UNITED STATES PATENT AND TRADEMARK OFFICE
`____________
`
`BEFORE THE PATENT TRIAL AND APPEAL BOARD
`____________
`
`POWER INTEGRATIONS, INC.,
`Petitioner,
`
`v.
`
`SEMICONDUCTOR COMPONENTS INDUSTRIES, LLC,
`Patent Owner.
`____________
`
`Case IPR2017-01328
`Patent 6,845,019 B2
`____________
`
`Record of Oral Hearing
`Held: July 19, 2018
`____________
`
`
`
`
`
`
`
`
`
`
`
`Before BRIAN J. MCNAMARA, JOHN F. HORVATH, and KAMRAN
`JIVANI, Administrative Patent Judges.
`
`
`
`
`

`

`Case IPR2017-01328
`Patent 6,845,019 B2
`
`
`
`APPEARANCES:
`
`ON BEHALF OF THE PETITIONER:
`
`
`NEIL A. WARREN, ESQ.
`Fish & Richardson P.C.
`500 Arguello Street, Suite 500
`Redwood City, California 94063
`
`
`ON BEHALF OF THE PATENT OWNER:
`
`
`ROGER FULGHUM, ESQ.
`Baker Botts LLP
`One Shell Plaza
`910 Louisiana Street
`Houston, Texas 77002-4995
`
`and
`
`BRIAN W. OAKS, ESQ.
`NICK SCHUNEMAN, ESQ.
`Baker Botts LLP
`98 San Jacinto Boulevard, Suite 1500
`Austin, Texas 78701-4078
`
`The above-entitled matter came on for hearing on Thursday, July 19,
`
`2018, commencing at 1:00 p.m., at the U.S. Patent and Trademark Office,
`600 Dulany Street, Alexandria, Virginia.
`
`
`
`
`2
`
`
`
`
`
`

`

`Case IPR2017-01328
`Patent 6,845,019 B2
`
`
`P R O C E E D I N G S
`- - - - -
`JUDGE MCNAMARA: Please be seated.
`Good afternoon, everyone. We have oral hearings in three cases
`scheduled this afternoon, IPR2017-01328, 01329, and 01453. We're going
`to do these cases in sequence. We'll start with 01328. Then we will move
`on to 01453, because I believe the reference is common to those two. And
`the last one we will do, then, will be 2017-01329.
`Beginning with the -- oh, a couple of things at the very beginning I
`want to mention. We have Judge Jivani and Judge Horvath participating
`remotely, and so, therefore, I would ask everyone to speak clearly from the
`podium and into the microphones so that they can hear you, and identify any
`demonstrative by number that you are referring to at any point in time so
`that they can pull it up on their screens and will have that available to them
`as well.
`
`Let's see. Beginning with the Petitioner, could you please
`approach the podium and identify who you -- and enter your appearance, let
`us know who's here.
`MR. WARREN: Your Honor, Neil Warren for the Petitioner,
`Power Integrations.
`JUDGE MCNAMARA: Mr. Warren, are you going to be arguing
`all three cases?
`MR. WARREN: Yes, Your Honor, I will be.
`JUDGE MCNAMARA: Okay, great. Thank you.
`And for Patent Owner?
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`
`3
`
`

