Reflecting on How Innovation Happens
People always ask me, “What made you think of that?” or “Where do these ideas come from?” Well the truth is, I have no idea. However, I spend a lot of time studying the markets, talking to customers, finding interesting bits of technology, and trying to absorb as much raw data as I can. Perhaps this is the modern analog to Edison’s quip, “To invent, you need a good imagination and a pile of junk.”
Of all the things we do, trying to do something that has never been done before is always the most rewarding – the riskiest, but also the most rewarding. Our new TracPhone® V11 is a great example of this. Something that everybody knew was impossible so nobody had even tried.
The V11 idea came to me shortly after we delivered the very first mini-VSAT Broadband product, the TracPhone V7 satellite communications system. Back in January of 2008, we had a regional maritime VSAT service and big plans to roll out our Ku-band network on a global basis. (VSAT is an acronym for ‘very small aperture terminal.’) We had developed the TracPhone V7 as a Ku-band spread spectrum solution, using ViaSat’s ArcLight® technology. The benefit is that by spreading the transmit signal across a broad range of frequencies, you could use a smaller antenna – always helpful for maritime vessels.
Our Ku-band service worked great – fast and reliable – but the coverage wasn’t fully global due to the regional beam design of Ku-band satellites. I wondered if we could apply the same principles to a different frequency, one that provided global coverage using only a few transponders. The obvious frequency choice was C-band. Cruise ships had relied on C-band’s global beams for decades. However, to close the link to these beams, you needed gigantic 3.7-meter antenna systems that literally weighed 2,000 pounds.
The Power of ‘What If’
What if we applied the technology of spread spectrum communications in use for our Ku-band TracPhone V7 to C-band? Could we get the size down to the point where this became practical? We started running the link budgets and the answer was very exciting. We could probably get the total antenna aperture into the same size as a conventional 1-meter VSAT!
But that wasn’t good enough. C-band, while having great coverage, doesn’t have as much capacity as Ku-band. The ideal product would have BOTH frequencies. We needed a dual-band antenna and a dual-mode feed system that could give us the best of both worlds: super high speeds at Ku-band and terrific reliability and true global coverage at C-band. A dual-mode antenna would solve another problem: C-band isn’t licensed for use in ports. We’d use Ku-band near land and in port, and C-band for those regions where we didn’t have Ku-band coverage. I liked C-band because it also gave better performance in bad weather. Now we would have two different satellites with different frequencies. Having two satellites in different orbital slots also helps if you have blockage from the ship itself. If one is blocked, the other satellite probably won’t be.
Reality Sinks In
For all these reasons, this was starting to look like a solid product concept. I then started looking for dual mode RF electronic components and quickly realized that they didn’t exist. We’d have to invent them from scratch. Our RF teams, both in our headquarters in Middletown and our fiber optic facility in Chicago, felt confident that they could develop this technology. They were very concerned, however, about the tradeoffs they’d have to make. Since satellite communications systems must transmit and receive, you don’t have a dual-band system, you actually have four bands (transmit and receive, both at Ku and C). I threw them a bone and said that since the antenna would actually be much larger at Ku-band than our existing V7, they didn’t need the same efficiency. Even at 50% efficiency, we would have the same gain at Ku-band on a one-meter antenna that we did with 70 percent efficiency with a 60 cm (V7) antenna. But, I knew they would never live with that and in the end, we ended up close to 70% efficiency in both bands. Given that the wavelengths are so different and the size of the sub-reflector for C-band is so much larger (big sub-reflectors create more blockage and more side lobes), it’s incredible they were able to pull this off.
Another major challenge was in developing the custom C-band components that could be built right into the antenna feed system and in developing an RF waveguide path that could accommodate all four bands. In the end, the key invention was developing a solid plastic rod that carried the Ku-band signals right down the center of the waveguide and let the C-band signals travel outside the rod, but still inside the metal tube that contained it all. There were four sweeping arms that carried the signals and combined them into the waveguide. The entire assembly looked more like a high-tech ray gun than an RF waveguide.
