CONSTELLATION DESIGNS, LLC v. LG ELECTRONICS, INC.

Authored by: Jeremy J. Gustrowsky

A recent Federal Circuit decision drew a sharp line between claims that recite an optimized outcome and claims that recite the specific thing the inventors actually built. The technology involves digital television broadcasting, specifically the “constellations” that communication systems use to translate digital bits into analog signals and back again. The inventors at Constellation Designs departed from the conventional wisdom of maximizing the distance between constellation points. Instead, they spaced the points unevenly and optimized them for something called parallel decode (PD) capacity, which compares information entering the transmitter’s mapper with the information leaving the receiver’s demapper.

Constellation sued LG Electronics in the Eastern District of Texas over televisions compatible with the ATSC 3.0 over-the-air broadcast standard. The district court granted summary judgment that all asserted claims were patent eligible, and a jury then found willful infringement of nine claims across four patents, awarding roughly $1.68 million based on a royalty of $6.75 per television. LG appealed on eligibility, infringement, and damages.

The Federal Circuit split the claims into two groups. The first group, from U.S. Patent Nos. 8,842,761 and 10,693,700, recited a constellation “optimized for capacity using parallel decode capacity” that delivers a given capacity at a reduced signal-to-noise ratio. The court found these ineligible. The claims describe a desired result without describing how to get there. The specification includes an iterative process for selecting constellation points, checking whether the constellation size supports the desired capacity, and looping until the input and output signal-to-noise ratios converge. None of that made it into the claims. Drawing on O’Reilly v. Morse and ChargePoint, the court explained that the claims cover every way of optimizing a constellation for PD capacity, and details in a specification cannot rescue claims that do not recite them. At step two, the alleged inventive concept was the abstract idea itself, which is not enough. Evidence that the patents survived obviousness challenges did not substitute for an inventive concept evident in the claims.

The second group fared much better. Claims from U.S. Patent Nos. 11,019,509 and 11,018,922 recite specific non-uniform constellations rather than an optimization goal. Claim 21 of the ‘509 patent, for example, requires unequally spaced unique point locations, distinct labels for each point, and the counterintuitive feature that at least two constellation points occupy the same location. The court held these claims are a concrete implementation of a technological solution to a technological problem, so the analysis stopped at Alice step one. The lesson for drafters is direct. Reciting the structural result of your discovery is safer than reciting the objective you were pursuing.

On infringement, LG argued that a patentee relying on an industry standard must show the standard meets every limitation of a claim. The Federal Circuit disagreed and endorsed a limitation-by-limitation approach. A patent owner may prove some limitations through standard compliance and others through direct evidence about the accused products, so long as the Fujitsu safeguards are satisfied for each limitation relied upon. The relevant portion of the standard must be specific enough that practicing it always results in practicing the limitation, and it must be mandatory or shown to be implemented. Substantial evidence, including expert testing, source code, LG’s internal documents, and testimony that the A/322 protocol is mandatory for ATSC 3.0 receivers, supported the jury’s verdict.

The damages challenge failed largely on procedure. LG framed an admissibility objection as a sufficiency-of-the-evidence motion, and the court agreed with the district court that this was the wrong vehicle. Treating the argument as presented, the panel found ample support for the damages expert’s reliance on comparable Zenith licenses and a built-in apportionment theory, given testimony on both technical and economic similarities. LG’s argument that a partial eligibility loss should undo the damages award came too late, raised for the first time at oral argument and therefore forfeited.

The result is a partial vacatur. The optimization claims go back to the district court, the constellation claims remain eligible, and the infringement and damages rulings stand. Notably, this opinion was modified and reissued following a petition for rehearing, underscoring how close the panel considered the eligibility question on the optimization claims.