`____________________________________________
`
`BEFORE THE PATENT TRIAL AND APPEAL BOARD
`____________________________________________
`
`Intel Corporation
`Petitioner
`
`v.
`
`Qualcomm Incorporated
`Patent Owner
`
`U.S. Patent No. 8,698,558
`Claims 15-20
`____________________________________________
`
`Case IPR2018-01154
`____________________________________________
`
`REPLY DECLARATION OF ALYSSA APSEL, PH.D.
`ON BEHALF OF PETITIONER
`
`Intel v. Qualcomm
`Exhibit 1228
`IPR2018-01154
`
`
`
`2.
`
`B.
`
`I.
`II.
`
`TABLE OF CONTENTS
`BACKGROUND ............................................................................................. 1
`CLAIM CONSTRUCTION ............................................................................ 2
`A.
`Patent Owner’s Proposed Construction Is Wrong ................................ 2
`1.
`Patent Owner’s Proposed Construction Contradicts The
`Plain Claim Language ................................................................. 2
`Patent Owner’s Proposed Construction Would Exclude
`Disclosed Embodiments .............................................................. 5
`Patent Owner’s Remaining Arguments Have No Merit ............. 7
`3.
`III. GROUNDS ...................................................................................................... 8
`A.
`Claim 15 is anticipated by Kwak .......................................................... 8
`1.
`Kwak’s feedforward path increases the inductor current ........... 9
`Petitioner Has Not Improperly Combined Two Different Kwak
`Embodiments ....................................................................................... 23
`1.
`Figure 5 Alone Anticipates Claim 15 ....................................... 23
`2.
`Figure 6 Shows A Detailed Implementation of a Figure 5 ....... 26
`Claim 16 Is Unpatentable .................................................................... 27
`Patent Owner Is Wrong That Petition Fails To Demonstrate a
`Motivation to Combine Kwak with Choi 2010 ................................... 28
`1.
`Patent Owner Misrepresents Petitioner’s Argument ................ 29
`IV. AVAILABILITY FOR CROSS-EXAMINATION ...................................... 30
`V.
`RIGHT TO SUPPLEMENT .......................................................................... 30
`VI.
`JURAT ........................................................................................................... 31
`
`
`C.
`D.
`
`i
`
`
`
`I, Alyssa Apsel, declare as follows:
`
`I.
`
`BACKGROUND
`1.
`I am the same Alyssa Apsel who submitted a prior declaration in this
`
`matter, which I understand was filed on June 28, 2018. I am currently the Director
`
`of the School of Electrical and Computer Engineering and a professor of electrical
`
`and computer engineering at Cornell University in Ithaca, New York. Between
`
`September 2016 to June 2018, I was a visiting professor at Imperial College in
`
`London, England, where I worked on low power RF interfaces for implantable
`
`electronics. My background and qualifications remain as stated in paragraphs 2-14
`
`and Appendix A of that declaration, filed as Exhibit 1203 in this case. My
`
`statements in paragraphs 17-19 of my prior declaration regarding my review of
`
`U.S. Patent No. 8,698,558 (“the ’558 patent”) and related materials also remain
`
`unchanged, as do my understandings of the relevant legal principles stated in
`
`paragraphs 20-31.
`
`2.
`
`Since my prior declaration, I have reviewed Patent Owner’s
`
`Preliminary Response of October 17, 2018 (“POPR”), the Board’s Decision to
`
`Institute of February 6, 2019, the transcript of my deposition taken on March 6,
`
`2019, the Patent Owner’s Response of April 15, 2019 (“POR”), the Declaration of
`
`Arthur W. Kelley of April 15, 2019 (Ex. 2002), the transcript of Dr. Kelley’s
`
`1
`
`
`
`deposition taken on June 21, 2019 (Ex. 1229), and the related district court
`
`litigation claim construction order (Ex. 1227).
`
`3.
`
`I confirm that everything included in my prior declaration of June 28,
`
`2018, and all of the testimony given during my deposition of March 6, 2019,
`
`remain true to the best of my knowledge.
`
`II. CLAIM CONSTRUCTION
`A.
`Patent Owner’s Proposed Construction Is Wrong
`4.
