I was just reading an article on the NMEA 2000 marine network standard over at the excellent Ocean Navigator blog. For those of you who are not aware, NMEA is a standard backbone cabling system that allows you to connect every piece of electronic gear on your boat together, regardless of what company manufactured each piece. Most every techy type in the marine journalism community has been waxing eloquent about how cool this is. And they are right…but.
You knew there was a “but” coming, right?
You see, the thing is that many devices use the NMEA backbone not only to communicate with other devices, but to communicate with their own sensors. And therein lies the problem that I have not seen one commentator mention: If there is a significant problem with the NMEA 2000 backbone, every piece of gear on the boat dependant on it stops working. What does that mean for a fully NMEA 2000 boat?
- The engine instruments don’t work.
- All of the devices that require a GPS position, including the plotter, stop working.
- There is no read out from the fluxgate compass.
- The multi function display in the cockpit that displays radar and AIS information goes black.
- The autopilot stops working because it has no input from the compass.
- The speedo/log, depth sounder, and wind instruments stop working.
Now let’s put this into prospective. Mr. Murphy will make absolutely certain that all of this happens just as you are making an approach to a tricky harbour crawling with traffic…in the fog. That’s going to be fun! As a famous ocean racing skipper was heard to say after nearly running aground on Cape Sable in the fog, “only your laundry knows for sure”.
So how likely is it that this will happen? Let me quote from the ON article (emphasis mine).
Because cabling problems are especially troubling for bus or linear topology networks such as NMEA 2000, the cabling should be checked first if any malfunctions start to manifest. This is because if there is a short or open in the trunk or drop lines it can affect the entire network, downgrading overall performance or even taking down the network. The problem, however, could be as simple as a missing/damaged terminator or loose/dirty connectors. If cleaning and reseating connectors or terminators cannot solve a problem, then by all means seek the professional help of an NMEA-certified electronics technician. Follow this link to find the closest CET to your homeport.
Hum, this is comforting. There are yards of cable with dozens of connectors snaking all over your boat and routed through inaccessible and damp places. And all it takes is one of those connectors to fail and it’s lights out.
I’m an electronics and computer technician by trade and I can tell you that network problems can be some of the most frustrating and time consuming problems to troubleshoot because there is no easy way to logically zero in on the failure point. No, all you can do is check every single connector and terminator until you find the bad one. And if the problem is intermittent, as it very likely will be, then the trouble shooting is going to be ten times more difficult.
But not to worry, all you need is a “NMEA-certified electronics technician”. Good luck with that in most of the places we voyage to. Even if you can find one, think about what his or her charge rate is likely to be. And think about the number of hours solving a difficult problem could take—ouch!
Now don’t get me wrong. I’m not saying that you should not use NMEA 2000 or that it’s a bad protocol. In fact I really like the idea of replacing the hundreds of different wires running hither and yon through most modern boats with a single network backbone.
But what I am saying is that we need to think about what a network failure will do to us on a dark night in a tricky situation and have a backup plan. And because of the issues with trouble shooting these complex systems in remote places, it would be good to have a plan for continuing our cruise without all this interconnection too.
I, for one, will make sure that:
- There is a plotter with a separately connected GPS sensor, or GPS with paper charts, available
- There are backup hard wired, or mechanically connected, engine instruments. We already have these.
- There is a properly swung mechanical compass in view of the helm.
- There is a separate AIS unit with its own screen and direct connection to its own GPS within sight of the helm.
- We have an independent and immediately accessible backup option.
The point of all this being that if we are going to be safe and competent, we need to really think about new technology and what its intrinsic vulnerabilities are and not just get swept up in the hype.
At any given time often one or more of the very few electric devices on my boat does not function. It seems to me that integration of many electric devices, some complex, in a marine environment is a guarantee of frustration & failure.
Integration requires backups that lie outside of this integrated system, like the few examples you provide. These can be very few – the essentials – handheld GPS, lead & line, ordinary compass, candles & flashlights, etc.
Integration also requires that the boat owner be a damn good electrician.
A wise and timely reminder – daisy chains linking all of the vital ingredients of your nav system offer all sorts of risks, and I don’t like them.
On Pelerin (and any other boat I advise on) we have one of the excellent Furuno GPS 32’s, which is not linked in to the Simrad network, which covers AIS, plotter, autopilot etc. Instead it is simply linked to the ICOM VHF/DSC radio and a cockpit repeater for SOG/COG log, etc. We also use the Furuno set for anchor watch. These sets are simple and reliable, and by keeping it simple we preserve the security offered by the DSC function of the radio.
Like Nick we also carry a complete manual back up – Walker log, sextant, lead line etc. – even if lost all of the electrics we could still keep going that way.
Hi Colin and Nick,
Thanks for the backup and amplification. We too still carry a sextant, lead line and Walker Log.
