Oral argument — Power Integrations, Inc. v. Fairchild Semiconductor

Before we begin with the cases this morning, I have the pleasure of asking Judge Chen to make a motion for admission of his law clerk, Luke Burton. Yes. Thank you, Judge Dyke. This morning I'd like to move the admission of Luke Burton, who is a member of the bar and is in good standing with the highest court of California. I have knowledge of his credentials and am satisfied that he possesses the necessary qualifications. I know these things because Mr. Burton has been my clerk for the past year. Like all of my clerks, Mr. Burton has done excellent work. He's proven himself to be a very good lawyer. He is creative. He thinks through the issues well and helps me think through the issues, writes well, has been a hard worker and willing to help his colleagues in the chambers. He's also someone that marches to his own drum. The first thing he did when he got out here was get a bright red Vespa, something Judge Clevenger knows well. And I wish him well on his return to California. So for all these reasons, I respectfully request that my motion to admit my clerk be granted. I'm delighted to vote in favor of the motion. All right, your motion is granted, Judge Chandler. We welcome Mr. Burton to the bar of the court. So you should take the oath. That's a necessary part of this. Thank you, Judge. Do you solemnly swear or affirm that you will conform yourself to the terms and conditions of this court, by the power vested in me by the court of law, and with the support of the Constitution of the United States of America? I do. Congratulations. Thank you. Okay, we have four argued cases this morning. The first one is number 162691, Power Integrations, Inc. v. Fairchild Semiconductor, Ms. Sullivan. Good morning, your honors, and may it please the court, Kathleen Sullivan for the Fairchild appellants. This appeal involves a $147 million judgment for infringement of two patents related to power supplies, an amount reflecting a royalty rate that is twice the selling price of the units, and that Power Integrations' own damages expert admitted would have bankrupted the defendant at the time of infringement. I'd like to focus on why you don't need to reach the damages, because the evidence is legally insufficient to support infringement on either patent. But I want to at least remind the court or advise the court that all the asserted claims of both patents have been canceled in a final action by the PTAB, and that those have been docketed as appeals to this court. Can you tell us the timing of that appeal? What stage is it at? Yes, your honor. When will the briefing be completed? It is docketed. It's awaiting the certified list coming over from the board, and the briefing has not yet begun. Weren't the board decisions issued over half a year ago at this point? Yes, your honor. Well, in September. So the final action. So, your honor, that will be before you, and there may be motion practice on the timing, but I know it doesn't affect the outcome on this appeal, so can I go back and cover the infringement issues? I'd like to begin with the 079 patent, your honors, and this is a patent where the asserted claims require a fixed switching frequency for a first range of feedback signal values. The district court construed the meaning of fixed switching frequency to be a non-varying number of switching cycles per second. And crucially, your honors, power integration has not questioned that construction on appeal. They have not sought to overturn the construction of fixed as meaning non-varying. So it's a simple road to a finding of insufficient evidence of infringement, because it's undisputed in the record that all of the accused products have a varying number of switching cycles. Do we have a claim construction issue in that respect before us? You do not, your honor. So we argued at claim construction. So it's entirely under Hewlett-Packard, then it's entirely a question of whether the claim construction that was given could reasonably be interpreted by the jury to reach these products. Exactly, your honor. And it cannot, because it's undisputed that the accused products have a varying number of switching cycles, even on a per second time scale. So there's a separate issue about the per second construction. You don't need to reach it if you agree with us on this, because we asked the court to construe... This is the 5 to 15 percent? That's right, your honor. Temperature variation gives you 5 to 15 percent. But the problem is you seem to admit, or counsel for Fairchild seemed to admit, that the claim construction should allow for some variability due to environmental factors. We did not, your honor. We argued that fixed means no variation whatsoever at Markman, and we maintain that argument now. So how is that possible to have a product that isn't responsive to environmental factors? Your honor, it may be a problem, but it's not a problem for Fairchild. It's a problem for Power Integrations. It may mean that they sought and obtained a claim that is inoperable. Why isn't inoperability your problem? It's not our problem, your honor, because they have not challenged the claim construction of fixed, meaning non-varying, on appeal. So for present purposes, the claim construction, if it renders it inoperable... Well, they're relying on Dr. Kelly's interpretation of those words. Your honor, Dr. Kelly can't give extrinsic evidence now to overcome the claim construction. The claim construction was fixed means non-varying. No amount of extrinsic evidence now can change that claim construction where Power Integrations has failed to challenge it on appeal. That's why we're in the world of Chef America and not the world of Ecolab. It's not extrinsic evidence. It's his view that the words say fixed means you count the cycles within the second. Your honor, you're on a separate issue. Let me separate two issues. We went on 5% to 15% because it's undisputed that... Wait, wait, wait. It's going around in circles. If your interpretation of the claim is that fixed means no variation whatsoever... That's correct. As everybody understood at Markman, you have an inoperable claim. That's not correct, your honor. Yes, it is. And your lawyer there threw Chef America under the bus by proposing an alternative claim construction. Your honor, we did not. At Markman, we said fixed means no variation. And you had a backup. You had a fallback. It's a fallback, your honor. But the construction is made. The construction of fixed as non-varying is made. That is the law that governs here because Power Integrations has conceded that that is correct. There's no evidence that the 5% to 15% is due to anything other than environmental factors, right? That's correct, your honor. But Dr. Kelly admitted that those frequency changes, 5% to 15%, happen even over a per second timescale. So he admitted it. The admission is at A630. He admitted it also in a part of the transcript that's not in the Joint Appendix, transcript 2939 and 2982. Dr. Kelly admitted that there is the 5% to 15% variation due to environmental factors. Now, your honor, the reason that's a problem for Power Integrations and not for Fairchild is if it renders the claims