`Patent Owner contends that the term of claim 19 “the envelope
`
`amplifier operates based on the first supply voltage or the boosted supply voltage”
`
`should be construed such that “the envelope amplifier must be able to operate,
`
`selectively, based on either the first supply voltage or the boosted supply voltage
`
`(referred to herein as a ‘selective boost’).” (POR, 11, 35.) In other words, under
`
`Patent Owner’s construction, an amplifier that received only the first voltage or
`
`only the boosted voltage would not satisfy this limitation. I have been informed
`
`and understand that this proposed construction is far from the broadest reasonable
`
`construction of “or,” is contrary to the plain meaning, and excludes disclosed
`
`embodiments, and, therefore, it should be rejected.
`
`1. Patent Owner’s Proposed Construction Contradicts The Plain
`Claim Language
`Claim 19 recites an “envelope amplifier” that “operates based on the
`
`5.
`
`first supply voltage or the boosted supply voltage.” Ex. 1201, 14:25-27. As Dr.
`
`2
`
`
`
`Kelley conceded at his deposition, the term “or” is a conjunction that identifies two
`
`alternatives: this “or” that. (Ex. 1229, 130:10-18 (“Q. I’m asking at the
`
`Schoolhouse Rock level, or is a conjunction that joins two alternatives, correct? A.
`
`Well, if we’re going to import Schoolhouse Rock into the deposition, in that
`
`context, yes, it is.”).) Under its plain meaning, the requirement for an amplifier
`
`that operates based on “the first supply voltage or the boosted supply voltage” is
`
`met by an amplifier that operates based on either one of those alternatives alone.
`
`(Id. at 130:19-131:2 (“Q.…If I said I would like coffee or tea, you could give me
`
`tea and that would meet my requirement, right? A. In that hypothetical abstract
`
`outside the bounds of the ’558, sure.”). Patent Owner has identified no sound basis
`
`to deviate from that broad plain meaning.
`
`6.
`
`To the contrary, Patent Owner concedes that the common meaning of
`
`“or” in patent claims is to recite alternatives. See, POR, 39 (“The use of ‘or’ is
`
`sometimes an acceptable mechanism for claiming alternatives such that only one of
`
`the limitations need be found in the prior art to support anticipation.” I have been
`
`informed and understand that this is exactly how Hon. Dana M. Sabraw construed
`
`“or” in the related district court litigation on the ’558 patent. (Ex. 1227 [Claim
`
`Constr. Order] at 5-6 (holding the limitation “a source receiving the boosted supply
`
`voltage or the first supply voltage” in claim 6 does not require “selective boost”).)
`
`3
`
`
`
`Indeed, Dr. Kelley admitted at his deposition that Patent Owner’s construction
`
`contradicts Judge Sabraw:
`
`Q. And just to be clear, you’re giving an opinion that is contrary to
`Judge Sabraw’s claim construction, right?
`I understand what the Judge did. And I’ve reached a different
`conclusion.
`
`A.
`
`Ex. 1229, 147:10-151. I have been informed and understand that Judge Sabraw
`
`applied the Phillips standard, which is a narrower standard than the broadest
`
`reasonable interpretation standard. But if “or” covers either alternative alone under
`
`the Phillips standard that Judge Sabraw used, I understand that it is certainly at
`
`least that broad under the broadest reasonable construction rule, which is
`
`applicable in this proceeding. Accordingly, an envelope amplifier that operates
`
`based on either one of the recited alternatives – e.g., the “boosted supply voltage”
`
`alone – would meet the claim.
`
`7. Moreover, other claims demonstrate that Patent Owner knew how to
`
`recite “selective boost” when it wanted – using language different from claim 19.
`
`Dependent claim 7 provides one such example. Claim 7 depends from
`
`independent claim 6. Much like claim 19, claim 6 recites “a source receiving the
`
`boosted supply voltage or the first supply voltage….” Claims 6 is therefore met
`
`
`1 Objections have been omitted in transcript quotations.
`
`4
`
`
`
`using only one of the boosted or first supply voltage. Indeed, as noted above,
`
`Judge Sabraw found that this claim does not require “selective boost.” (Ex. 1227,
`
`5-6.) By contrast, dependent claim 7 adds the limitation “wherein the supply
`
`generator is operative to generate the second supply voltage based on the envelope
`
`signal and either the boosted supply voltage or the first supply voltage.” If claim
`
`6 already required “selective boost,” then claim 7 would add nothing to claim 6.