While the R&D guys were solving the difficult technical challenges, we encountered a much bigger problem: C-band global beams were simply not available – and we needed three of them in unique orbital slots 120 degrees apart to cover the entire surface of the earth with no gaps and no unnecessary overlaps. Periodically, a C-band transponder would come on the market, but not in the right spot, or only one, not three. And even if they had been available, our CFO cautioned me to not commit to global bandwidth until we had a product to ship. So, we played a cautious waiting game. By building out our Ku-band network region-by-region and adding subscribers at a record pace, I knew we’d be in a position to cover the costs for our C-band capacity.
Finally, the RF team’s hard work paid off. They demonstrated all four bands working with efficiencies of 70 percent and the entire assembly would fit into a 1-meter dish. In the interim, we had developed a very robust robotic mechanism for our TracVision® HD11 satellite television system.
This pedestal was over-designed for the TV-only application, because we knew that this was also the platform that would support our V11. Given the extra weight of the RF components that were going to be mounted on the dish, we needed to save weight.
The answer was to switch the entire dish and support bracketry to carbon fiber. This material is common in sports and yacht racing, so I reached out to a good friend who was running a carbon-fiber sailboat mast company to design and build our dish and RF electronic support structure. The solution worked perfectly. Even though we added about 20 pounds to the antenna with the electronics, the lower mass moment of inertia enabled us to provide perfect stabilization even under the most extreme conditions.
Now we had the antenna, the RF electronics, and the pedestal all in full-scale development. We’d invested close to four years and two million dollars in product development –and we still didn’t have any C-band satellite capacity. My worries played like an endless loop: “What if we couldn’t find the capacity we need? We are going to have the satcom equivalent of the bridge to nowhere. Great antenna but no service? Please don’t let this happen…”
Fortunately, we were starting to make some progress with major satellite provider Intelsat. They had some global beams coming on the market and they were beginning to see us as more of a strategic customer. It helped that Inmarsat had just announced their own global VSAT solutions, which made KVH a key potential customer to companies like Intelsat…(the enemy of my enemy is my friend). Also, we had recently become the world’s market share leader in maritime VSAT, which helped our credibility with the satellite owners. With Intelsat’s help, we were getting close to putting a deal together to give us the full global coverage we needed, in near perfect 120 degree segments around the world.
In the meantime, we had a major regulatory hurdle to get over. C-band is unlicensed in the sense that it requires coordination – not by country but by individual locations. Getting a “license” would require coordination in every port in the world. Obviously a non-starter. But since we had the Ku-band licenses for most of the key countries already, we reasoned that we wouldn’t need to have a C-band license inside territorial waters. Our general counsel had started making contact with the FCC, and we used various consultants to float the idea of a dual-mode product by the authorities. But, we didn’t want to tip our hand to competitors yet, so we initially filed for a C-band license and a Ku-band license separately. We figured that if they approved them separately, there would be low risk in getting the combined license.
Of course, if the FCC rejected our license requests, millions of dollars and years of work would go down the drain. Given all the risks we were already taking, this one was one of the largest. We actually signed the contracts with Intelsat prior to getting the formal licenses we needed. It was a calculated risk. We knew that the combined product would appeal to regulators because it solved the problem of land-based interference in a novel way. In the end, they asked for a carve out of C-band around the U.S., which seemed reasonable so we implemented this in our design. We also needed to get an experimental license both in Rhode Island and in California so that we could do our C-band testing during the development phase. Fortunately for us, we were able to get all the licenses in time to begin production and shipments.
The product looked like it was proceeding on schedule and we were ready to go public with a big announcement. Only one slight problem remained. We had never actually tested the product live over a satellite. All the link budgets said it would work, but I started to panic. What if we just couldn’t close the link? Nobody had ever done this before with such a small antenna…maybe that’s because it simply couldn’t be done? Maybe we weren’t smarter than everybody else. Maybe we just didn’t understand the technology limitations. So hours before our public launch, we begged Intelsat to give us access to one of their satellites for a quick sanity check. Our entire design team was huddled around the antenna waiting for the service to go live and see if we could close the link. Anyone who has ever worked on this kind of a complex product knows that it takes weeks of debugging. We plugged it in, and after fiddling with some of the settings on the hub side, we instantly got online. Big cheers and high fives all around.
It’s Out the Door!
After four years, we had a product that we now knew could work. The press release went out, and we announced the world’s first dual-mode C/Ku-band product. Just a few weeks ago, we started shipping the first production units and I couldn’t be more proud of our team. Now, about my next idea…