inoperable, there may be an invalidity for lack of enablement, but it doesn't change the claim construction that governs the infringement judgment in this case. Infringement has to be adjudged under the unchallenged claim construction that fixed means non-varying. I see you're focused on the word fixed, but I mean, I'm wondering why shouldn't we be really thinking more about fixed switching frequency and trying to understand what does that phrase mean and is there anything in the spec that channels us to command that fixed switching frequency to one of skill in this art after reading this patent would be compelled to think of it as a completely rigid non-varying frequency that couldn't be subject to normal variations due to environmental conditions. So your honor, first of all, Markman at Markman, after Markman, the claim construction was fixed means non-varying and that was not challenged. Right, but the entire phrase... That's not the limitation that's at issue. Yeah, the entire phrase construction... Okay, your honor, fixed switching frequency was construed as non-varying. Non-varying... Non-varying number of cycles per second. Correct, your honor. What's a fixed number? If you disagree with us that the temperature variations are not... temperature variation was not included in the claim construction. Power integrations didn't say... look at what power integrations asked for, your honor. What they asked for was a construction that said fixed means substantially. And everybody understood that substantially is such a squirrely word that nobody wanted to live with that. Your honor, if you don't agree with me on temperature variations controlling, then you need to reach the jitter products. The jitter products, which are a substantial subset of the products, infringe for the separate reason that the per second addition to the switching frequency limitation was wrong as a matter of law. It was wrong as a matter of law... I don't see that you argued in the opening brief, and maybe I'm wrong and you can show me where, that the error was adding the per second limitation to the claim construction. We did, your honor. We argued it in the opening brief. First of all, we're entitled to stick to our construction, which is fixed means non-varying over any time scale. We did argue in the opening brief that adding the per second limitation was legal error, and that's in the blue brief at 31 to 34. And, your honor, the reason we argued that is that... Hold on, hold on. Okay. I'm going to run out of time to talk about damages, your honor. We'll give you some time. Thank you, your honor. We argued that the addition of the per second limitation was error, and the reason, your honor, is that the specification refers to time scales of microseconds. The specification we counted, it refers to the microsecond time scale... You could have made that argument either to Judge Ware or Judge Chesney below when they invited you. Your honor, we weren't required to because your case law is very clear. We cite 02 micron, and we cite the Pabst case. It's very clear that if we make the same claim construction argument on appeal that we made at Markman, and we do, we are entitled to object to the claim construction and seek its overturning as a matter of law, whether or not we objected to the instructions. And Pabst is the identical case because there was a limitation added at claim construction that was not part of the defendant's proposed claim construction. The defendant didn't make further objection but was able to assert on appeal that the addition of the additional element was wrong. That's our case. We think addition of per second was self-evidently wrong here because the spec refers 58 times to microseconds. But as I understand the way claim construction was argued below, it was your position was consistently always this concept that it should be non-variant. Yes, your honor. Fixed switching frequency. That's correct, your honor. But then when the per second term got introduced into the claim construction, there was never an attempt to do some kind of 02 micro clarification of the construction or construction of the construction in order to get a better understanding of what per second ought to be. So perhaps I accept your idea that you did not waive the right to appeal the question of whether or not fixed switching frequency ought to be a completely non-variant frequency. But I'm not sure that you've in any way preserved an argument over whether per second can't be there versus per microsecond or something like that. Your honor, I believe we did under Pabst. In other words, the claim construction chart is at 2125 of the joint appendix. And power integrations lost its claim construction. And it does not try to revive that on appeal except indirectly. It argued for fixed switching frequency, meaning the target switching frequency is intended to be substantially fixed. Judge Ware plainly rejected that. He chose our meaning of fixed to mean non-varying. We proposed fixed switching frequency means the switching frequency does not vary. We're getting confused here because there are two different arguments. One is the 5 to 15%. Let's assume that we reject your position on that. But that leaves the jittering issue. And what you're arguing there, as I understand it, is that the district court improperly included the per second limitation, which allowed them to argue to the jury that within a particular second, any variation didn't count, correct? Exactly right, your honor. And we should prevail, we believe, on the first argument. But if you don't agree with us on the first argument, then we should prevail on the frequency hopping or jitter problem. But let's assume for a moment that you preserved the idea that the per second limitation should have been eliminated from the claim construction. What is it in the intrinsic evidence that suggests that that's the correct construction? That the elimination of the per second construction. And the answer to that is, your honor, that jitter occurs on a microsecond scale. The number of times the power switch oscillates is happening on the scale of 100,000 or 68,000 times a second, or varying between 62,000 and 68,000 times a second. So it makes no sense. It's legally erroneous to construe a claim whose specifications speak in microseconds 58 times as number of switching cycles per second. Do they speak of microseconds in terms of the thing being fixed? Your honor, the claim is fixed switching frequency. And we think it's legally erroneous to read into that a per second limitation. A fixed velocity for a car would be... No, I move. You know the car analogy. I don't need to go over that. I'm sorry, are you trying to say that the construction ought to be amended from per second to per microsecond? Your honor, that would be acceptable, or the elimination of a time frame. So to simply say fixed switching frequency... Right, because as I understood it in the specification, they were talking about a given cycle being 10 microseconds. And then maybe if you have a 10% duty cycle, it would be the length of one microsecond. So you would not have any cycle on a microsecond. You would have at best a tenth of a cycle in a microsecond. Please don't