This really is an excellent “new” marine-electronics network protocol. So why isn’t the industry getting behind it?
It was almost two years ago that I first got out on the water with an NMEA 2000 navigation network and, though the trial rig looked a bit funky (see above), I became a believer. I was bowled over by how easy it was to screw together a multi-manufacturer system of sensors and displays, and by how easily they all shared data, with high levels of reliability and redundancy built in. This, I thought, is the bright future of navigation aboard everything from my 14-foot testing skiff up into superyacht territory. Moreover, NMEA 2000 promised to move beyond integrated navigation to true convergence, so that eventually a skipper might mix in engine control and monitoring systems, power distribution, lighting, security…whatever…to his heart’s content. Hallelujah!
I’m a bit less sanguine about NMEA 2000 today. If it once seemed merely amusing that actual certified gear was coming out years after the protocol’s namesake year 2000, in 2007 it’s painful to see how few boats are taking advantage of its capabilities. Though the major electronics manufacturers, not to mention the Coast Guard and other benevolent authorities, worked hard to create NMEA 2000, today, if they support it at all, they’ve generally taken a push-me-pull-you approach, often rebranding it as their own protocol and messing with features as obviously beneficial as a common plug type. It’s a further disappointment that the boating press has done a poor job of cutting through the hype—they have often, in fact, helped spread misinformation—so it’s quite understandable that most sailors are thoroughly confused about NMEA 2000 and marine networking in general. But before I expand on that, let’s look at what NMEA 2000, aka N2K, can do for you.
NMEA 2000 goodness
My first NMEA 2000 experience started with a box of made-up “Micro”-size cables and tee fittings, plus some sensors—a GPS receiver, a compass, and a depth/
speed/temp transducer—–as well as a gray-scale display and PC gateway, all from Maretron (see below). It came with a useful guide to various 2000 network configurations, but this relatively simple test setup was pleasantly reminiscent of working with Tinker Toys—literally, it was a plug-and-play installation. You have a trunk line, in this case quite short, but it could snake up to 325 feet from the windlass locker to the rudderpost (see diagram p. 89). Wherever you need a network device, you put in a tee drop. The cables and tees have keyed male and female ends, thus are virtually foolproof, and they’re both waterproof and heavily shielded from electro-mechanical interference (or EMI, which is often a real bugaboo on today’s hyperelectronic vessels). Raw wire and straightforward field-attachable plugs are also available.
Now note the yellow cable; it’s a 12-volt power tap that feeds 4 amps to each side of the trunk (or you can get a heavier “Mini”-size cable capable of 8 amps per side and longer runs). Thus all the Maretron gear was powered off the network cable (with, it turned out, plenty of amperage to spare). All I had to do was screw together the cables and connect the one power feed, and the display lit up with all the sensor data. So far this network, though particularly rugged and simple, only offered me what a lot of familiar single-brand instrument systems already offered. But who gives you the potential to add on gear from other manufacturers, up to a total of 50 physical data nodes, all certified by NMEA to play together nicely?
Thus the “Wow!” came when I plugged the Raymarine E-120 and Simrad CX34 multifunction displays into the backbone. Never mind for now that I needed patch cables to do it, because an N2K loophole permits manufacturers to use proprietary plug and cable designs within their own 2000 networks (or that those companies mention N2K only in the fine print, branding their data networks as SeaTalk2 and SimNet, respectively). Regardless, both plotters understood most of what the Maretron sensors were saying and began to display position, heading, depth, and so on, as if the whole system had come from the same engineering team.
And it got better. The Simrad had its own GPS plugged into it (via NMEA 0183), but it recognized both and offered me a choice via a thorough interface menu (see above). The Raymarine automatically selected the Maretron, which I—simulating a failure—proceeded to cover with a coffee can. The E120 complained that it had lost position, then moments later picked up the Simrad output. Nice! I’d thus seen two easy ways that NMEA 2000 could support redundant devices without relying on a possibly unreliable, and probably manual, data switch (dual N2K compasses also work).