`
`But the additional language “and either” of claim 7 has meaning. For that reason,
`
`claim 7 could not be read to require only one of the boosted or first supply voltage.
`
`It is therefore that additional language appearing in claim 7 (which does not appear
`
`in either claims 6 or 19) that adds a requirement for the amplifier to be able to
`
`select between “either” the boosted voltage “or” the first supply voltage. Because
`
`claim 19 (like claim 6) lacks this “and either” language, it has no such requirement.
`
`2. Patent Owner’s Proposed Construction Would Exclude
`Disclosed Embodiments
`I have been informed and understand that Patent Owner’s construction
`
`8.
`
`should also be rejected, because it improperly excludes disclosed embodiments.
`
`9.
`
`Specifically, the ’558 specification (starting at column 8, line 24)
`
`teaches “another design of supporting operation with a lower battery voltage[.]”
`
`(Ex. 1201, 8:24-25.) In this embodiment, “the entire envelope tracker is operated
`
`based on the Vboost voltage from boost converter 180” alone (i.e., solely based on
`
`“boosted supply voltage”), without ever operating based on battery voltage (i.e.,
`
`5
`
`
`
`never based on “first supply voltage”). (Ex. 1201, 8:25-26.) As Dr. Kelley
`
`admitted at his deposition when asked about claim 13, which recites language
`
`similar to claim 19, Patent Owner’s construction would exclude this disclosed
`
`embodiment:
`
`Q.
`
`If you’re right that the selective boost and the or means I have
`
`to be able to use either boost or first, then under that
`
`circumstance, claim [6] and 13 would not cover the
`
`embodiment at column 8, line 24 that uses Vboost alone. Is
`
`that fair?
`
`A.
`
`I think that’s fair.
`
`(Ex. 1229, 134:12-18.) Because Patent Owner’s proposed construction
`
`would exclude this disclosed embodiment, I have been informed and
`
`understand that it should be rejected.
`
`10. Patent Owner’s argument demonstrates that it is improperly
`
`attempting to limit claim 19 solely to one disclosed embodiment (to the exclusion
`
`of others). For example, Patent Owner contends that column 1, lines 42-50
`
`discloses a form of “selective boost.” Whether the specification here means what
`
`Patent Owner says is debatable. But what is beyond dispute is that column 1
`
`discloses merely “one design” as an “example” that in no way limits the claims:
`
`6
`
`
`
`In one design, the envelope amplifier may further receive the first
`supply voltage and may generate the second supply voltage based on
`either the first supply voltage or the boosted supply voltage. For
`example, the envelope amplifier may generate the second supply
`voltage (i) based on the boosted supply voltage if the envelope signal
`exceeds a first threshold and/or if the first supply voltage is below a
`second threshold or (ii) based on the first supply voltage otherwise.
`
`(Ex. 1201, 1:42-50. (emphasis added).) Patent Owner’s citation to column 8, lines
`
`55-62 is similarly misplaced. (Id., 8:55-62 (“In one design….For example,…”).) I
`
`have been informed and understand that limiting claims solely to one embodiment
`
`disclosed in the specification in not appropriate. Accordingly, Patent Owner’s
`
`attempt to do so here should be rejected.
`
`3. Patent Owner’s Remaining Arguments Have No Merit
`11. Finally, Patent Owner suggests that I somehow “agreed that it would
`
`not make sense to interpret the boosted supply voltage as purely optional in the
`
`context of the claims” (POR, 40), because “when asked about a similar ‘based on’
`
`limitation in claim 6,” I supposedly “admitted that it makes sense to interpret
`
`‘based on’ as requiring a boosted supply voltage to be an available supply
`
`voltage.” POR, 40.
`
`12. The language from claim 6 that I was asked to address in this
`
`testimony recites “generat[ing] a second supply voltage for the power amplifier
`
`based on the envelope signal and the boosted supply voltage.” (Ex. 2003 [Apsel
`
`7
`
`
`
`Transcript], 41:21-23; Ex. 1201, 11:49-51.) This language is not similar to claim
`
`19, which plainly recites “or.”2 (Ex. 1201, claim 19.) By reciting “or,” claim 19 is
`
`clear that either one of the “boosted supply voltage” or “the first supply voltage” is
`
`alone sufficient to satisfy the claim’s requirement.