shake your head. That's not helpful. I'm sorry? Gentlemen in the audience, please don't communicate to the court that way. I'm sorry, your honor. I just want to make sure you weren't addressing me. So here's the argument. There's no question that the switching cycle frequency in the jitter patents varies and is not fixed on the microsecond level. If you eliminate the gratuitous claim construction, and we believe erroneous claim construction, of fixed switching frequency... We just asked for construction of fixed switching frequency. The court went on to define switching frequency as the number of cycle switches per second. The per second... And to be clear, you never objected to the per second? Until here. Your honor, we didn't object to it. I just want to make the best answer to my question. Yes, and I'm asking you to just allow us to argue that our claim construction, which didn't have that limitation, is the one we can assert now under PAPS. You're saying you preserved it by asking for construction, which wasn't adopted. That's correct, your honor. Exactly what happened in PAPS. I'm concerned I need to get to damages. May I simply leave you with the argument? I want to be clear, we should win on all products because of the 5 to 15 percent, which the other side admitted exists. If anyone wanted to say that the claim should be amended to include that as natural variation, they have waived their opportunity to do that because they haven't challenged the claim construction on appeal. And we know that the judge at Markman was construing fixed to mean non-varying by reference to a dictionary where he said it's not subject to change or variation. So to answer your question, Judge Chen, we believe either you would replace per second with per microsecond or eliminate the time element and simply construe it as fixed. You never asked for the microsecond construction, right? I'm sorry, your honor? You never asked for a microsecond. That's correct, your honor. That's correct. We're relying on the argument that we can stick with our initial proposal, which is fixed means not subject to variation. And the addition of elements was error. But if you disagree with us on the infringement, the absence of legally sufficient evidence for the infringement of either patent, 908, I'll leave to the briefs unless there are questions. It's a straight argument that I'm DOE. Why don't you go ahead. Yes, sir. I want to simplify the damages argument into two. First, we have an argument that the evidence is legally insufficient to support an entire market value rule royalty base where the evidence doesn't come close to satisfying the requirements this court set forth in Laser Dynamics. The evidence doesn't show that the patented feature, the 079 feature, constituted the basis for consumer demand. We have no quarrel with the jury instruction. The jury instruction at appendix 1954 is correct. What did their expert testify about this, the one watt feature, whatever you want to call it? Your honor, the one watt feature. Did he say that it was responsible for the demand, that it drove the demand? I don't recall if the, there was a lot of testimony about the one watt feature. And yes, there was testimony both lay and expert that it drove demand. But here's why that is not legally sufficient. First, the one watt requirement applied only to government purchased products. And I can't think of any example of the entire market value rule in which a major, a submarket's requirement for a product governs the rest of the market. So that's the first problem. But the second problem is think of the implausibility of the argument under your Laser Dynamics reasoning. What did their expert witness say about these other features, which were characterized and apparently agreed by both parties, characterized by the judge, as valuable features? What did their expert witness say about that? That the expert claimed in conclusory fashion that the 079 feature, the power switching regulator feature, drove demand. Notwithstanding that, number one, on jitter features, power integrations, as this court well knows because these cases came to you before, and the jitter features, they've sued us saying the same products infringe jitter features. It's inconsistent to say that- Did their witness say that the other features, including the jitter feature, was irrelevant to demand? No, their witness to the contrary. Their CEO at Page Appendix, page 1603, admits that other features, he specifies jitter and he specifies soft start, which helps prevent overheating at startup. He specifies both of those at 1603, the CEO, as valuable to consumers. What were we supposed to think about the law of entire market value rule, which maybe ought to be called entire market value exception to the rule of apportionment? In the sense that we're always going to be in the context of a multi-component product, and there's a patented feature inside that larger product. It's only logical that every other component of that multi-component product has some value, some utility, some contribution, and some thing in there that consumers are going to want. And yet, nevertheless, we have this rule that says there are times when the patented feature can be deemed legally to be the basis for the consumer demand. So there has to be some room for allowance of fact patterns in which the entire market value rule will apply, even if other components, the non-patented components, are regarded as useful, important, etc. So can you help me think through what are the instances where a patented feature will qualify as serving as the basis for consumer demand? Yes, Your Honor. It's certainly something that you visited on the injunction standard with causal nexus. And let me try to answer your question here. There will sometimes be features that are the driver of consumer demand and warrant the application of the entire market value. Think of the pharmaceutical context. If there's a new patented feature which provides all the therapeutic value of a new drug, that will drive demand, apart from the form factor or the digest and tolerability of the casing. So I think there are certainly in the pharmaceutical instances, many instances in which you might find entire market value derived or attributable to the new patented feature. In multi-component products, it's going to be harder. But we think here, as in Lucent, you should reverse the denial of JMOL because here there was plain evidence of other features driving consumer demand. We've talked about two of them. There's evidence of jitter. They sued us on the jitter patents for the same products. That seems like a constructive admission that the 079 patent doesn't drive all of demand or else they wouldn't be suing us on the jitter patents for damages. And second, SoftStart conceded by the CEO to be an additional feature. So here there's evidence in the record of additional features. Right, but I guess I'm still stuck with trying to figure out, yes, there are other features in this multi-component product, just as there are in every multi-component product. And yes, those features are going to be important. And any time somebody says that a non-patented feature is important, that defeats entire market