`

`Case IPR2017-01328
`Patent 6,845,019 B2
`
`
`MR. FULGHUM: Good afternoon. Roger Fulghum, lead counsel
`for Patent Owner. For the 01328 case, that case will be argued by Brian
`Oaks, and for the 01329 and 01453 case, that case will be argued by
`Nicholas Schuneman. Also with us today is IP counsel for ON
`Semiconductor, Josh Engel, and our legal intern, Leonard Cortelli.
`JUDGE MCNAMARA: Okay. While you're here, let me ask you
`one other question in terms of procedure today. I believe the Patent Owner
`has a motion to exclude in each of the cases, and it's essentially the same
`motion. Are you planning to present oral argument on the motion to exclude
`today?
`
`MR. FULGHUM: We were not planning to do so.
`JUDGE MCNAMARA: Okay, great. Then that will at least cut
`off a little bit of the extra jumping up and down, having everybody address
`those questions. All right, thank you very much.
`MR. FULGHUM: Thank you, Your Honor.
`JUDGE JIVANI: Lead counsel, just to follow up on that, you
`recognize this is your only opportunity to present that argument if you
`wanted to, correct?
`MR. SCHUNEMAN: We do appreciate that, and notwithstanding
`that, we will not plan to present argument today.
`JUDGE JIVANI: Great, thank you.
`JUDGE MCNAMARA: All right. If everyone is ready to begin,
`welcome to the Patent Trial and Appeal Board, and we will begin with
`Petitioner in case 1328. In each case we have allocated 45 minutes per side,
`so is there some amount of time you would like to reserve for rebuttal?
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
`4
`
`

`

`Case IPR2017-01328
`Patent 6,845,019 B2
`
`
`MR. WARREN: Yes, Your Honor. I would like to reserve 15
`minutes for rebuttal.
`JUDGE MCNAMARA: All right. So what I will is just put this
`up on the screen, and it will alert you when you have used up 30 minutes.
`MR. WARREN: I appreciate that.
`JUDGE MCNAMARA: All right. Please proceed.
`MR. WARREN: Thank you, Your Honor. Neil Warren for the
`Petitioner.
`As you know, this matter deals with the '019 patent, and in this
`matter, we have a motion to amend that's been presented by the Patent
`Owner, so the two claims that are at issue are claims 16 and 17 from the
`motion to amend, and, Mr. Sayres, if we can go ahead and go to slide
`number 2.
`In the Demonstrative Number 2, we see the front cover and the
`first figure from -- or primary figure from the '019 patent. The '019 patent
`refers to a switching power supply. On the far right side of this figure, you
`see something indicated as the load. The load is something like a laptop
`computer that's connected to a wall through a power supply, and in this
`patent, the invention, what it really relates to is something that in the field
`we call cable compensation.
`So, you know, the length of the actual cable that comes out of your
`power supply can vary by manufacturer, so different manufacturers may use
`the same controller chip, but they may have different lengths of cables that
`they want to use. And as we know, a cable, even though it's copper wire,
`has a very small amount of resistance, so proportional to the current, you're
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
`5
`
`

`

`Case IPR2017-01328
`Patent 6,845,019 B2
`
`actually going to get a cable drop, what we call a cable drop, which is
`actually a voltage drop across that cable.
`So in order to allow a single chip to be used with many different
`lengths of cables, a process was developed that allows you to compensate for
`that drop of the output voltage so that the voltage you see at the output of
`your power supply is actually the voltage that you see at your load of your
`laptop computer regardless of the length of the wire.
`And as I said, that -- that drop is proportional to the current, that's
`the output current, and in these types of power supplies, as is explained in
`the patent, that's proportional to the primary side switching current, and so
`there's a relationship between the current that's developed on the primary
`side and the output current. So the basic idea is that on the chip side, you
`detect what that peak primary side current is, and you compensate your
`feedback signal to provide for this cable drop compensation.
`So if we go on to slide number 3. Yeah, slide number 3, this
`shows the prior art. This is the Bonte patent.
`JUDGE JIVANI: Counsel, before we get into Bonte, would you
`go back a slide for me, please, and identify for me output at the feedback
`signal from your feedback generator in this figure?
`MR. WARREN: Well, that's actually an interesting point. So I
`was going to get into this later, but -- so this is actually the challenged
`patent, and in my opinion -- so, there's something called VFB, which is the
`feedback, which is at the bottom of the bottom left corner.
`JUDGE JIVANI: Yes.
`MR. WARREN: Now, one of the issues that we have in this case
`is that the claims present new matter, and part of that new matter is that what
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`26
`
`6
`
`

