Wednesday, September 19, 2018

Qualcomm switch patent asserted against Apple in Germany should be invalidated, Swedish patent office says

Yesterday's Qualcomm v. Apple trial took twice as long as the average Mannheim patent trial. In fact, the ventilation system was switched off in the late afternoon, so for the last hour, two doors had to be kept open. The courtrooms at the Mannheim Regional Court, Europe's leading venue for wireless patents, are famously windowless.

Presiding Judge Dr. Holger Kircher forthcomingly stated at the outset that this case was, in my words, too close to call (unlike the one that Qualcomm agreed to stay in June), thus the court had to elaborate on all our of Apple's defenses: non-infringement, invalidity (which German district courts don't determine, but they can and often do stay cases pending a parallel nullity or revocation proceeding in another forum), abusive conduct (antitrust), and licensing (through one or more contract manufacturers). I'll address the first two--the traditional defenses to patent infringement--in this post, and the affirmative defenses (the remaining two) in a subsequent post since there's an abundance of interesting things to report and comment on.

The patent-in-suit, EP2460270 on a "switch with improved biasing" ("biasing" in this context basically meaning that one voltage gets to control another), is not standard-essential. Essentiality hasn't been alleged by any party to the German Qualcomm v. Apple cases that have been heard so far. Nor is it related to wireless baseband processors: it's a general circuity patent covering a type of switch. It was mentioned during yesterday's trial that the chip allegedly infringing on the patent is supplied to Apple by Avago/Broadcom. But all of the accused devices come with an Intel baseband chip, a fact that will be relevant to the antitrust part of the next post.

In the combination of Judge Dr. Kircher's summary of the issues in the case and what Qualcomm's German counsel, Quinn Emanuel's Dr. Marcus Grosch, said yesterday, it turned out that Apple originally didn't even deny infringement but focused on invalidity. Then, after Qualcomm's reply in support of its complaint, Apple presented a non-infringement theory in a sur-reply. It wasn't said whether Qualcomm's reply brief or a change of mind triggered this, but normally a defense has to be presented in the defendant's first pleading, so I would attribute this to something Qualcomm said. This much is certain: the non-infringement argument is effectively a disputed construction of a claim limitation, not a factual question concerning the accused products.

As far as I understood, Apple argued that the patent-in-suit covered a certain chipset architecture involving a constant bulk voltage, with ground also being constant. But Qualcomm disagreed, arguing that "AC ground" stood for only the AC part, not DC. In any event, the court appeared less inclined to agree with Apple. But... and this was something very special:

Judge Dr. Kircher, who unlike German criminal judges does not have a duty to "investigate," felt the parties had not focused enough on an inconsistency between claim 1 (an apparatus claim; Qualcomm is asserting one or more apparatus claims) and independent claim 16, a method claim that refers--unlike the apparatus claim(s)--to only "the" (a single) control signal in connection with what turns on the switch, resulting in an input signal being passed through a plurality of transistors. The accused chip, however, has two control signals and would therefore, at least potentially, fall outside the claim scope.

I was profoundly impressed and still am. Apple v. Qualcomm is the commercially biggest patent-related dispute ever (at a minimum, the biggest one in the history of the mobile device industry). This patent is just a limited part of the wider dispute, but still this (unlike what Judge Alsup once mislabeled Oracle v. Google) is truly the World Series of IP. Yesterday one could see sky-high piles of case documents behind the bench. Still, with all that had been written in the build up to the trial, the presiding judge identified an issue that has the potential to be outcome-determinative and had not received much if any, attention.

It's not just because of one demonstration of extraordinary competence and dedication that I couldn't imagine a future Unified Patent Court (UPC) without this judge, despite fundamental disagreements in the past and the present with respect to the bearing that antitrust issues should have on patent infringement cases.

Unsurprisingly, Qualcomm argued that the distinction between one or two control signals was an irrelevant labeling question. I lack the expertise in electrical engineering to form an opinion, but I've watched numerous patent cases in (mostly) two major jurisdictions and this is the typical difference between an accused product and a properly-construed patent claim that routinely becomes a defendant's "get out of jail free" card. In fact, the world is littered with complaints that have been dismissed for lesser reasons.

However, Judge Dr. Kircher said that case law and literature were inconclusive as to whether or not a parallel claim could help a court construe another claim. Apparently, reasonable and reputable people disagree on this legal question because it hasn't been resolved. Better late than never. Here, it's a given that Qualcomm would appeal a stay by "complaining" (that's the term used here) to the Karlsruhe Higher Regional Court.