value rule? No, Your Honor, let's look at it the other way around. I think that the problem here, and let's go back to the one watt efficiency standard. The problem here is that it's simply not legally sufficient to say, oh, there's a government regulation that wanted more efficiency in government products. And that made the 079 the driver of the demand for power switching circuits. Think of an analogy. So if this were the only market, if the only market for these things had been in the government for some strange reason, then you would not be objecting, right? No, Your Honor. We would still object, and here's why. A government regulation that says miles per gallon for fuel efficiency needs to go up from 20 to 21. Suppose that leads to a new generation of cars that have 21 miles per gallon efficiency if people buy those cars. You use a car for a lot of other things. I'm sorry, Your Honor? You use a car for a lot of other things. What I'm trying to get at is if these chips, if the only person that was buying them was the government. The answer is no, Your Honor, because the government regulation could be satisfied by other methods, other products, other approaches than the 079 patent. We don't need to go that far. Your argument is that proof of driving demand in a submarket is insufficient for an entire market. We do argue that, but we argue that additionally here, even if the entire market were the government, the 079 wasn't proved to drive demand because you could reach the government regulations requirement through other methods, burst mode, for example. If you've got a valuable feature that's valuable to consumers, isn't that going to affect demand automatically? Why is it necessary to show something more than there are other valuable features to it which are perhaps unique to this particular product? If they're valuable to consumers, doesn't that mean that they affect consumer choices? The other features? Yeah. We think the fact that the other features drive consumer choices defeats entire market value. Yes, I understand, but what I'm saying is if the district judge here said that you have shown that there are other valuable features to the Fairchild products, and I don't understand that to be disputed. I'll ask about that. Correct, Your Honor. I don't understand that to be disputed, but why isn't the fact that there are other features which are valuable to consumers in and of itself be sufficient to defeat the entire market value rule? I believe that it would and does. There's just one more point. It's not that the other features exist. It's that it's conceded by power integrations that the other features like jitter and soft start drive consumer demand. That means that power integrations has failed to show that the 079 is the driver of consumer demand. What if that wasn't the exact testimony, that there are other features that drive consumer demand? What if it's just that there are other features that consumers would regard as important? Any power supply controller chip has to have certain components in order to serve its basic function, and so those would always be in any controller chip. Now there's this patented feature that's introduced into their controller chips that, save for the sake of argument, there was a paradigm shift in the marketplace. Everybody needed these one-watt energy-efficient controller chips, and that was just the new world everyone was living in. Then every controller chip has standard, routine, conventional, well-understood features. Now their chip has the standard, well-conventional features, plus this patented feature that everybody wants and needs now. Would that be enough for the entire market value rule? No, Your Honor. Not without changing the laws that was announced in Laser Dynamics, because Laser Dynamics, of course, said valuable, important, and essential is not enough. And, crucially said, even if practicing the disk discrimination method would be commercially unviable, it's not enough. So it doesn't matter if there—let me just go back briefly to the car example. If miles per gallon in a government regulation goes up from 20 to 21, and everybody buys the new, more fuel-efficient cars, it still doesn't show that the mileage efficiency feature drove the purchase of the car. No, but cars are $30,000, $50,000. We're talking about chips that, as I understand it, are less than $1, like $0.90 or $0.50. Correct, Your Honor, but where there were— The scale is way off. They were not infringing substitutes. It's admitted by the CEO, by the articles that they— It seems to me that there are three different boxes here. One is you've got a feature which is important to consumers and affects purchasing choices, and that's in the context of other features which may be valuable but are, as Judge Chen said, routine and conventional and this kind of device or chip or whatever. And under those circumstances, as I understand it, you admit that the entire market value rule would be appropriate. The other end of the spectrum is where consumer choices are directly affected by other features, but it seems to me there's a middle ground where you have a device where there are other important features which are unique to that device, in other words, that they aren't routine and conventional. They don't affect—they don't appear in every chip or every device of this sort, and that seems to be perhaps what we're struggling with here is that middle box as to whether that is a situation where the entire market value rule is appropriate. And what's your position on that? Well, if you're asking whether the 079 feature is unique— No, I'm asking a hypothetical. You've got a feature which is important to consumers, which is the patented feature, but you've got other features which are valuable, but there aren't just routine and conventional features that you would find in any embodiment of this device that's on the market. They're unique to the particular product that's being accused of infringement. And under those circumstances, is it appropriate to use the entire market value rule? No, because the other features defeat the notion that the patented feature is what's driving consumer demand. Here, for example, the— So you're seeking really an exquisite test, right? Basically, the patented feature has to be the sole reason for the purchase. Yes, Your Honor, and I'm just relying on laser dynamics that said there has to be proof that any one of those features alone drives the market, but if it's a sole test, then any evidence of another feature can't overcome it. You say, well, if you have to show you can't get entire market value unless you show the only reason why anybody could have possibly bought this product was for the patented feature. Well, Your Honor, I think that is correct because— If there are other features in the patent, you're never going to be able to make that showing, right? Well, Your Honor, I think there are times— Is that right? Isn't that right? You're never going to be able to make the showing under your test. No, I think you can make a showing, as I gave you the pharmaceutical example, where the therapeutic value of the patent is the driver of consumer demand, and nobody would credit trivial advantages about the