`
`13. Patent Owner’s proposed construction should thus be rejected.
`
`III. GROUNDS
`A. Claim 15 is anticipated by Kwak
`14. As I explained in my first Declaration, claim 15 is anticipated by
`
`Kwak. Ex. 1203, ¶¶87-109.
`
`15. Patent Owner raises only two arguments for why claim 15 is allegedly
`
`not anticipated. First, Patent Owner argues that Kwak’s feedforward path does not
`
`increase the inductor current. POR, 24-33. Second, Patent Owner argues that
`
`Petitioner improperly combines two different embodiments of Kwak. POR, 17-24.
`
`As I explain below, both arguments are incorrect.
`
`
`2 This language is also different from the relevant part of claim 6 that Judge
`
`Sabraw interpreted, which recites “a source receiving the boosted supply voltage
`
`or the first supply voltage.” (Ex. 1201, 11:58-59.)
`
`8
`
`
`
`1. Kwak’s feedforward path increases the inductor current
`16. Kwak discloses switching amplifiers in Figures 3(a) and 5. Kwak’s
`
`Figure 5 modifies Figure 3(a) by adding a feedforward path. Ex. 1211, 2668-2669.
`
`In both Figures 3(a) and 5, the output current (io) equals the sum of the linear
`
`amplifier current (ia) and the inductor current (id), i.e., io= ia+id. As Patent Owner
`
`has conceded, the use of the feedforward path does not change the output current,
`
`io. POR, 27. Patent Owner has also conceded that it was known in the art that
`
`switchers are more efficient than linear amplifiers, and that it was therefore known
`
`to be desirable to maximize the inductor current and reduce the linear amplifier
`
`current. Ex. 1229, f-186:10.
`
`17. Kwak explains that, in theory “the linear amplifier only delivers the
`
`switching ripple current of the switching amplifier because the switching amplifier
`
`supplies most of the output current through the relation of io=ia+id =(1+β)‧ia.” Ex.
`
`1211, 2666. But as Kwak recognizes, “[i]n reality, however, [] the output current
`
`of the switching amplifier (id) is slower and less than the output current (io) because
`
`of the finite loop gain β. Thus, the linear amplifier must provide some amount of
`
`signal current in addition to the ripple current to compensate for the distortion that
`
`results from the phase lag of the switching stage in the high-frequency region.” Ex.
`
`1211, 2666. Kwak thus states that the linear amplifier does not merely provide
`
`ripple cancelation current–it also provides “some amount of the signal current.”
`
`9
`
`
`
`18. Kwak’s feedforward path shown in Figure 5 has a purpose to reduce
`
`the signal current from the linear amplifier. See, Ex. 1211, 2668 (“Hence, an
`
`auxiliary circuit [feedforward path] is necessary to alleviate the burden of the
`
`linear amplifier.”) Use of the feedforward path reduces the burden on the linear
`
`amplifier (i.e., decreases ia) . According to Kwak’s equation (io=ia+id), the only
`
`way that total current (io) can remain constant when linear amplifier current (ia) is
`
`decreased is if the inductor current (id) increases. Indeed, that is the point of
`
`Kwak’s feedforward path––to reduce the burden on the linear amplifier (i.e.,
`
`reduce ia) by increasing the inductor current (i.e., increase id).
`
`19. This interpretation of Kwak is clear and Patent Owner cannot escape
`
`from it. Consequently, Patent Owner relies on misleading diagrams from Dr.
`
`Kelley, on a flawed visual inspection of Kwak’s Figure 11, and on
`
`mischaracterization of Kwak’s equation 4.
`
`(a) Diagrams in Dr. Kelley’s Declaration Do Not Depict
`Kwak
`In his declaration, Dr. Kelley presents three graphs to support his
`
`20.