`

`Case IPR2017-01328
`Patent 6,845,019 B2
`
`the patent describes is generating a feedback signal -- sorry, what the claims
`describe is generating a feedback signal and then correcting that feedback
`signal.
`
`And so if we look at this, the mechanism that's used to correct that
`feedback signal is this current that you see in the middle that's actually going
`into the gain regulator. The gain regulator is syncing a current, and it's
`syncing a current across the resistor R120, and that's correcting the feedback
`signal.
`
`And so what we have is we have in the specification and in the
`proposed claims, we have this disconnect between what we see in the
`specification and what's in the claims. So I would agree that in the
`specification, you really don't see a separation between this generated
`feedback signal and this corrected feedback signal. It's all -- as you can see
`from this structure, the thing that is generated, that VFB, is generated by the
`voltage drop across that resistor, which is inherently generated, in part, by
`that current that's being sunk there.
`And so as I'll discuss later when we get to the new matter issue,
`that actually presents a problem with the challenged claims because there
`isn't a separation. As the claims recite, generating a feedback signal,
`correcting it, and then outputting a corrected feedback signal, that's actually
`not in the specification, and that's why we say it's new matter. So it's a -- it's
`actually not there, and that's a big problem for this patent.
`So if we -- does that answer your question or can I move on to
`slide number 3?
`JUDGE JIVANI: You can move on.
`MR. WARREN: Thank you.
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`26
`
`7
`
`

`

`Case IPR2017-01328
`Patent 6,845,019 B2
`
`
`So slide number 3, this is the Bonte prior art, and what Bonte
`describes is exactly the same type of invention, exactly the same thing as is
`in the '019 patent, and in the bottom left, you see this box, load regulation
`compensation circuitry, which is coupled to the peak detector, 670, and this
`is from Figure 6B of Bonte.
`And just like we see in the '019 patent, that is syncing a current -- I
`note that on this figure there's an arrow that kind of looks like it's pointing
`towards the feedback. When you look at Figure 8, you can see that it's
`syncing a current. It's doing exactly the same thing as the '019 patent. It's
`syncing that current and -- to do the cable compensation.
`So at a high level, you know, just looking at the figures, we have
`basically -- we have basically the exact same thing. And so in the original
`petition, what we had was an anticipation claim, and that -- the Board found
`certain claims were anticipated, certain claims were not anticipated, and the
`Petitioner or the Patent Owner --
`JUDGE JIVANI: Let me pause you there for a moment. Certain
`claims -- we found a reasonable likelihood that certain claims may be
`anticipated.
`MR. WARREN: I apologize. You're right, Your Honor.
`And so taking some of the elements, not all of the elements, but
`some of the elements where it's found that there is not a reasonable
`likelihood that they would be found anticipated, certain elements were
`incorporated through the motion to amend into the claims 14 and 15, but
`they were modified.
`And so in response to that, what the Patent Owner didn't do was
`they only considered anticipation at that stage; they didn't consider
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`26
`
`8
`
`

`

`Case IPR2017-01328
`Patent 6,845,019 B2
`
`obviousness. And so while from a high level these two inventions seem
`very similar, as we'll talk about, there was one element that was found to not
`be disclosed in the Bonte figures that we cited, and as we'll discuss in detail,
`it was found to be -- or it is disclosed in Figure 3.
`So if we can go on to Figure or slide number 6, this is --
`summarizes that. This is from the institution decision. Figure 5 was the
`originally cited figure from the petition, and what we're talking about here is
`basically this -- the claims recite a driving power that's used to actually
`power the chip, and it also describes that the driving power is what's used to
`generate that feedback signal.
`And if you look at the figure that's in the bottom right here, we see
`that the -- there's actually two separate chains. There's diode 501 and diode
`503 that each develop their own signal, one of them generating the supply
`voltage for the switching regulator; the other one developing the feedback
`signal.
`
`And in the institution decision, it was found that because the
`driving power and the driving power signal, as a matter of claim
`construction, would be the same thing, that we have two different things
`going on here. We have a separate signal that's driving the power and a
`separate signal that's being used to generate the feedback signal.
`If we can go on to the next slide. Sorry, the next slide. Sorry. So
`what we have and what Patent Owner didn't note was that Figure 3 of
`the Bonte prior art actually shows exactly the same thing that's in the '019
`patent, and it shows a single signal that is being used to develop the supply
`voltage and is being used to correct the feed -- and to generate the feedback
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
`9
`
`