casing for the pill as driving consumer demand, but that's quite different. Well, what happens if you had some evidence that somebody did? Somebody said, I switched to this one because it was sweeter. Your Honor, the— And that little bit of evidence is the one that destroyed the entire market value. Well, I do want to go back to— Right? I mean, I'm just trying to get to— It seemed to me that your position is extremely rigid. Well, Your Honor, don't take it from me. Take it from your decision in Laser Dynamics, which I think correctly states the law that goes back to Gerritsen v. Clark. Remember, there is an obvious legal error here, which is the court, the district court, thought that— Well, Gerritsen v. Clark didn't give us a lot of analysis, right? It didn't, Your Honor, but I think you cashed it out correctly in Laser Dynamics. Remember, the district court erred by saying that some intervening cases between, obviously, Gerritsen in 1884 and Wright-Hite in 1995, they say that Bose and Fonar and Tech Air show that laser dynamics is not the law. Let's just clear that out of the way. The application of the standard was the same. You can throw that out of the way, but we can't. You can, Your Honor, because the legal standard is absolutely the same in Gerritsen, Wright-Hite, Bose, Fonar, Tech Air. So Bose was a frolic and detour. It wasn't. Well, Your Honor, it was an— We throw Bose under the bus because of Gerritsen? No, you don't throw Bose under the bus. It was on different facts. There was not evidence, as there is here, that other features were important drivers of consumer demand. Remember, we were sued on jitter features for the same products. How can you come in and say, hello, I'm in this case saying the 079 feature is the entire driver of consumer demand? Pay no attention to that other lawsuit where we sued you on the jitter features in the same products. That is not a case for an entire market value rule. There may be another case where there isn't such an obvious competing feature that the patentee views as valuable because they're suing you on their other family of patents. Okay, I think we're about out of time here. I'm sorry, Your Honor, there's one last damages argument. Well, you can do that on the reply. We'll give you two minutes. Can you just flag what your other damages argument is going to be? Yes, Your Honor. Is it going to be the double-counting one? Double-counting. Okay. It's going to be double-counting, Your Honor. All right, that's fine. Thank you. Thank you for letting me speak to that on rebuttal. Thank you. Mr. Schickenbach. Thank you. Good morning, Your Honors. May it please the Court. I understand there's a lot of questions this morning. I actually intended to cover the client construction issue per second. I think that's an important issue to talk about, and then I'm happy to talk about EMVR in as much detail as I can. Well, let's begin with the 5% to 15%, which everybody seems to agree is due to environmental factors. Yes. And they say that they didn't agree that there was any room in the claim construction that the district court gave for a 5% to 15% variance based on environmental factors. What's your answer to that? Well, everyone below, the parties and the district court, understood that the issue was applying the claims as construed to the parts and whether people of skill and the art understood applying the claim construction, whether there was a fixed switching frequency or not. So I think this is a new argument on appeal that they're making. If you look at the transcript of the claim construction hearing, the judge actually tried to make very plain in his questioning of Fairchild Council, it was a different council, said, well you're not saying that normal environmental variation would take something outside the scope of the claim, are you? And so it's very clear what the district court understood. And in fact, the district court then invited, construed the claims, invited counsel, if anyone had any issues with the claim construction, wanted clarification, tell me. That's the way, in fact, that Judge Ware conducted claim construction and said, you know, I want to know this is a process. If you think I got something wrong, talk to me. Nobody did. And then during trial, again, this issue came up. And at that point it had become Judge Chesney. And Judge Chesney was at pains to say, look, no one's asked me to change this. You can't question a witness on the basis of some different claim construction. So the construction is what the construction is, and it's a question of application of the construction. And the last point on that is actually. I guess it's fair to say that the two expert witnesses had competing conceptions of what the construction really meant. Is that fair to say? And then they hinged their analysis based on their alternate constructions of the construction. They did, I think, with one important addition or footnote, if I might. Our expert, Dr. Kelly, explicitly applied the court's construction of a fixed number of cycles per second. Okay? And he explained in great detail all the parts infringed under that construction. Dr. Wee. What did he say about the 5 to 15%? He said that two things. Number one, these are, if you look at the data sheets, these are extreme ranges. So is there temperature variation? Yes. Maybe if you go from minus 30C to 150C, there might be some variation, theoretically. And the same thing with the input voltage varying wildly. You might get some variation. But what Dr. Kelly said is two things. Number one, you're still not going to have a varying number of cycles per second, according to the construction. No, but put aside. Okay. I think it's very confusing if we conflate these two separate arguments. One is the 5 to 15%. Yes. And the other one is the per second. And I'm just focusing now on the 5 to 15%. And did your expert say that 5 to 15% is within the normal variation due to environmental factors? Yes. Yes. And in addition on that point, if I may, there was no showing below by Fairchild that, in fact, in the real world, any of these parts is subjected to such temperature variation or such variation of the voltage on the input that there would, in fact, be a varying number of switching cycles per second. This is an entirely sort of theoretical argument by Fairchild without any evidence in the record to support it on this 5 to 15% issue. Okay. Let's turn to the per second. Do you agree that by asking for a claim construction that didn't include the per second limitation that they preserved their right to object to that? Not at all, Your Honor. And I think it's really important to look at what actually happened here at A2125. No, but first tell me why they don't preserve that by asking for a claim construction that didn't include the per second limitation. Because some of our cases seem to say that if you ask for a claim construction that is rejected by the district court, that you preserved it, and you don't have to go the extra step, for example, at the jury charge conference to objection to an instruction. Right. So let me take those in reverse order. That's the Pabst case, which is, of course, they're trying to embrace this Pabst case. The problem is the premise