For the development of German (and, by extension, European) patent case law it would actually be good if the Mannheim court was prepared to take the risk of being overruled on appeal. I have no idea how the appellate panel under Presiding Judge Dr. Andreas Voss ("Voß" in German), who previously presided over another patent infringement chamber in Mannheim, or, in the ultimate consequence, the Federal Court of Justice (both in the city of Karlsruhe) would decide. My guess is that neither would pronounce a bright-line rule declaring it categorically impermissible to get mileage out of something that is found in the patent document itself. Why rely on descriptions, which aren't even updated as the claims are narrowed (which is also a huge issue in this case, with Apple alleging that Qualcomm pointed to a drawing that relates to something that the original application covered but the actually-issued patent does not), but not on claims, which the examiners focus on? Much less on anything found outside the patent document? It would probably the most likely outcome that some weight (within reason) would be given to what can be inferred or deduced from parallel claims, and in a case that is too close to call, it doesn't take too much to tip the scales.

It's fair to stay that no one in the room appeared to have a better idea of the issues in the case and the evidence in the record, across the board, than Judge Dr. Kircher. But in connection with invalidity, something apparently entered the record rather late, and while the court initially showed an inclination not to grant Apple's motion to stay since it's based on lack of an inventive step, while German courts normally grant stays only if there's a strong case for anticipation. The aforementioned Judge Voss ("Voß") always reminded counsel that a stay required a prior art reference constituting a "photographic image" of the patent-in-suit.

That's because German patent infringement courts don't like to take a position on whether a person of ordinary skill in the art would be likely to combine two or more prior art references: they're jurists, not engineers, unlike patent examiners.

Yet the hurdle for getting German patent infringement cases has, fortunately, become lower and lower over the years, as the regional and higher regional courts saw all too many cases in which they issued injunctions and later the patent died or was defanged as its claims got narrowed.

The court agreed with Apple that a new prior art reference (with an Asian name I didn't fully understand, but I'll find out whenever I find the time to request access to the nullity file), coupled with one named "Brindel" or "Brindle" that had been considered by the examiner in connection, represented a relatively stronger case for invalidation than one that the examiner found wasn't likely to become part of a combination. Dr. Grosch's efforts to deny or downplay this didn't persuade the court.

Freshfields Bruckhaus Deringer's Prince Wolrad of Waldeck and Pyrmont argued Apple's invalidity contentions. Toward the end I got the impression that the court was potentially having second thoughts. Just potentially, but that means a lot in a case where the original position was that a stay would not be warranted here.

The potential silver bullet here is that Apple had submitted the original patent and its proposed combination of prior art references (which are, according to Apple, also at the heart of its German nullity complaint pending with the Munich-based Federal Patent Court) to the Swedish Patent & Registration Office ("PRV"), and received an expert opinion from an examiner according to which the patent-in-suit does lack an inventive step in light of the prior art.

I haven't been able to find that opinion on the website of the Swedish patent office yet, but I'll keep trying.

Judge Dr. Kircher asked a key question: how many sucb opinions had Apple requested before obtaining this one? Apple's counsel, under the usual obligation to speak the truth, said it was just this one, and stressed that Apple wanted a neutral opinion based on just the prior art and no interaction with the examiner himself. That's the way it works in Sweden according to Apple. The Austrian and Swiss patent offices were also mentioned as providers of such opinions, but one of them no longer offers that service and the other one lets the petitioner communicate a lot with the examiner, which Apple's lawyers didn't want to do because they have a stronger case based on an uninfluenced opinion.

Apple didn't say so, but I might add that Sweden is a country with a lot of expertise in this field of technology, especially because of Ericsson.

It was again no surprise that Qualcomm's counsel would be dismissive of Apple's position. Dr. Grosch said that it would obviously simplify all patent cases if one could just fill out a web form somewhere. And it's not like the court indicated a change of mind. But I got the impression that Judge Dr. Kircher may indeed take the independent Swedish evaluation seriously. Maybe--just maybe--this does constitute a basis on which the court would make one of those rare exceptions in which a patent infringement case is stayed over inventiveness, not novelty, concerns.

There'll be a formal post-trial brief for which Apple was granted leave, and a legal brief (no new facts) which Qualcomm announced. The court will announce a decision on December 4, 2018. That one could be a final ruling, a stay, or whatever other procedural order.

Share with other professionals via LinkedIn: