The company, led by CEO Mani Ramachandran, formerly CTO of Synchronous Communications, has been working under the radar since 2006 to develop and supply a new type of laser transmitter that exploits the distance and multiple wavelength advantages of delivering amplitude modulated (AM) signals at 1550 nanometer wavelengths within the traditional cost parameters of 1310 nm. technology. With in excess of 1.5 million homes now passed by networks utilizing the InnoTrans platform, it appears the technology is proving itself as a way for MSOs to overcome cost barriers that have blocked aggressive expansion of narrowcast capacity over HFC networks.
íƒ˙Our Chromadigm forward path DWDM (dense wavelength division multiplexing) solution is being widely adopted by the MSO community in the U.S. and Canada,íƒ˘ Ramachandran says. íƒ˙The solution provides operators the ability to do full eight-way node segmentation in both forward and reverse directions on one fiber, for distances up to 36 Km.íƒ˘
The company is now showing 16-wavelength capabilities to customers in preparations for commercial introduction, he adds. The 16-wavelength solution for transport at 36 Km. requires use of an EDFA (erbium doped fiber amplifier) in the field. Four- and eight-wavelength transmissions at that distance are done without use of field-deployed EDFAs.
The four-way or eight-way node segmentation provides a means by which each downstream wavelength carries the full load of broadcast and narrowcast signals designated for a given node, in contrast to using high-power 1310 lasers, which require a separate fiber to serve each node. Thus, in the case of node splitting, a fiber serving one node now can accommodate division of that serving area into four or eight separate serving areas each with its own narrowcast signals, thereby greatly expanding the ratio of narrowcast channels to households. Or, where fiber reclamation is involved, one fiber can be used to reach as many as eight nodes, freeing up multiple fibers for other uses.
On the upstream, the same fiber carries all the return signals from the segmented nodes in AM mode. Ramachandran notes the ability to operate in the return over AM leaves open the use of mid-split spectrum for the return, in contrast to use of digital signaling in the return, which caníƒÙt operate above 65 MHz.
Node segmentation was the application Ramachandran and his team originally envisioned for the new technology, but he says the range of applications quickly expanded as operators discovered the versatility of having a 1550 DWDM solution that doesníƒÙt incur former cost penalties and overcomes a key limitation that plagued earlier attempts to use the technology in AM access networks.
íƒ˙Initially, we thought node splitting would be the big opportunity for this technology,íƒ˘ Ramachandran says. íƒ˙But then operators started to use it for fiber reclamation to free up fiber for cellular backhaul and business services. Now theyíƒÙre using it for N+1, 2 or 3 [which designates the maximum number of amplifiers between the node and any given subscriber] and RFoG (RF over glass for fiber-to-the-home).íƒ˘
Overall, the ratio of deployments so far is about 40 percent for fiber reclamation, 40 percent for node segmentation and 20 percent for fiber deep, with a smattering of RFoG and super-trunking deployments. The technology is also finding use in ring-in-ring AM optical architectures where fail-safe mechanisms require that if a point of failure occurs on one of the rings the optics can support instantaneous redirection of signals at new distances to keep service alive to all nodes. íƒ˙Our technology isníƒÙt loss constrained the way others are for this application,íƒ˘ Ramachandran notes.
The ability to serve a wide range of distances at acceptable performance ratios wherever a signal is terminated is one of the major advances claimed by the InnoTrans 1550 nm. AM technology. In the past, when operators tried to use 1550 DFBs for AM transmissions over HFC they had to tune each transmitter precisely to a given distance, which, even with remote on-screen tuning capabilities, proved nearly impossible because, in most instances, engineers did not know precisely what the hub-to-node distances were.
íƒ˙Distance-independence allows for a universal deployment as long as the optical input power is designed to the node manufactureríƒÙs recommended input range,íƒ˘ Ramachandran says.
He adds that InnoTrans has overcome cost barriers by allowing use of off-the-shelf DFBs rather than requiring selection of extremely low-noise devices. The companyíƒÙs optical transport platform supports up to four ITU (International Telecommunications Union) wavelengths in a single rack unit with optical output powers ranging from +4dBm to +18dBm.
One great advantage of deploying 1550 transmitters for AM optical distribution is the fact that distances can be extended with the use of EDFAs. For example, Ramachandran notes, íƒ˙As long as you run fiber to the last active (coaxial amplifier), you can activate those links later for fiber deep.íƒ˘
InnoTrans can use the lower cost 1550 DFBs with their higher noise levels because it implements electrical and optical processing to achieve a higher optical modulation index (OMI) than would otherwise be possible. Higher OMI produces a higher signal-to-noise performance at a given level of optical power, which means the transmitter can deliver the power required for longer distances without generating too much noise.
íƒ˙Normally, you need to hit the receiver hard and use low-noise lasers,íƒ˘ Ramachandran says. íƒ˙We doníƒÙt use low-noise lasers. We improve the signal level by increasing modulation depth.íƒ˘
InnoTransíƒÙ trick is to overcome two impediments to good performance known as laser chirp and clipping. Chirp, which results from the unintended frequency modulation of each amplitude modulated optical signal, causes variations in optical power from one wavelength to the next that become disruptive to performance of the combined wavelength transmission over distance. InnoTrans, using an undisclosed mode of adjustment to the transmitter output in the optical domain, gets rid of the chirp, Ramachandran says.
The other important step is mitigation of clipping, which InnoTrans does with electrical circuitry. Here the problem is that with the stacking of multiple QAM inputs the optically modulated signals of the different 6 MHz channels line up periodically to cause a very strong but very short composite triple-beat pulse that can be 20 to 30 dB above the noise floor. While the instruments used by operators to measure modulation error rate (MER) might produce an average signal reading that suggests everything is working fine, the bit error rate (BER) with each pulse can be well below acceptable levels, causing tiling or blocking of the picture on customersíƒÙ TV screens.
íƒ˙As operators add more QAMs clipping becomes one of the main impairments to performance,íƒ˘ Ramachandran says. íƒ˙When we put in clipping mitigation circuitry the MER doesníƒÙt change much, but the BER becomes zero.íƒ˘
The combination of chirp and clipping mitigation not only reduces costs of the transmitters but eliminates the need to fix distances precisely or purchase receivers precisely tuned to the performance parameters of the optical signal at a given node, Ramachandran adds. íƒ˙You doníƒÙt need to make adjustments for distance, so you can buy passives from anybody,íƒ˘ he notes. íƒ˙All we ask is that you hit the optical receiver in a given power window. It works with every node weíƒÙve tried.íƒ˘
Along with selling its gear directly to operators, InnoTrans has begun supplying OEMs (original equipment manufacturers) with the technology, especially for RFoG applications, notes George Vasilakis, vice president of sales and business development at InnoTrans. íƒ˙WeíƒÙre selling directly to OEMs when they need a downstream transmitter,íƒ˘ Vasilakis says.