doesn't exist in this case. In Pabst, the party, in fact, asked for a construction without a particular limitation, made clear they objected to the inclusion of that limitation, and the judge put it in. And the question was then, well, is there a waiver if that party then doesn't subsequently renew the same position either when the jury instructions are read or otherwise? And the answer is no, no problem with that. But that principle doesn't apply here, because Fairchild never asked for a construction without per second. This is a little bit of a... Well, I don't understand how you can say that, because they did ask for a construction without per second. The question is whether they had to do something more to preserve that later on. What they asked for was that they argued about that fixed meant non-variant. But they never engaged on switching frequency. Their proposed construction did not include the per second limitation, right? True, but it just didn't address the issue one way or the other. Well, the question is whether proposing a claim construction that didn't include the per second limitation entitles them to argue later on that their construction should have been adopted and the per second limitation should have been omitted. But again, the relevant part of the construction they're challenging now has to do with what switching frequency means. No, they seem to be arguing also that it was error to include the per second limitation in the claim construction. They are, correct. But the district court actually gave two constructions below with per second. One was just switching frequency, and one was fixed switching frequency. They both had this per second term. And Fairchild did not propose a construction of the switching frequency part that omitted per second. Well, perhaps implicitly they did by saying the claim construction should be no variation at all. With respect, I would say that only goes to the fixed versus non-varying part of the construction. It doesn't say anything at all about what switching frequency means, which is you have to construe the entire term. But we have a lot more than that here. Is it possible that maybe what is really on appeal that's preserved for us to consider on the claim construction question is just the basic argument that fixed switching frequency has to be non-varying frequency, a very absolute fixed frequency. But any other more detailed arguments and criticisms about the introduction of the phrase per second is off the table because that wasn't ever specifically raised or explored down below. And so what's in front of us right now, whether it's the 5 to 15 percent or the per second language, is just the single thrust of whether the claim term requires absolute no variance at all. I think at a minimum what you said is correct, that certainly the most they could do is contend for absolute fixedness without any variation at all. But there is, if I could just finish this waiver point, because I think it is important. It's important because this issue never got vetted below. And the reason it never got vetted below is Fairchild not only didn't object, they embraced this claim construction. You look at the summary judgment briefing they filed. This is all on the record. You look at the summary judgment hearing transcript. You look at the J-Law brief they filed. You look at their J-Law argument. They said, we love this construction. This is a great construction. Well, wait. If somebody says the way this ought to be construed is X and the district court rejects that, my understanding of the cases is that you can then assume later on that you preserved your objection and that you can argue the case under the claim construction which you think is wrong. And where did they say, where did they give up the objection that they earlier made by proposing a claim construction that didn't include per second? Where did they give that up? They gave it up when they embraced the per second construction subsequently. Could you show me a page? Sure, absolutely. In their summary judgment motion, I'll give you two sites or three on each one. In the summary judgment motion, A3954. 3954. That's the motion itself. And then also at the hearing. Wait, wait, wait. Okay. 3954, line what? 3954. Well, you can begin. It's actually in three or four spots, beginning at lines five and six. No accused product includes a fixed switching frequency as that term was construed by the court. Well, that hardly says that they agree with the construction from the court. Where do they say they agree with the court's construction? Their argument is premised on the fact the court's construction is correct. Of course, but that's what happens all the time. You make a claim construction argument at Markman, you lose it, and you've got to live with it, and you've got to brief the case based on that. What I thought you were saying is they said they loved the construction and they approved it. Where do they approve the construction? They say that as well at the bottom of the page, beginning at line 25. There's a sentence, instead, the court agreed with Fairchild and construed the term switching frequency to be the number of switching cycles of the power switch per second and so forth. I mean, they're saying we were right, power integrations was wrong, the district court agreed with us, and I can give you five or six more spots where they did the same thing in arguing it orally and where they did it on JMA. It's the same issue. So this is not at all like the PAPS case or any other waiver cases they rely on. There was a different understanding as to what where's words meant, right? There was a dispute, exactly. The case was tried on the basis of an agreed construction. Nobody challenged the construction, and the experts disagreed over what that meant and how it applied, which is a government-related injury. The other side was saying right when they're going to the jury, this is a conference that was held on the 19th of February in front of Chesney right before Dr. Way testified. There was a colloquy back and forth. The other side is saying we absolutely are sticking to the claim construction. Without doubt, yes, they did. They embraced the claim construction, which is why Judge Chesney said on more than one occasion, nobody's asked me to change it, so that's the construction. And it's a question of how the experts apply it. And, in fact, that's one of the reasons the district court denied J-Mall, was because everyone agreed that it was a dispute among the experts. Okay, well, let's assume, I understand the argument that they didn't preserve the issue. Suppose they did preserve the issue. Why is the claim construction correct that would ignore variability within each second? Well, it's correct because as the district court observed when reaching the construction, you look at the specification and all over the place there are references to switching cycles per second, either in hertz or kilohertz. We actually collected these in the brief in 40, and the district court cataloged several of them and said this is the plain and ordinary meaning of the term. Yeah, but that's not the same thing as saying what your experts said, which is that it can still be