`

`Case IPR2017-01328
`Patent 6,845,019 B2
`
`signal. So that cures the one issue that was raised, and we can see that in the
`comparison between these two.
`And it was discussed in Bonte -- and we'll discuss the reasons,
`because the reason is very important in this case -- why Bonte considered
`what's in Figure 5 an improvement over what's in Figure 3. As it turns out, it
`has nothing to do with the load compensation circuitry that's at issue in this
`case, but I'll get into that in more detail, because that's really one of the crux
`issues in this IPR.
`If we can go on to the next slide. So the obviousness combination
`that was proposed in the opposition to the motion to amend was basically a
`very simple combination of the circuitry. So you see that in the right-hand
`side, which is Figure 5, there is the feedback circuitry, 414, and you have a
`circuitry that performs the identical function, just in a slightly different way,
`on the left side in Figure 3.
`And as you can see, both have an input coming in off of the
`winding, off of 404; both generate a supply voltage; and both output the
`feedback signal. So at the core you have exactly the same inputs and
`outputs. The structure is just slightly different, and the structure actually
`matches up exactly with what Patent Owner suggested was not in Figure 5
`and what the Board -- the reasoning for the Board finding nonanticipation of
`the claims in the original petition.
`And so if we go on --
`JUDGE HORVATH: Mr. Warren, does Bonte say anything about
`the size of -- anything about -- any difference between diodes 501 and 503,
`their operating characteristics or anything of that, or are they the same type
`of diode?
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`26
`
`10
`
`

`

`Case IPR2017-01328
`Patent 6,845,019 B2
`
`
`MR. WARREN: So I don't believe it goes into specifics, which is
`actually an interesting point when you get to that, because these are basically
`generating -- the diode's purpose there is really rectification, so it's rectifying
`there, but then when you compare -- you have two capacitors that are at
`issue here that are doing somewhat of a filtering operation.
`And so without getting too much into the weeds, when we look at
`Figure 3, the prior art, what Bonte noted was that that capacitor there is
`actually performing a filtering operation, which would basically peak detect
`that signal for the feedback signal.
`And as was noted by our expert, you know, if you were to compare
`those two circuits, you could actually get fairly similar outputs generated by
`them by changing those relative capacitors. So, for example, if the capacitor
`was made in 330 very small, so that it's not doing very much of a
`peak-detection operation, you can get very similar behavior as compared to
`what you would see in Figure 5. But to answer your question, no, I don't
`believe Bonte does go into details. Those are primarily there for
`rectification.
`If we can go on to the next slide.
`JUDGE JIVANI: Actually, let's stay there for a minute.
`MR. WARREN: Yes.
`JUDGE JIVANI: So, Counsel, Patent Owner contends that Bonte
`teaches the feedback signal -- if you were to sort of make the substitution
`that you propose, the feedback signal that would be presented to amplifier
`350 will be artificially large and the output voltage will sag.
`How do you respond to that?
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
`11
`
`

`

`Case IPR2017-01328
`Patent 6,845,019 B2
`
`
`MR. WARREN: So that's actually not true. So that's not what
`Bonte says, and that's not what happens here. So the difference is that Bonte
`has a separate invention in here that has to do with -- actually, one of the
`issues that's at issue in one of the other IPRs we're going to talk about today,
`which is the spike blanking. So you get this initial spike that happens based
`on ringing from that the winding, and so it's not really that big of a deal
`when you're peak-detecting, as in the left side, and that is conventional. It's
`prior art. It works just fine for the feedback portion, but it has the problem
`of that inductance spike at the very beginning.
`Bonte proposes the circuit on the right side so that you can separate
`your feedback signal generation from the big capacitor that has to power
`your chip, so that you can get much more sensitivity, but that also brings
`into it the inductance spike, and so he has a separate circuit.
`But as explained by our expert and as, you know, a person of
`ordinary skill in the art would understand, both of these circuits, especially
`Figure 3, will work just fine in terms of feedback. It does provide a
`peak-detected feedback signal, but that's perfectly fine for doing the basic
`regulation function, which is what is being dealt with in the cable
`compensation issue of the '019 patent.
`Your feedback will respond. You will be able to determine it.
`And as I said earlier, based on how big or small that capacitor on 330 is, you
`can actually get very similar results. It is peak-detected, it is filtered
`slightly, but not very much. You can still get a lot of wiggle in there if that's
`a relatively small capacitor and still big enough to power your device.
`If we move on to the next slide, so this slide shows the simple
`obviousness combination that we've proposed. This is slide number 10, and
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`26
`
`12
`
`

