Modulation approach set to solve spectrum congestion
4 mins read
As mobile communications become more pervasive, the amount of data flowing around the various networks and – importantly, the various parts of the spectrum – grows inexorably. The world, it is claimed, is moving ever closer to spectrum congestion.
When there's limited availability of something, engineers look for a way around the problem. The current solution is quadrature amplitude modulation (QAM). But even QAM in its more exotic flavours is beginning to run out of steam, according to some.
A company which has recently exited 'stealth mode' believes its technology could be the solution for which the industry is looking. WAM – or wave modulation – is said by MagnaCom to be a 'game changer' for the telecommunications industry. "What we have is a technology that will be as big a shift as upgrading analogue to digital TV," said Yossi Cohen, the company's cofounder and chief executive.
MagnaCom believes it has the ability to change the telecomms industry 'in a big way'. Cohen, an industry veteran with spells at Broadcom, National Semiconductor and Motorola Mobility under his belt, explained why.
"QAM is at the heart of all digital communications," he said. "It's pervasive; it's in anything and everything." He listed such sectors as mobile phones, Wi-Fi, wireless backhaul, satellite communications and fibre networks. "You name it, it's in it."
QAM as a technology dates back some 40 years and was developed originally to cope with the problem of white noise in data communications. "For the last 40 years, the industry has been developing various flavours of QAM," he continued. "However, we believe we have developed something which can be perceived in the same way as the jump from AM to FM radio."
WAM is said by Cohen to be a different way of modulating information. "Today," he said, "it's QAM. This says that every signal has a phase and an amplitude. The technology allows both to be recovered at the receiving end." As data communications demands increased, QAM became more complex. "Over the last 40 years," Cohen continued, "it's moved from QAM4 to QAM64 to QAM256 and so on." The number following QAM refers to the number of signals that can be passed over the particular link; for example, QAM64 handles 64 signals. The latest development has extended the technique to handle 4096 different signals, although devices featuring this are yet to ship in volume.
"But it's a two dimensional approach," Cohen contended. "Using WAM, we can place the coordinates in multiple dimensions, rather than being limited to two." He gave the example of putting 1000 dots on a piece of paper. "The dots would be much closer together than if you had 1000 points in a cube with the same footprint. That's the fundamental difference between WAM and QAM."
While QAM was ostensibly developed to deal with the problem of white noise, it's now facing other challenges. Cohen continued: "When QAM was invented, it was only intended to address white Gaussian noise. But now it has moved up by several orders of magnitude, white noise is getting dwarfed by non linearity. This is far more dominant and we are now at the point where we can ignore white noise. WAM solves both problems, offering better performance in the presence of non linearity."
He used the example of an audio amplifier. "When you keep turning an amp up, you reach a maximum – a saturation point. It's the same with wired communications; you can only turn up the power to a certain point. WAM was designed from the start to deal with non linearity."
He added that QAM fails when the power curve starts to bend. "It can only operate on linear channels. WAM doesn't need linearity; in fact, we welcome non linearity as this allows the approach to work with a much wider range of amplifiers."
WAM is said by MagnaCom to be a pure digital modulation scheme which uses spectral compression to improve spectral efficiency. Spectral compression enables the signalling rate to be increased. This, in turn, allows the use of lower order alphabet and a reduction in complexity. The company adds that WAM provides inherent diversity of time and frequency domains and uses non linear signal shaping. The non linearities are handled digitally at the receiver side, allowing a lower cost and lower power transmitter design.
More technically, WAM is a multidimensional signal construction, operating at the Euclidean domain, which breaks the orthogonality of signal construction (zero inter symbol interference in single carrier systems and zero inter channel interference in OFDM systems). It has been shown, says MagnaCom, to increase capacity and to provide an optimal handling of nonlinear distortion. Ultimately, it says, this results in 'significant improvements' versus legacy QAM systems.
Craig Mathias, principal of wireless and mobile advisory firm Farpoint Group, said: "With wireless applications often dominated by conflicting requirements, what's needed is a fundamental advancement in communications technology with the potential to address all of these requirements and more. MagnaCom's WAM [is] a fascinating approach to advancing the state of the art in wireless communications that might even have the ability to unify the local area world of wireless and the wide area world of cellular in a single technology."
One of the main benefits of WAM is that it 'plugs into' the existing signal chain (see fig 1). Because WAM is licensed as an IP block, customers can remove an existing QAM block and replace it with WAM. If this route is chosen, MagnaCom claims it's possible to get a 10dB performance benefit compared to QAM. That benefit is expected to increase as QAM grows in complexity. Other benefits include a halving of power consumption and less use of the spectrum.
The 10dB performance advantage is claimed by Cohen to be equivalent to a '20 year leap forward'. "It translates into 400% for distance, so you can transmit a signal over four miles instead of one and still get the same quality. And it's the same with Wi-Fi; you can be four times further away from the router and get the same quality and speed."
"Cost won't be a factor," said Cohen," as it's a pure digital design. You don't need to touch anything, just add a digital block which is easy to implement on any silicon process."
However, Cohen also points to WAM's backwards compatibility. "At least 95% of connections over the next few years will need backwards compatibility with QAM.
There are 7billion devices out there which are QAM compatible and they aren't going to disappear overnight," Cohen admitted. "But if they add a WAM modulator, they can connect it to the same place in the signal chain."
WAM was demonstrated at the recent Consumer Electronics Show in Las Vegas. "We know people will be sceptical," said Cohen, "so we've taken time to implement WAM in hardware so people can actually see it working and can verify the technology in their labs."
MagnaCom is now looking to license the technology. "Our approach is much like that taken by ARM with its processor cores. Our intent is to proliferate the technology as quickly as possible by making it easy to license," Cohen concluded.