fixed as long as it doesn't vary outside of an individual second. Why would it be a correct construction here to say that all sorts of unlimited variations within a second are permitted? Well, I guess you have to be careful with unlimited. The way the experts explain how this works is you can have some... I'm asking about the claim construction. Why is it that it's a correct construction of the claim to say that variations within a second don't count? Because that's the plain and ordinary meaning of the term. This is on a per second basis. Other than the plain and ordinary meaning, do you have any argument as to why it's the correct construction? Yes, because again, the specification repeatedly refers to frequency in terms of cycles per second, not microseconds. Yeah, but that's a definition of frequency measurement. It's not a definition of the variability that's permitted here. Well, but you have to... Okay, but you have to have some unit of time. You can't define frequency. I mean, I think it was suggested here this morning you could define frequency without reference to any particular time frame. That's a new argument, never raised below, and doesn't make any sense. Technically, it's not consistent with what the patent shows. The patent shows per second. And this is the way... There's a mountain of extrinsic evidence on this that's uncontested. Everyone in the field uses it in the same way. You pick up anybody's data sheet or application note, and that is the way the term is used. It is cycles per second. And so it would be incredible to construe the claim now to exclude that. I mean, you might say, well, it can include other things too. But it certainly, I think, would, on the basis of the intrinsic evidence, would have to include at least that, cycles per second. Ms. Sullivan says their original position at Markman, their preferred position, was no variation, period. That seemed to be correct. That was their preferred position. Yes. And she's now maintaining that their position now equates to the same thing, no variation, period. And so she said we're making the same argument, using different words at page 32 to describe no variation. So is the patent inoperable? No, Your Honor. Because here, again, the issue comes down to what fixed or non-varying means to somebody of skill in the art. I mean, the Chef America. Well, say, for example, we agreed with her that the claim should mean fixed, no variation whatsoever. Right. Not even natural variation, no variations at all. But it's the whole term. And I believe there was a question here earlier this morning to make that clear. It's not just fixed. It's fixed switching frequency. And the construction of that whole term is non-varying number of switching cycles per second. And so we get into a question of applying that to the accused products, an issue that never was reached in Chef America. It didn't have to be because you're— Well, what I'm trying to say is her interpretation would result in inoperability. That's true. And so, therefore, it doesn't fly. That's a pretty good reason not to adopt the construction. That's for sure. I mean, as Chef America laid out in some detail, it is certainly a last resort. I mean, if there is any plausible construction that avoids inoperability— Well, their own lawyer suggested an alternative. So clearly there's an alternative construction to what was there. Clearly, clearly. And, in fact, there was really no dispute about that until this appeal, frankly. Nobody thought for a minute that at least the environmental variation, to put it that way, was excluded by this claim construction until they filed their opening brief. So it's one of the many reasons we have a problem with the way this is being— I thought Dr. Way—is it Dr. Way or Dr. Wee? It's Wee. Is it Wee? Yeah. Wee. All right, if you say so. All right, Dr. Wee. We spent some time with Dr. Wee. Dr. Wee. Yes. I thought his argument was you don't ever count cycles in a second. That's not how you understand either the claim construction or the term fixed switching frequency. Full stop. So, therefore, that was his argument. It wasn't some acknowledgment of, hey, I understand and appreciate and am bound by a construction that allows for some natural variance. Again, correct with a footnote. He actually never applied the court's construction. He was asked about this, and he said it's not important given the way he viewed the claim. That's at A950. It's not important. He instead kept trying to come back to this driving analogy, which counsel tried again. I guess my point is that in this trial, there wasn't any argument over what per second meant as a claim construction question. It was always dealt with on a trial factual level of how to understand and apply the court's construction. Absolutely. That's right. And that was a dispute between the experts, and, of course, the jury found in power integrations favor on that. Shelly, I see I'm a little over time. Go on to EMVR. Yeah, and the court here at 1764 said there's evidence in the record that other features are important and are highlighted by the respective parties. So we have a starting point here. You don't disagree that the evidence was that there were other features in the Fairchild product which were valuable? Don't dispute that at all. Just like every other EMVR case that this court has faced, you don't even get to EMVR, of course, unless it is a multi-feature product. And those features here were unique to the Fairchild product? No, they weren't. They were not? No, actually, the two they point to regularly are jitter and soft start. Not unique. In fact, those were both prior power integrations features. They were in the prior part, the FX. So getting to Ms. Sullivan's earlier argument about frequency jitter is something they've already had to pay for in an earlier patent infringement action. Yes. That obviously suggests that the frequency jitter feature of these controller chips are pretty darn important and valuable because they had to pay several million dollars on those for that particular feature. Had some value. Yes. How much? Do you recall what the judgment was for? Well, the unfortunate reality is they haven't paid anything for any feature so far. Well, what was the damages verdict on that one? Well, in the first case, the damages verdict was $32 million premised not entirely on jitter, largely on the basis of the jitter features, the 876 and the 851 patents. But that was also premised on a third patent having to do with device structure. So you can't really tease out a specific number. I guess the point is that they are liable, apparently, for several million dollars on the frequency jitter feature of these Fairchild controller chips. So why isn't that create some doubt, maybe considerable doubt, about whether we can really hold that all of the value of these controller chips are legally attributable to the 079 patented technology? So this, I think, is the Fonar case, right? So an oldie but a goodie where there was multiple patents in that case. One of the patents in particular was found to have created the basis of customer demand, EMVR applied. There was one of the other patents had value. There was a separate damages award for that other