`

`Case IPR2017-01328
`Patent 6,845,019 B2
`
`this comes straight out of the opposition to the motion to amend. Essentially
`all we've done is we've taken the circuitry that was in Figure 3, that
`corresponds to the exact same circuitry from Figure 5, and we've replaced
`the circuitry from Figure 5.
`So you can see in Figure 5 -- this is Figure 5, but modified Figure
`5, obviously -- what's now inside the box, 414, the feedback generation
`circuitry, that's actually the feedback that took these same inputs and
`generated the same outputs from Figure 3. So it's a very simple obviousness
`combination. It's a very simple copy-and-paste actual obviousness
`combination.
`And so what we've heard from the Patent Owner is this argument
`that a person of ordinary skill in the art would never do this. So there has
`been no argument from the Patent Owner that it doesn't teach all the
`elements of the claim. There's no dispute in this IPR right now that all of the
`elements of the claim of this combination are taught. The dispute is a purely
`legal argument on the obvious -- on the standard for obviousness, as to
`whether a person of ordinary skill in the art would make this combination or
`could make this combination.
`And if you look at the language that --
`JUDGE JIVANI: But they are two different questions, aren't they,
`would and could make this determination -- this combination?
`MR. WARREN: Absolutely, they are two different questions, but
`if you look at the language or KSR, KSR discusses that if you take
`conventional circuitry and you use it in a conventional way, unless there's
`some significant, unpredictable results, or if the method of the combination
`is beyond -- the "how" is beyond the skill of a person of ordinary skill in the
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`26
`
`13
`
`

`

`Case IPR2017-01328
`Patent 6,845,019 B2
`
`art, it's obvious. Those are the two grounds that KSR discusses, and that
`hasn't been proposed by the Patent Owner.
`So it's never been proposed that this has some kind of an
`unpredictable result. We know it doesn't. This is the Bonte -- Bonte says it's
`conventional. It can't generate unpredictable results. We're going
`backwards. And it hasn't been suggested that this copy-and-paste, you
`know, that we can make on a PowerPoint slide is somehow beyond the skill
`of a person of ordinary skill in the art.
`What they've suggested is that if you were to do this combination
`and you went back to having the conventional peak detected feedback signal
`generation, that it wouldn't work with the leading edge -- the blanking, the
`spike blanking circuit. If you look throughout their papers, the only
`functionality that they claim would be hurt by this combination is the spike
`blanking.
`And as I discussed before, Bonte has numerous inventions.
`Remember, we looked at that first Figure 6 that had two separate boxes. The
`box that had the error amplifier is where the spike blanking circuit is, and the
`load compensation circuitry is where the circuitry for doing what's in the
`'019 patent is. They're completely separate.
`We know they're completely separate because Bonte tells us
`actually that they are completely separate. And so, Mr. Sayres, if we could
`go back to slide number 4, this is -- slide number 4, Bonte, Exhibit 1002,
`column 8, lines 1 through 9, and Bonte numerous times expressly states that
`it's disclosing numerous inventions, a variety of inventions, as we'll see the
`language is in other places.
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
`14
`
`