`
`argument that decreasing the linear amplifier current (ia) does not increase the
`
`inductor current (id). Ex. 2002, ¶¶77-80 (Examples 1-3); POR, 28-29. As I explain
`
`below, Dr. Kelley relies on improper constraints to generate these graphs –
`
`constraints that are contrary to Kwak’s teachings. In fact, at his deposition, Dr.
`
`10
`
`
`
`Kelley conceded that those graphs do not represent Kwak and instead merely
`
`represent mathematical addition of sine waves.
`
`Q. So the assumption that the amplitude, the peak to peak
`amplitude of Id equals Io, Kwak at Page 2673 bottom of the left
`column says that it is not correct, right?
`A
`Again, my waveforms are an illustration of how to do
`math with sine waves. They are not meant to directly reproduce
`Figure 11 of Kwak.”
`
`Ex. 1229, 232:19-233:5.
`
`21. Because Dr. Kelley admits that his graphs do not represent Kwak, this
`
`argument from Patent Owner is entitled to no weight.
`
`(b) Patent Owner Misinterprets Kwak’s Figure
`22. Patent Owner argues that “a comparison between Figures 11(a) and
`
`11(b) reveals that the feedforward path causes a phase shift in supply current (id)
`
`[], but no increased magnitude [].” POR, 25. Specifically, Patent Owner argues
`
`that even though the addition of the feedforward path in Figure 11(b) results in a
`
`decrease in the linear amplifier current, Figure 11(b) does not show any increase in
`
`the inductor current. See, id., 32. Patent Owner is wrong. In particular, this
`
`argument is based solely on a visual inspection of Figures 11(a) and (b). But those
`
`figures are not properly scaled to show changes in the inductor current, as Dr.
`
`Kelley conceded. Ex. 1229, 204:1-205:3.
`
`11
`
`
`
`23. First, Kwak’s Figure 11 uses different scales to represent the
`
`waveforms of the linear amplifier current, ia, and the inductor current, id. That is,
`
`Figure 11 uses a step-size of 200mA for the linear amplifier current, ia, but uses a
`
`step-size of 500mA for the switcher current, id. These scales are clearly shown in
`
`Kwak’s Figure 11.
`
`Ex. 2011, Fig. 11
`
`
`
`24. Therefore, in Figure 11, a change of one division for the inductor
`
`current, id (i.e., a change from one dashed horizontal line to the next) represents a
`
`change that is 250% greater than a change of one division for the linear amplifier
`
`current, ia. Because of these different step sizes, it is much easier to see small
`
`changes in the linear amplifier current, ia, than it is to see the same small change in
`
`12
`
`
`
`the inductor current, id. Dr. Kelley admitted this at his deposition. Dr. Kelley
`
`admitted that “[c]ertainly at this particular resolution you wouldn’t see a 1 percent
`
`change.” Ex. 1229, 205:1-3. See also, id., 204:1-205:3.
`
`25. Second, any decrease in the linear amplifier current, ia, caused by
`
`Kwak’s feedforward path is balanced by an identical increase in the inductor
`
`current, id. This is so because, as Patent Owner concedes, use of the feedforward
`
`path does not change the output current, io. POR, 27 (“The output current (io)
`
`remains constant”). Therefore, because io=ia+id, and because io remains
`
`unchanged, if ia decreases, id must increase by the identical amount. Because of
`
`the different scales used in Figure 11, the decrease in linear amplifier current, ia, is
`
`more visible than the identical increase in inductor current, id. But Kwak’s
`
`equation confirms that inductor current id must increase.
`
`26. Third, a person of ordinary skill in the art (“POSA”) would have
`
`understood that Kwak’s inductor current, id, was much larger than the linear
`
`amplifier current, ia, and indeed this is what Figure 11 shows. This is so because,
`
`as Patent Owner concedes, switchers are more efficient than linear amplifiers and it
`
`was known in the art to maximize the contribution of the switcher. For example,
`
`the inductor may provide 80% or more of the total current, where the linear
`
`amplifier may provide only 20% or less. Therefore, any decrease in the linear
`
`amplifier current, ia, caused by Kwak’s feedforward path will be much easier
`
`13
`
`
`
`visible in Figure 11 than the corresponding increase in the switcher current, id. For
`
`example, using the 80-20% example, suppose the switcher current, id, is 80mA and
`
`the linear amplifier current, ia, is 20mA without the feedforward path. Suppose
`
`further that activating Kwak’s feedforward path decreases the linear amplifier by
`
`2mA. That would be a 10% reduction in linear amplifier current, ia. However, the
`
`corresponding 2mA increase in switcher current, id, would be a 2.5% increase,
`
`which would be harder to perceive than the 10% decrease in ia.
`
`27. Notwithstanding these differences in the scales, close examination of
`
`Figure 11 leads to the opposite of Patent Owner’s conclusion. Patent Owner
`
`provides an annotated Figure 11 (reproduced below) and argues that “[a]s can be
`
`seen by comparing the vertical red lines to the vertical blue lines, the phase lag of
`
`(id) is noticeably reduced in Figure 11(b) with the feedforward path compared to
`
`the phase lag in Figure 11(a) without the feedforward path. Meanwhile, the
`
`magnitude of supply current (id) remains the same in both Figures 11(a) and 11(b),
`
`as denoted by the horizontal blue lines.” POR, 31 (internal citations omitted.)
`
`14
`
`
`
`
`
`POR, 32 (Fig. 11 as annotated by Patent Owner)
`
`28. Patent Owner misrepresents Kwak’s Figure 11. Patent Owner argues
`
`that because the maximum value of inductor current id (indicated by the blue
`
`horizontal line) is the same in both figures, the magnitude of the inductor current
`
`remains the same. However, as I explain below, while the magnitude of the
`
`highest peaks in the inductor current, id, are the same in Figures 11(a) and (b), the
`
`valleys, or lowest peaks, of id are not. Because the valley of id is lower in Figure
`
`11(b) than it is in Figure 11(a), and because the highest peaks are the same, the
`
`peak-to-peak magnitude of the inductor current, id, is larger in Figure 11(b) than in
`
`Figure 11(a).
`
`15
`
`
`
`29. As Kwak explains, the magnitude of the AC component of the current
`
`is measured by its peak-to-peak value. Ex. 2011, 2673 (“Without the feedforward
`
`path, as shown in Fig. 11(a), the output current of the linear amplifier is
`
`approximately 120mAPP. In contrast, as shown in Fig. 11(b), this value is reduced
`
`to about half when the feedforward path is used.”) The “PP” subscript in Kwak
`
`corresponds to “peak-to-peak.” Patent Owner admits “that the magnitude and
`
`phase of output (Vo) remain constant with and without the feedforward path.”
`
`Accordingly, the minimum value of signal Vo is at the same point in both figures,
`
`as Dr. Kelley testified: “Q. But the minimum values of Vo are the same in 11(a)
`
`and 11(b), right? A. I understand that they’re the same.” Ex. 1229, 207:9-13.
`
`30. This is illustrated in the excerpted detail from Patent Owner’s
`
`illustration below, in which the orange lines identify the valleys of the output
`
`voltage, Vo, and the green lines identify the valleys of the inductor current, id.
`
`16
`
`
`
`
`
`POR, 32 (Fig. 11, excerpted and annotated)
`
`31. According to Patent Owner and Dr. Kelley, the orange horizontal lines
`
`are at the same points in Figures 11(a) and 11(b) (The figures are not perfectly
`
`aligned in the page. Fig. 11(b) is slightly higher). POR, 31; Ex. 2002, ¶87. In the
`
`illustration above, the distance between the orange and green lines is smaller in
`
`Figure 11(b) than in Figure 11(a). That is, the distance between the valleys of the
`
`output voltage, Vo, and the inductor current, id, is bigger in Figure 11(a) than
`
`Figure 11(b). Indeed, Dr. Kelley admitted that the inductor current id valley
`
`(green) is closer to the Vo valley (orange) in Figure 11(b) than it is in 11(a),
`
`thereby conceding that the peak-to-peak distance in Figure 11(b) has increased:
`
`17
`
`
`
`Q. …Is it true, yes or no, that the vertical distance between
`the lowest point of Vo and the lowest point of Id in (a) is greater than
`the vertical distance between the lowest point of Vo and the lowest
`point of Id in (b)?
`A. Yes. That’s what produces the current Ia.
`
`Ex. 1229, 217:12-19.3 This confirms that the peak-to-peak value of inductor
`
`current id has increased in Fig. 11(b), because the valley, or lowest peak, of id
`
`(green line) has decreased, while the highest peak of id (blue line) remained the
`
`same.
`
`32. Accordingly, even Patent Owner’s illustration confirms that Kwak’s
`
`feedforward path increases the inductor current.
`
`(c)
`
`Patent Owner’s Argument That Kwak’s Feedforward
`Path Affects Only The Phase Of The Inductor
`Current Is Wrong
`33. Patent Owner argues that Kwak’s feedforward path changes the phase
`
`of the inductor current, but not its magnitude. POR, 29. As explained below,
`
`Patent Owner is mistaken.
`
`
`3 In redirect, when asked about the peak-to-peak magnitude of inductor current id,
`
`Dr. Kelley contradicted himself by arguing there is no change in the peak-to-peak
`
`magnitude.
`
`18
`
`
`
`34.
`
`In Kwak’s Figures 3(a) and 5, the output current, io, equals the sum of
`
`the linear amplifier current, ia, and the inductor current, id. Ex. 1211, 2668-2669.
`
`Kwak’s equation 4 shows this and further shows that the inductor current, id, has
`
`two components: (1) the sensed current from current sensor AS, “ASꞏiaꞏC,” and (2)
`
`the offset current, “AFꞏvinꞏC,” from the feedforward path AF, where “C=AIꞏAM
`
`ꞏ(1/(ZL+sL)).” Id., 2669. Therefore, the output current depends on the linear
`
`amplifier current (ia), the sensed current (ASꞏia), and the offset current (AFꞏvin).
`
`35. As Patent Owner concedes, use of the feedforward path does not
`
`change the output current, io. POR, 27. As Kwak states explicitly, use of the
`
`feedforward path decreases the linear amplifier current, ia. Ex. 1211, 2668-69.
`
`Decreasing the linear amplifier current, ia, decreases the first component of the
`
`inductor current: ASꞏiaꞏC. Therefore, maintaining the same output current requires
`
`increasing the second component of the inductor current: AFꞏvinꞏC. The increase in
`
`this component offsets the decrease in the linear amplifier current, ia. Said more
`
`simply, because io=ia+id, and because io remains constant, any decrease in ia must
`
`be offset by an increase in id.
`
`36.
`
`In his analysis, Dr. Kelley assumed that the magnitude of the inductor
`
`current, id, remains constant. Ex. 2002, ¶77. But by imposing this constraint, he
`
`assumed the conclusion he was trying to prove.
`
`19
`
`
`
`37.
`
`In any case, Dr. Kelley repeatedly admitted that neither Kwak nor a
`
`POSA would require inductor current to remain constant. See Ex. 1229, 252:3-4
`
`(“A. The switcher is trying to provide most of the current Io by way of Id.”); see
`
`also id., 184:11-185:4; 185:19-186:10; 195:9-15; 196:7-18. For example, Dr.
`
`Kelley testified that a POSA would try to increase the inductor current to the extent
`
`possible:
`
`Q. And a person of ordinary skill would understand that it
`would be desirable from an efficiency standpoint to have the switcher
`produce as much energy as possible, right?
`A.
`I think it would be better phrased the switcher provides
`as much current as possible.
`
`Ex. 1229, 245:15-21.
`
`38. The graphs that Dr. Kelley provided relating to the mathematical
`
`addition of sine waves (Ex. 2002, ¶¶ 77-80) violate this teaching of Kwak.
`
`Specifically, in his sine wave analysis, Dr. Kelley imposed the counter-factual
`
`constraint that the inductor current magnitude must remain constant as the linear
`
`amplifier magnitude decreases. Ex. 2002, ¶ 77. Dr. Kelley admitted at deposition,
`
`however, that Kwak does not impose any such constraint. Ex. 1229, 199:9-17.
`
`Rather, it is Kwak’s output current (io) that remains constant with the introduction
`
`of the feedforward path. Enforcing only this constraint from Kwak’s equation
`
`io=ia+id, when the phase difference between the total current and the inductor
`
`20
`
`
`
`current decreases – e.g., with the introduction of the feedforward path – the
`
`inductor current magnitude will increase.
`
`39.