patent, and the two coexisted. And I think it's a necessary, a logical conclusion of what EMVR means. But that creates a situation where you could have overlapping damages between two cases, one feature in one case, one feature in another case. And in one case, you apply EMVR and you recover for, on the entire market value theory, a value that's attributable to another feature where this should have been apportionment. And if I could just tag along on that. Isn't this a royalty stacking problem? You've got a lot of different patents on a given multi-component feature or multi-component product. And now manufacturers are having to pay out royalty on this, royalty on that, royalty on the other thing. And, oh, now we've got to pay entire market value rule for this particular patent. Isn't there some kind of problem there, a disconnect? I don't, certainly not in this case and on this record, right? I mean, this was never raised. This is another argument, this whole reference to the jitter verdict and the fact they lost this other case and may have to pay for that. Never raised below, never vetted, no discussion about it. So there's nothing, zero in the record on that. But just hypothetically, speaking hypothetically, could there be a case where there'd be a royalty stacking issue? Sure, maybe. And in a case where that's a real issue, the party who thinks it's an issue raises it and it gets vetted and the experts take it into account to ensure it doesn't happen. That was not the situation here. And one of the reasons, with respect, I think it was not the situation here is because of the uniqueness of the facts of this case. This is not sort of some runaway train on EMVR. You have a patent in front of you. Was one of the other features that was identified as valuable here the jittering feature? Yes, it was. It was in the parts and it had some value and we don't dispute that. There's soft start, frequency jitter. Soft start and frequency jitter are the- Milliwatt, saver, safety features. Well, that's a sort of an umbrella term, milliwatt, that refers to a sort of a collection of features in the chip. But this goes back to, I think, some of the earlier questions. You know, you have to have a lot of parts to have a car or any particular device. But if one of those features is so important, it's the reason people buy the device in the first place, that's an application of EMVR that's totally appropriate. And in this case, we had unique facts that indicated that was the case. One point I think that got lost in the questioning, I just wanted to emphasize, I know I'm way over time. It's not just a question of the government standards because that came up. The government standard was critical. It drove not just government purchases, but there was undisputed testimony. It drove the whole market. And here's the kicker. There was undisputed testimony from our expert, Dr. Kelly, that one could not meet those standards without using this patented feature. You had to have this patent to meet those standards. Burst mode could have done it. It just would have been noisy. Well, that's right. And the evidence is that no one would use burst mode after the 079 came along. Nobody used it. Fairchild never used it again or any other alternative. Power integrations never used it again. And as far as we know… Well, we were still making sales on the FX, right? It wasn't zeroed out on the graph. In applications where it was already designed in, right? I mean, this was an issue that Judge Chesney wrestled with. She really wanted to understand the facts around that. And so you had a large customer who designed in the FX before GX came along, Motorola, and it was a big cell phone charger program. And you buy those chargers as long as those phones are supported. It can be a long time. So sales don't go to zero. But the critical consideration is, what is driving purchases of the new parts for the new design-ins? And on this record, it's undisputed that once GX came out, there was not a single design-in for the FX part. It absolutely disappeared overnight. Like the Bose case. We could go on forever. This is fine. I think we won't go on forever. Thank you. I want to make sure the other questions are answered. Thank you. All right. Ms. Sullivan, you have two minutes. Two points only, Your Honor. I have one word to leave you with on the seemingly inoperable claim construction of no variation whatsoever. ZOW. The ZOW prior art reference was the problem here. And I would just refer, Your Honors, to the pages where the ZOW reference appears. That's at Appendix... And there you'll see the problem. The problem is ZOW had not a fixed line, a horizontal line, but a rising line for the frequency variation in the first range. To overcome it, if you look at Appendix 3260, the power integration had to make the argument, no, no, we're fixed, meaning non-varying. So we overcome ZOW. That's the reason why they made what sounds perhaps like an inoperable claim construction. But second, Your Honor, they waived any objection to our claim construction on non-varying. They waived it because they haven't challenged it on appeal. So you don't have a Chef America Ecolab problem here. The construction to circumvent ZOW is their problem, not ours. If it renders the patent invalid for lack of enablement, that's their problem, not ours. Did you argue ZOW below? Absolutely, Your Honor. We did. Where? Well, I'm seeking that page, Your Honor. Can I just finish on double counting so I don't lose all my time with you? Because I promised to come back to you on double counting. The argument is simple. If you take Dr. Putnam, the damages expert's three buckets, they appear in his demonstrative, reproduced on page 20 of the Blue Brief. He did count, in violation of Azutech and Lucent, lost profits and royalty on the same infringing acts. He double counted lost profits and additional royalty on the same infringing acts. If you look at the price erosion bucket that's depicted here on the left, price erosion on power integration sales is a result of all sales of Fairchild's infringing products. So he's counted a price erosion lost profits damages measure based on all our sales. He then, on the right-hand bucket, the $47.1 million for the supposed license fee on the additional Fairchild units that Power Integrations wouldn't have sold, he counts it again. He says they infringed and you get a license. That's the double counting, Your Honor. Okay, I think we're about out of time. Do you have the reference or do you want them to submit it? Your Honor, I don't have the reference. I can't answer your question affirmatively right now. I don't recall you answering the waiver argument the red brief made in your gray brief, which is why I was wondering, did Briseau ever actually put it up? Your Honor, we think it's there in the patent history. We're not relying on it. It's explanatory of why there may be a seemingly inoperable construction. Thank you for giving us all the extra time on the argument. I respectfully suggest that you reverse. But certainly this is not a case to amend entire market value rule law. This was a case for a portion of it. I think we're out of time. Thank you. Thank both counsel. The case is submitted.