`

`Case IPR2017-01328
`Patent 6,845,019 B2
`
`
`It's not the case that Bonte says that you have to use my Figure 5 or
`the whole thing doesn't work. Figure 5 is only ever -- the changes made to
`Figure 5 are only ever associated with the spike blanking, and the patent
`itself says that "The present invention includes, alone or in combination, a
`unique error amplifier" -- which deals with the spike blanking, or,
`separately -- "a unique load regulation compensation circuit," which is what
`we're dealing with here with the cable compensation in the '019 patent. So
`column 8, you know, lines 1 through 9, it says expressly that these two
`separate inventions can be used alone or in combination.
`And if we go on to the next slide, slide number 5, this is a snippet
`from Exhibit 1002, at column 2, line 64, to column 3, line 24. What we see
`is that Bonte is very express, that Bonte knew that he had created multiple
`inventions and included them in the same specification. He describes them
`as a variety of novel circuits. And when describing the variety of novel
`circuits, he very carefully singles out both the unique sampling error
`amplifier circuit for tracking the magnetic flux sense feedback signal -- that's
`what I was discussing earlier, that tracking as being, you know, more
`specific, but having that spike -- and separately, it discusses the load
`compensation circuit, which is included for providing current compensation
`to cancel additional dependent errors caused by parasitic impedances in the
`output circuitry. That is -- those are Bonte's words for the cable, and the
`parasitic impedances are the -- is the voltage drop across that cable.
`So what we have here is Bonte itself telling us that these are
`separate, and at the end, in the column 3 portion, we see that Bonte describes
`that the circuits can be incorporated together into the switching regulator or
`it can be used in various combinations with other types of circuitry. So
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`26
`
`15
`
`

`

`Case IPR2017-01328
`Patent 6,845,019 B2
`
`throughout Bonte, we see this teaching that you don't have to use my Figure
`3 or my Figure 5 in the error signal -- unique sampling error amplifier circuit
`with my load compensation circuitry.
`Throughout the petition, the circuitry that we cite is only related to
`the unique load compensation circuit. That's the same as in the opposition to
`the motion to amend. So the argument that we've heard from the Patent
`Owner is really neither here nor there when you actually look at the
`teachings of Bonte.
`So the other issue that's at issue -- first off, before I move on, are
`there any questions on that?
`(No response.)
`MR. WARREN: If there aren't, I'll move on to the second issue
`that's at issue in this IPR, which is the new matter issue, and we have -- the
`Petitioner has identified several areas where we believe that the claims
`introduce new matter into the claims, and, therefore, the motion to amend
`should be denied.
`Mr. Sayres, if we could go to Demonstrative Number 13, this is a
`table, what we see on Demonstrative Number 313, which was provided in
`the opposition, which is a redline comparison of one of the key or the
`primary addition, the feedback signal generator, and what the Patent Owner
`said in their motion to amend was we're doing what's in claim 1. The Board
`found claim 1 wasn't anticipated, so we're going to do what's in claim 1 and
`incorporate it into claim 14.
`We can see that that's not really what they did. There's several key
`differences here, and I'll discuss them both. So the first difference is that in
`the original claim 1, we have a very specific reference to a very specific
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`26
`
`16
`
`

`

`Case IPR2017-01328
`Patent 6,845,019 B2
`
`current signal, and that is the italicized portion of this, the "corrected current
`signal output from the gain regulator," and they've replaced that with a much
`broader term, the "current input signal."
`And as the papers describe, this is a broadening, and we think it
`introduces new matter because there isn't matter to support it, but the much
`more fundamental problem with their claim gets back to what we were
`talking about earlier at the very beginning. So if we look at the claim that's
`on the right side -- and, actually, why don't we go ahead and go back to -- if
`we could, I think it's the prior slide that has the language on it, because I
`really want to discuss with you -- slide number 7.
`So this is Demonstrative Number 7 that we see here, and this is the
`claim -- this is the feedback unit element, this is from the motion to amend,
`and what I really want to point out here is what I discussed earlier, is that the
`claim, if we look at that top portion, it recites several steps which we cited in
`the petition as being problems under IPXL, but, you know, that's not an issue
`for here, but we've got the step of generating a feedback signal. That's
`something that the feedback unit has to be coupled to do. It then has to take
`another step. It has to detect a switching current of the switch to correct the
`feedback signal. And that phrase, "to correct the feedback signal," was
`construed by the Board. And so I'll get back to what that construction is
`why it fully demonstrates that their argument is nonmeritorious.
`Then the last step that has to be done is you have to output the
`corrected feedback signal. So that's really the issue here, because when we
`look at the very bottom, the very bottom is the feedback signal generator.
`The feedback signal generator performs those steps of correcting the
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`
`17
`
`