At this point, if you’ve ever used the old NMEA 0183 data standard to connect multiple devices, you are likely appreciating N2K’s power. And if installing electronics isn’t your thing, remember that you pay for complexity and unreliability sooner or later. There’s not enough room here to get into NMEA 2000 subtleties like error protection and prioritized messaging, but note that the standard is based on CANbus technology that’s used in mission-critical applications like engine controls and robotics. Yes, good installers can make 0183 work across complex systems—they’ve had to—but it generally involves multiplexers, lots of wires, and hours of fiddling around.
By contrast, my test network combined data from multiple sources seamlessly. For instance, both plotters were able to show me the set and drift of the current I was motoring in, a sophisticated calculation (comparing COG/SOG and Speed Through Water/Heading vectors) that I’ve rarely tested before, because temporarily rigging all the necessary sensors was too much hassle, even for one machine, let alone two plotters of different brands plus a PC running various charting programs.
I’ve had many more NMEA 2000 “wow” moments since that initial test, like riding aboard a center-console powerboat with Evinrude Etec outboards that were talking via an N2K backbone to a set of Lowrance digital gauges and displays (page 90), which were able to show complex diagnostic info. I also successfully hooked a whole network of notably inexpensive LowranceNet (NMEA 2000) sensors and displays into the Maretron/Raymarine network. Sure, another damn patch cable was needed, but this year Lowrance is supposedly switching to standard plugs and cables.
And it was child’s play to install Maretron’s nifty new ultrasonic wind-and-weather sensor, which played very nicely with the E-120 (see Freeman K. Pittman Innovation Awards 2007, page 71). As complete as Raymarine’s electronics line has become, you won’t find one of those in their catalog. Maretron—whose principles are, I think, the angels of NMEA 2000, because they’ve focused on maximizing the standard’s best qualities—has also introduced N2K tank sensors and is working on a sophisticated engine-display program.
There are many more NMEA 2000 connections I’d like to see in action. For instance, some Yanmar diesels are already 2000-certified, and Westerbeke just introduced two generator models that can power a 2000 backbone you can use to install control panels wherever you want; tee-and-drop simple. DNA and Moritz offer N2K-based digital power-distribution and switching systems. And many more companies—Volvo Penta, Furuno, and Northstar/Navman, to name a few—say they plan to get on board “soon.” In fact, Garmin, generally a do-it-their-own-way operation, has just introduced two new NMEA 2000-capable multifunction displays, even though it has no 2000 sensors.
Yet the tipping point, when N2K gear starts snowballing and becomes the true industry standard, has not yet been reached. For example, only a couple of powerboat builders, and no sailboat builders, are installing N2K backbones (with plenty of extra tees) in new vessels. Why is that?
Networking confusion
I once heard an executive at a major electronics company exclaim, “Hell, these days electronics networking is like the Gold Rush!” I’ve come to appreciate this in ways he probably didn’t intend. What was San Francisco like in 1849? Picture not just extraordinary excitement, rapid change, and lots of fortunes being made, but also lots of confusion, hype, and wasted energy. Another electronics manager frankly told me that while some of his colleagues believe that “open” networking— where data can be easily be shared among gear made by different manufacturers—will ultimately be good for sales, others think the opposite. This sort of strategic schizophrenia may be contagious, which explains why I sometimes get different answers from different representatives of the same company.
Rapid change and general confusion also explain why I’ve seen so much misinformation in the press. “SimNet is like NavNet.” “Buy electronics that use Ethernet, because it’s an open system.” Dead wrong and mostly wrong! And don’t look for perfection here. I’m trying hard to figure out what’s going on, but understanding marine networking today is a little like stepping into a 49er saloon and asking where the gold is.
It helps to categorize the growing clutter of “nets” and “buses” into three types based roughly on the volume of information each can handle. Call it the “three porridge bowls” approach to marine networking. The big bear’s bowl is Ethernet, which can pass electronic cartography, radar imagery, and much more between devices at lightning speed. Ethernet has become the de facto standard for this sort of thing, and it is an open-communications standard, but the data communicated is almost always proprietary. Thus, Raymarine E series talk only to each other, or to a PC running RayTech 6.0. Ditto Garmin’s MarineNet, Northstar’s N2, and so on. The only exception I know of is Furuno’s FAX-30, whose weatherfax imagery can be read by any Ethernet device running a Web browser. But the rest of Furuno’s NavNet, including the link with MaxSea, is closed.
We may eventually see more-flexible systems, like generic Ethernet radar scanners, but don’t expect Ethernet GPS receivers, compasses, and the like; the protocol is too costly and power-thirsty. These devices have instead been the province of small-bowl data networking, principally NMEA 0183, which is electronically simple. Too simple, in fact; the engineers will tell you that 0183 is more an interface than a networking standard, hence the subculture of multiplexers and the existence of improved, but proprietary 0183-based networks like SeaTalk and Navbus.
But modern electronics are rapidly outgrowing 0183’s slow speed and low data volume, and NMEA 2000 now offers both in momma-bear portions. N2K will never carry charts or radar, but it can handle many more small talkers and listeners than 0183. Plus—besides having a standardized “physical layer” (there are no common 0183 plugs and cables) and built-in power—N2K’s common messages are much more refined. Route details, for instance, are easier to share (0183’s weakness in this regard led to numerous custom formats). And there are other momma-bear-size CANbus networks getting onto boats—like the J1929 used by many electronic diesels and Xantrex’s Xanbus power-control system—but they can be gatewayed fairly easily into N2K for whole-boat integration (if the manufacturer is willing).
The Raymarine E-Series—which can talk 0183, SeaTalk, SeaTalk2 (NMEA 2000), and SeaTalkHS (Ethernet)—is a good example of how complex marine networking has become. It sounds messy, but in fact there are good uses for all these protocols, whether it’s incorporating an old Raymarine sensor into a modern system, adding someone else’s cool new 2000 device, or sending everything to another E device over a single Ethernet cable (all data nicely migrates up through all the network types).
But why SeaTalk2, SimNet, LowranceNet, and so on? In what’s already a welter of proprietary networks in all three “bowl” sizes (I’ve left many names out), why have the manufacturers put their own brands, and often plugs, on the one new network that’s not proprietary? Maybe it’s that Gold Rush mentality. During the early years of NMEA 2000 development, the typical electronics system was “best of breed”—a mix, say, of B&G instruments, a Northstar GPS, and Furuno radar. Creating a more reliable, more flexible way for all these devices to make small talk was going to be good for everyone. Then along came the one-brand-whole-boat theory of marine electronics, wherein you choose a favorite multifunction display and then get all related peripherals from the same company. The original purpose of NMEA 2000, at least as far as manufacturers are concerned, evaporated even as actual working N2K products finally materialized.
But I was inspired by NMEA 2000’s capabilities, and now, hopefully, you will be, too. No, it’s not perfect. The standard messages, for instance, will never catch up to the possibilities (a certain percentage of proprietary signals are allowed on the network), and no single display will manage all the data you might want it to in a multi-manufacturer system (thus, you’ll need Maretron’s display or PC software to calibrate its Weather Station). But, oh, the possibilities for much-reduced and -improved wiring, easy redundancy, and openness to new products. Could there be, say, an N2K windlass switch that requires almost no extra wiring and is controlled by multifunctional digital switching already installed at your integrated helm? Yes!
If I were fitting out a new boat, or refitting an old one, there’d surely be a NMEA 2000 trunk running through it, and I’d be looking at electronics, engines, and any other systems that can take advantage of it. I implore you to look deeper into the details of NMEA 2000—www.nmea.org and www.maretron.com are good places to start—and, if you too like what you see, start asking electronics dealers and boatbuilders hard questions about why they aren’t supporting it with greater enthusiasm. It may take some sailor activism to reach the tipping point.
