Magnacom emerges from stealth mode today, claiming it has a new twist on digital modulation technology that provides big gains for any wired or wireless network. The company will license its technology on which it already has 14 US patents, all granted in less than a year.
The startup will demonstrate in a private suite at the Consumer Electronics Show an FPGA board using its technology to deliver a 10dB signaling advantage compared to QAM4096, the most powerful version of the quadrature amplitude modulation scheme widely used in communications today.
The improvement could be used to cut power or spectrum requirements in half or to send data up to four times further than with the existing QAM approach, the startup claims. It requires changes in hardware that use less than a square millimeter of silicon in a 28nm process, it said.
The company’s so-called wave modulation technology “is practical and it’s fairly surprising to people who have been in the field that you can get these kinds of gains,” said Jason Trachewsky, a communications expert and serial entrepreneur compensated by Magnacom to review the technology.
“Some of the concepts are familiar — no one invents anything from thin air — but it is a pretty novel approach, focused on a computationally effective way to get capacity gains on non-linear channels,” he said.
The technology was developed by Amir Eliaz, the startup’s CTO, who was vice president of R&D for Provigent, a wireless backhaul startup acquired by Broadcom. At a conference in Israel, a mutual friend introduced Eliaz to Yossi Cohen, an engineer who rose to executive positions at National, Broadcom, and Motorola Mobility.
Cohen was seeking new opportunities after Google acquired Motorola Mobility and installed new management. Eliaz wanted to share his work on modulation.
“I intended to meet with him for an hour — there was a fireside chat with Israeli President Shimon Peres coming up at the conference — but I ended up staying four hours,” said Cohen who became co-founder and chief executive of Magnacom. “My friends tease me that I gave up a meeting with the president of Israel,” Cohen said.
“I was skeptical initially due to the claims [Eliaz] made — he came across as a scientist,” said Cohen who admitted he had “a tough time assessing” the technology.
Cohen hired a number of technologists including Trachewsky, a former Broadcom fellow, to review the technology. “I was honestly pleasantly surprised it passed the due diligence, [and] I gave the founders a loan so we could get going right away,” Cohen said.
The group formed Magnacom and closed a round of funding, mainly from Cohen’s business contacts, to fuel its first couple years. Cohen decided the startup needed working silicon to show other skeptics and hired a legal firm to nail down patents on the technology.
An Altera board is now running in Magnacom’s Israeli lab. A team of seven lawyers at McAndrews, Held & Malloy Ltd. helped guide Eliaz through filing in January 14 US patents using a relatively new fast-track service. They were all issued over the last few weeks.
“It’s pretty amazing,” said Trachewsky who has nearly 100 patents, the fastest of which took two years to be granted. “You have to be careful on how you file, what you are covering, how it’s novel, and what’s the prior art,” he added.
The US patents cover areas such as receiver design, equalization, forward error control, and error handling. The startup filed more than 50 other patents in agencies around the world and expects licensing to be worked out on a case-by-case basis.
“We want to be perceived as charging fair value or even less than that — we want this technology to catch like forest fire,” Cohen said.
Trachewsky describes wave modulation as “deliberately introducing correlation between signals — which sounds like a bad thing but is not — [because] dependence between transmitted instances and time [helps] shape and compress spectrum effectively [and] it isn’t distorted much by non-linear elements,” he said.
Part of the innovation is a novel receiver design “that pulls apart [the] dependency [between transmitted data sets] and extracts the information stream,” he said. “You have to do it exactly right not to increase capacity — that’s part of it too,” he added.
In its press materials, Magnacom described its technology as a “multi-dimensional signal construction operating at the Euclidean domain… breaking the orthogonality of signal… to increase capacity and provide an optimal handling of nonlinear distortion.” It “uses nonlinear signal shaping… digitally at the receiver side,” it added. The approach is backwards compatible with QAM.
A big hurdle for Magnacom is that new communications technologies such as modulation schemes need to be standardized before they are deployed. The process can take many years and force disclosures that help competitors catch up.
Thus Magnacom is likely to try to get a foothold in significant but smaller markets such as point-to-point microwave backhaul or satellite links. Here what Trachewsky called “the standards burden” is not so great.
Ultimately the big win for Magnacom is to get cellular and WiFi networks using its approach. But that will take longer efforts with groups such as 3GPP and the IEEE and involve competing — or partnering — with giants such as Qualcomm, Broadcom, and others.
“The real challenge is finding the right combination and sequence of market entry points,” said Trachewsky. “Yossi has good plan and is aware of these challenges,” he said of Cohen.