`

`Case IPR2017-01328
`Patent 6,845,019 B2
`
`feedback signal, and yet it's used to generate the feedback signal, which has
`an antecedent reference back to the originally generated feedback signal.
`So in this claim, putting aside anything else, we know two things.
`We know that a feedback signal has to be generated, and then a corrected
`feedback signal has to be created, and that corrected feedback signal has to
`be output.
`And so if we look at slide number 14, please, what we have here is
`the claim construction for what that phrase "to correct" means, because I'll
`get into what their argument is in a minute, but, briefly, they basically claim
`that the feedback signal and the corrected feedback signal are exactly the
`same thing and that they're merely different by a simple adjective.
`And so that doesn't make sense in the context of the claim or in the
`context of the claim construction, because the step of correcting the
`feedback signal, you have to be adjusting something. There has to be
`something that gets adjusted, and in the context of this claim, correcting the
`feedback signal means adjust the feedback signal to compensate for errors in
`the output voltage detection. And they have not challenged this claim
`construction.
`So what we have is we have -- we have the claim and the
`construction that describe that there must be something done. There must be
`an original feedback signal, and then there must be something that is done to
`correct it.
`JUDGE JIVANI: Counsel, let me stop you for a minute. You
`have actually omitted from this slide our construction, and yet you just
`argued that that's our construction. It is not. The omitted sentence reads
`from our decision on institution:
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`26
`
`18
`
`

`

`Case IPR2017-01328
`Patent 6,845,019 B2
`
`
`"For purposes of this decision, we construe the term 'correct the
`feedback signal' to encompass at least adjust the feedback signal to
`compensate for errors in the output voltage detection."
`We then continue to direct you to further construe and argue this
`phrase if it was needed, and we directed the parties to address the broadest
`reasonable construction. So for the clarity of the record, that demonstrative
`there -- what slide is that, 14 that you have? -- that is not our construction.
`MR. WARREN: All right, I appreciate that. So the -- the claim --
`the words that are there were quoted directly from the institution decision.
`JUDGE JIVANI: Yeah, and what I'm saying is that you didn't read
`far enough down.
`MR. WARREN: All right.
`JUDGE JIVANI: The immediately following sentence is our
`construction.
`MR. WARREN: All right. I'm sorry if I misunderstood the order.
`In any event, what I want to get to, if I may, is if we can go to slide
`15, this is the argument that was presented by the Patent Owner, and what is
`highlighted is the key portion, which says, "Substitute claims 16 and 17 thus
`refer to a single 'feedback signal,' and 'corrected' is simply an adjective
`modifying the same 'feedback signal.'"
`So at bottom, in order to find that they have adequately disclosed
`or that there is adequate support in the original matter, it must be found that
`that claim element that we looked at before has only a single feedback single
`and that the word "corrected" is simply an adjective modifying it.
`Now, I'm not aware of any Federal Circuit law that has simply an
`adjective exception to it. Typically, when we distinguish claim elements, we
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`26
`
`19
`
`

`

`Case IPR2017-01328
`Patent 6,845,019 B2
`
`use adjectives. We say things like a first or a second, and historically,
`adjectives always distinguish between two different kinds of things.
`In the context of this claim, I would submit that it doesn't make
`sense that they are the exact same thing, and so in the context of the claim
`itself, it doesn't make sense that the thing that would be output would be the
`feedback signal.
`Now, as you noted earlier and as we discussed, what's described in
`the specification, there is just the VFB