`
`In response to Dr. Kelley’s diagrams, I provide the graphs below
`
`(Examples 1 and 2) to illustrate this relationship. In Example 1, the phase of
`
`inductor current id lags the total current io phase by 20 degrees. The magnitude of
`
`the inductor current id has a peak-to-peak value of 0.9397x2=1.8974Ipp and the
`
`magnitude of the linear amplifier current ia has a peak-to-peak value of
`
`0.342x2=0.684Ipp.
`
`
`
`21
`
`
`
`Example 1
`
`40.
`
`In Example 2, the phase of inductor current id lags the total current io
`
`phase by 10 degrees. The magnitude of the inductor current id has a peak-to-peak
`
`value of 0.9848x2=1.9696Ipp and the magnitude of the linear amplifier current ia
`
`has a peak-to-peak value of 0.1736x2=0.3472Ipp.
`
`Example 2
`
`
`
`When the phase difference is reduced from 20 to 10 degrees, the linear amplifier
`
`peak-to-peak current is reduced by 49.3% (0.684Ipp to 0.3472Ipp), while the
`
`22
`
`
`
`inductor current is increased by 3.8% (1.8974Ipp to 1.9696Ipp). As explained above
`
`in Section III.A.1.b, because the inductor current contribution to the total current is
`
`much higher than the linear current contribution, percentile changes in the linear
`
`amplifier current are more visible. As also explained above, this is compounded in
`
`Fig. 11, because the scale for the id waveform is 2.5 larger than the one used of ia.
`
`For example, in the above example, the 3.8% change in inductor current would be
`
`shown on Figure 11 as a mere 1.5% increase in the peak-to-peak magnitude of
`
`waveform id because of Figure 11’s different scales (i.e., 3.8%/2.5 = 1.5%).
`
`B.
`
`Petitioner Has Not Improperly Combined Two Different Kwak
`Embodiments
`41. Patent Owner alleges that “Petitioner relies on the hybrid switching
`
`amplifier shown in Figure 5 of Kwak as allegedly disclosing the bulk of this claim
`
`element, but then relies on Kwak’s Figure 6 for the claim requirement of ‘the
`
`switcher adding an offset to the input current.’” POR, 17.
`
`1. Figure 5 Alone Anticipates Claim 15
`In its attempt to create (non-existent) distance between Kwak and
`
`42.
`
`claim 15, Patent Owner misrepresents Petitioner’s and my arguments.
`
`Specifically, Patent Owner is wrong that Petitioner relies on Figure 6 for “the
`
`switcher adding an offset to the input current” limitation. The Petition and my first
`
`Declaration presented arguments that independently show how Kwak’s Figure 5
`
`and the relevant disclosure teach this limitation.
`
`23
`
`
`
`43. For example, the Petition states “[a]s shown in Figure 5 [] Kwak
`
`discloses the internal structure of the supply generator, including the miscellaneous
`
`claim elements that the Examiner found were missing from the prior art during
`
`prosecution (operational amplifier, driver, and PMOS and NMOS transistors).”
`
`Petition, 31. See also, Ex. 1203, ¶¶71-72.
`
`44.
`
`In the grounds section (pp. 42-49), both the Petition and my first
`
`Declaration provide an annotated Figure 5 (p. 42). Referring to the “switcher”
`
`limitation, the Petition states: “Kwak discloses this limitation. Kwak discloses a
`
`switcher highlighted in yellow below:”
`
`Ex. 1211 at 2668 (“Fig. 5. Hybrid switching amplifier with the feedforward path.”)
`
`Petition, 42. See also, Ex. 1203, ¶¶92-93.
`
`
`
`24
`
`
`
`45. The Petition then continues: “The switcher adds an offset to the input
`
`current, which will ultimately generate a larger supply current via the inductor
`
`than without the offset,” (Petition, 43) and points explicitly to the feed forward
`
`path of Fig. 5. See, Id.:
`
`Kwak discloses that the driver AF (highlighted in orange) supplies
`an increase in current—i.e., an offset current—in a “feed forward”
`path. Ex. 1211 at 2668 (“If we add a feedforward path, like the one
`shown in Fig. 5, the input signal can directly control the switching
`amplifier. Such a path is faster than the feedback current path formed
`b