, Stephens does not have claim 2’s “fixed burst length.” Instead, as noted above,
`
`Stephens uses a “mode register” to set the burst length: when a read command is sent, the device
`
`checks the mode register to see how many chunks of data it is being asked to send in that
`
`particular burst. The mode register can be changed up to the very moment that the device begins
`
`9
`
`Page 15 of 68
`
`
`
`reading data. By definition, the mode register does not and cannot give the Stephens patent’s
`
`part a fixed burst length. CX-428C Q/A 62.6
`
`Second, Stephens nowhere says that its device is an integrated circuit, as required by
`
`claim 14. Respondents’ argument is that the Stephens invention might have “inherently” been a
`
`single chip, because integrated circuits were in existence at the time. Remand Pet. at 16. This is
`
`mere supposition by Respondents’ expert. It is undisputed that nowhere does Stephens state that
`
`the device is—or even could be—an integrated circuit. This is not clear and convincing.
`
`Finally, Stephens is similarly silent about whether it can free up its control wires while it
`
`is sending data so another request can be queued. This is an important and explicit requirement
`
`of claim 15, speeding up the part by allowing the processor to stack requests to the memory chip.
`
`Respondents’ expert claims that while the Stephens part is running, the control lines are not in
`
`use. But that, too, is mere supposition. As Cypress’s expert noted, the patent nowhere says that.
`
`CX-428C Q/A 64. Speculation by Respondents’ expert does not meet their burden of clear and
`
`convincing evidence.
`
`In short, Respondents have not shown that Stephens has a pre-determined burst length,
`
`nor that it cannot be interrupted, nor that it is fixed, nor that it is a single chip, nor that it frees up
`
`its control and address lines to queue another command.
`
`E.
`
`Respondents Also Rely on “Undisputed Limitations” with Takasugi
`
`Respondents’ attempt with the Takasugi patent fares no better. As with Stephens,
`
`Respondents argue that there are limitations Cypress has not disputed. And once again, that is
`
`nothing more than an effort to reverse the burden of proof.
`
`6 Respondents argue, as above, that while the burst is in progress, its length is fixed. But that
`ignores the independent claim’s mandate that the length be “pre-determined”—i.e., determined
`before the burst begins. see CX-428C Q/A 61.
`
`10
`
`Page 16 of 68
`
`
`
`It was Respondents’ burden to show every limitation was present. They failed to do so,
`
`and they are unable to prove that the Takasugi device has a pre-determined burst length because
`
`of the way the Takasugi invention operates. In Takasugi, a control line called “ CAS” creates the
`
`burst. As long as it is kept at zero, the burst continues. When CAS is changed from zero to one
`
`by the CPU, the burst stops. RX-475 at col. 5, ll 35-50; CX-428C Q/A 72-73; CDX-595C;
`
`CDX-630C.
`
`Because the control line is used throughout the burst, and can be held low for as long as
`
`the CPU wishes, there is ipso facto no pre-determined burst length. Respondents suggest this is
`
`an “undisputed limitation,” but that is