CAFC Punts on Court vs. USPTO Validity Determinations

(1) In Re Translogic Technology Inc. (2006-1192), October 12, 2007

(2) Translogic Technology, Inc. v. Hitachi, LTD, et al. (2005-1387) (nonprecedential), October 12, 2007

This case illustrated the ongoing tension between district court proceedings and parallel USPTO reexaminations, where each proceeding ended with a different result. In Translogic, the district court, after waiting 3 years on 5 reexamination requests, decided to reopen the case and found the patent-in-suit not invalid and infringed, and levied a $86.5M judgment and a permanent injunction (which was later stayed). In the meantime, the PTO found the patent obvious and invalid - for more background info, see earlier 271 Blog post here.

Despite the PTO's findings, the district court stuck to its guns and maintained its ruling, noting that "[t]he Federal Circuit is the proper forum to resolve conflicts between this court's claim construction and the Board's claim construction." One of the primary questions presented on appeal was "[s]hould the judgment of the district court be vacated and the case remanded for summary dismissal because the claims of the '666 patent are unpatentable on reexamination and the patent is void ab initio?"

To answer this question, the CAFC turned to BPAI proceedings, and found that the patent was invalid for obviousness. One of the sticking points was the interpretation of "coupled to" and "coupled to receive" - Translogic argued that the BPAI should have followed the district court's interpretation of "coupled to receive" to mean "connected to receive . . . directly or through one or more intervening inverters." The BPAI interpreted "coupled to receive" to mean "capable of receiving" signals.

Interestingly, the CAFC didn't base its decision on a "broadest reasonable interpretation," but instead engaged in a Phillips analysis of the claims, and found that the BPAI was correct:

The term "coupled to" in the phrase "second stage input terminal 'coupled to' the first stage output terminal" defines a connection between the TGMs. In other words, "coupled to" in the context of this claim phrase defines the connection between two of the TGMs essential for a series multiplexer.

On the other hand, the term "coupled to receive" in the phrase "input terminals 'coupled to receive' first and second input variables" does not specify a particular connection. In other words, the claimed circuit does not require any specific input or connection. Instead, like any logic circuit, this part of the circuit only accepts inputs from an external source. As such, "coupled to" and "coupled to receive" are clearly different in the claimed '666 patent multiplexer circuit . . . Even if the input terminals are not connected, the circuit claimed in the '666 patent defines a series multiplexer. Therefore, this court agrees with the Board's construction that "coupled to receive" means "capable of receiving."

On the issue of obviousness, Translogic argued that the prior art (Gorai) was not applicable, because it pertained to an algorithm to design logic circuits based on functional parameters, while the patent-in-suit improves a known circuit (i.e., multiplexer). The CAFC rejected this argument:

In its prior art argument, Translogic is making the same error corrected by the Supreme Court in KSR. Translogic mistakenly argues that variants of a circuit connecting 2:1 multiplexers in series are not relevant prior art with respect to the '666 patent because these variants do not address the same problem, namely an improved multiplexer circuit. However, this argument overlooks the fundamental proposition that the series circuits in Gorai are prior art within the public domain and the common knowledge of a person of ordinary skill in the art. Thus, the Gorai reference is a relevant prior art reference with respect to the '666 patent and clearly discloses a series 2:1 multiplexer circuit.

[C]laims 47 and 48 of the '666 patent define a series multiplexer circuit. Gorai Fig. 3 discloses the same serial multiplexer circuit. As any person of ordinary skill in the art would understand, the inputs to a circuit do not change the circuit itself. Therefore, this court finds that Gorai discloses a series multiplexer circuit as claimed in the '666 patent. Thus, the Board was correct in concluding that the Gorai reference discloses a series multiplexer circuit.

Regarding the question "[s]hould the judgment of the district court be vacated and the case remanded for summary dismissal because the claims of the '666 patent are unpatentable on reexamination and the patent is void ab initio?", the CAFC issued a short statement, answering this question in the affirmative:

In light of this court's decision in In re Translogic Tech., Inc., this court vacates the district court's decision and remands this case to the district court for dismissal.

When Art. III and APA Judges Disagree on Validity, Who Wins?

Assume the following facts:

1. Patentee sues XYZ in district court for infringement.
2. XYZ files multiple reexamination requests in the PTO seeking to invalidate patent.
3. District court gets tired of waiting for the PTO after 3 years, and resumes litigation.
4. In the meantime, PTO finds patent invalid, and patentee appeals to the BPAI.
5. District court finds patent not invalid and infringed by XYZ.
6. BPAI affirms invalidity, XYZ files motion to grant JMOL in district court.
7. District court refuses to recognize the PTO's findings and enters judgment in favor of patentee.

This is the situation facing the defendant in Translogic Technology v. Hitachi, (D. Oregon, No. Civ. 99-407-PA). Reading the court's opinion from Dec. 2005 (404 F.Supp.2d 1250), you can see the tough position the court was in:

Plaintiff filed this action in March 1999. Because of the reexamination proceedings that began in July 1999, in September 1999, this court granted defendants' motion to stay.

In December 2002, the court lifted the stay of litigation because the reexamination proceedings appeared likely to continue for years. This court determined that plaintiff deserved its day in court, despite the possibility that the reexamination proceedings might some day produce rulings that were inconsistent with this court's rulings. The Special Master reviewed the reexamination file up to that time in writing his report on claim construction. This court held the trial on validity in October 2003.

Meanwhile, reexamination proceedings proceeded at a somewhat slower pace. Cf. 35 U.S.C. § 305 (reexamination proceedings must be conducted with "special dispatch"). The PTO granted the request for reexamination in July 1999. Six years and five reexaminations later, the Board issued the claim construction on which defendants now rely.

If a district court was obliged to alter its claim construction to conform to a later claim construction issued by the Board, then there would be no reason to proceed with litigation while reexamination proceedings were pending. District courts are not, however, bound by the Board's rulings, and district courts are not required to stay litigation under these circumstances. See Vitronics Corp. v. Conceptronic, Inc., 36 F.Supp.2d 440, 441 (D.N.H.1997). The Federal Circuit is the proper forum to resolve
conflicts between this court's claim construction and the Board's claim construction.

Thus, the case is now before the CAFC (oral arguments were heard April 3, 2007 - listen to the arguments here). One of the questions presented is:

1. Should the judgment of the district court be vacated and the case remanded for summary dismissal because the claims of the '666 patent are unpatentable on reexamination and the patent is void ab initio?

This will be a very important decision that affects the relationship between the USPTO and the district courts when defendants challenge validity in parallel proceedings. This is also the type of situation the Patent Reform Act of 2007 is trying to avoid by terminating reexaminations when a "judgment by a district court" is rendered (as opposed to a "final decision").

Download/view Hitachi (Appellant) brief here.

Download/view Translogic (Appellee) brief here.

Similar Problems in the EPO

Opposition Proceedings at the EPO are often not concluded until six or more years after a patent has granted. Recently in the case of Unilin Beheer BV v Berry Floor NV (link) a UK court found a patent valid and infringed, while the EPO found the patent invalid. The UK Court of Appeals held that the UK Court trumped the EPO:

"If the patent is revoked, the way is cleared; if it is upheld and held infringed then compensation will be payable for past acts. And an injunction will run unless there is a later revocation by the EPO. Subject to that last point, the effect of all this is that one does not have to wait to find out who has won until the slowest horse in the race gets there . . . [It does not] help to ask whether a national court or the EPO is 'top'. It all depends on the circumstances, as the two following scenarios illustrate: the patent is still under opposition when a national court holds it valid and the EPO then revokes. So the EPO is 'top'. [Second,] the EPO holds the patent valid and a national
court subsequently revokes it (there is no estoppel created by an EPO decision as to validity […]). So the national court is 'top' . . . In truth asking which tribunal is 'top' is simply not helpful – there is just the untidy compromise inherent in the EPC and one which cannot be properly resolved unless and until a rational patent litigation system for Europe is created."

See Out-Law.com: "Patent damages not refunded if EPO cancels patent" (link)