TQ Delta, LLC v ZYXEL Communications Ltd & Anor

In patent construction, a technical term must be construed purposively and in the context of the patent. Where the patent’s technical objective can be achieved only by one meaning, that meaning should generally be preferred. A claim is not rendered obvious merely because prior art addresses the same general problem. The court must identify whether the prior art actually teaches the claimed technical solution without hindsight.

The skilled person is deemed to read and assimilate pleaded prior art properly, including lengthy incorporated documents, but does not approach it with knowledge that it contains the solution. An invention may lie in perceiving a technical problem, although that will be unusual where the skilled person regularly encounters and solves such problems.

TQ Delta brought a technical patent claim against ZyXEL concerning two DSL patents: the 268 Patent, relating to variable overhead allocation, and the 430 Patent, relating to CRC anomaly counter normalisation. The dispute concerned construction, validity and infringement.

For the 268 Patent, ZyXEL relied principally on the Alabama proposal as prior art. For the 430 Patent, ZyXEL relied on the ADSL2/02 standard and argued that the claimed normalisation mechanism was obvious once the alleged problem had been perceived.

The court therefore had to determine the meaning of key claim terms, whether either patent was obvious, and whether the patents were essential to relevant DSL standards.

1. 268 Patent—construction. The term “frame” in Claim 1 meant a mux data frame. Its meaning was context-dependent, and the patent’s technical objective could be achieved only at the mux data frame level, before scrambling and interleaving. The term “specifies” was clear and required no paraphrase. The reference to n_max was non-limiting, whether treated as a reference sign under Rule 29(7) of the Implementing Regulations to the European Patent Convention or construed more generally.
2. 268 Patent—validity. Alabama addressed the same broad problem of excessive overhead at low data rates, but retained overhead at the beginning of each mux data frame. It did not teach a parameter specifying which mux data frames contained overhead and which did not. Treating Alabama as an equivalent solution was hindsight reasoning. The 268 Patent was valid.
3. The alternative obviousness attacks were not fairly available because they were advanced only in closing submissions and raised technical issues not addressed in the evidence. In any event, they failed on construction, scrambling and the absence of an identified qualifying parameter.
4. 268 Patent—infringement. The patent was essential to the relevant standards. In VDSL2, the parameter G_p performed the required function where G_p was less than T_p. The 268 Patent was therefore infringed.
5. 430 Patent—construction and infringement. “Normalising” did not require CRC anomalies to be calculated and reported consistently for every connection. ZyXEL’s narrower construction was unsupported by Claim 1. If valid, the 430 Patent would have been infringed because it was essential to the relevant standards.
6. 430 Patent—obviousness. The skilled team included a person concerned with ISP error reporting and management. Such a person would properly assimilate ADSL2/02, focus on the mandatory SES reporting provisions, and recognise that the variable PERp range created inconsistency in the significance of SES reports. Once that problem was appreciated, the claimed normalisation was an obvious solution. The 430 Patent was invalid.
7. The evidence did not establish a longstanding unresolved problem sufficient to support inventiveness. The 15–20 ms range was not regarded as requiring normalisation, and the later widening of the range was accompanied by the introduction of normalisation.