However, maybe the three of us should keep a little quieter about this for fear of ending up stuffed and mounted in some museum. 🙂
All kidding aside, it’s a logical progression: the more you have been out there, the more times stuff has gone wrong on you at a bad moment, and the more sceptical of complicated equipment you get.
Three sextants here: the family that plots together…
My first exposure to computers was programming (on cards!) in FORTRAN in the 1970s in high school. The key word then (and still) is syntax: the terms were important, but the order in which they were given was also critical to effective task execution. Viewed metaphorically, the task of cruising is to arrive safely, keeping as much of the water out as is possible. The syntax of this is variable, but instrumentation, which provides a representation of the world, cannot take precedence. If all the screens went black, I’m pretty sure I could find, via means either analog or Walker log, our way to a safe anchorage.
So if integration is a weak link in the chain of instrumentation, why integrate at all? I prefer to optionally integrate via a PC, like having AIS targets on an Open CPN display, when needed. Otherwise I risk chasing down physical connectivity problems precisely when I should be on deck, actively steering. This is why I prefer discrete devices that don’t actually integrate unless I plug in a USB hub or an Ethernet cable. I can do “overlay”, after all, in my head, and analog dials or buzzers for engine functions or Tank Tenders for tankage are quite adequate and don’t require the movement of electrons in an inherently hostile environment.
Don’t get me started on slaving the AP to a GPS with a waypoint close to a nav aid, using last year’s electronic charts…anyone else notice that navigational errors resulting in total hull loss are becoming more egregious in the last decade?
After much hand wringing and procrastination we just upgraded our electronics with a new Raymarine chart plotter and AIS. I found out not only do “they” want you to link everything and the coffee maker to their system but that NMEA 2000 requires its own separate power source. Apparently, it cannot get power from the instruments it talks to. Just one more (weak) link in the chain to go wrong which I find puzzling.
The positive side to selective upgrading is, as you mentioned, different systems. Our other instruments work on an older protocol and when the marine tech I worked with asked me, “Do you want me to hook it all up and the autopilot together?” I gave him an emphatic, “No!”
We carry paper charts and separate GPS units on board as well.
Hi Derek H,
That makes a lot of sense: keep some of the old stuff as you transition to the new.
I agree completely with you about not connecting the Autopilot.
Thank you for the timely piece on nema 2000.
We are taking the plunge on the new boat, but not all the way, not everything will be connected into one system. We will keep the NKE Auto pilot and depth sounder separate from the Garmin chart plotters, AIS and HD radar. We are not crazy about the new way of doing things but that is what is out there. Having crossed the Pacific in the early 70’s with a sextant and no electronics I’ll take my chances on the new stuff. Having crossed again from 2007 to 2009 with two chart plotters that talked to each other through just one cable, a separate radar screen and a separate AIS screen life was good. But now that is not so easy to do so we are going to put our faith in the new system and see what happens.
Back up will be hand held GPS that will take our chart chips along with 500 paper charts. Lead line and a battery powered hand held depth finder, we use that for exploring Atolls and uncharted rivers with the inflatable but great for a backup. Yes the sextant still travels with us and the only time we use it is when we want some fun, but again it is there for serious back up.
Instruments and their sensors, the most unreliable navigation equipment onboard my boat.
PC has been pretty good. Its not marine though (just as well it seems) and never sofar failed on passage.
The boat (marine) GPS/plotter has probably one of the most dangerous bugs you can get… the GPS location stops working and the instrument thinks that the boat has simply stopped moving, so… no warnings given.
The PC, boat GPS/Plotter and the cockpit plotter all have seperate GPS receivers.
I would never wire everything up to one sensor for GPS, but for other things like depth, you sort of accept the one sensor. I do however carry a spare and some spare older instruments that could be bodged together.
I just thought of another important point. The modern PC usb GPS sensors work very well down below, even on my aluminium boat. There is no reason to run everything off one sensor, a usb gps device can simply plug in to the pc. Nearly instant startup, the PC by the way, uses only 8 watts not counting the screen.
You are right. There are some things that should be outside the NMEA 2000 bus. Compass signal to the course computer is one. GPS signal is the other one. At least have a backup GPS signal connected direct to your plotter. I would also advice not mixing other NMEA 2000 data like tank sensors etc. in the navigation system
I absolutely agree with your skepticism. A boat awash with interconnected electronics is great fun at the marina. I’m against it for the same reason I’m against having a “Stepford Wife” (for those of you old enough to remember that movie – the others can Google it). Simple redundancy over abundant complexity is my motto. Where’s the fun in a playstation boat?
Hello all, John is smirking now because he knew I could not resist taking a shot at this avenue of “progress” in marine wiring technology. Personally I will resist this modern trend until impossible to do otherwise which may be soon. Once one gets past exposed wire looms “keeping it simple stupid” is still the way to go. Wire runs that are exposed are much easier to adapt, inspect, and maintain. On my fish boats (offshore, North Atlantic, year round) each piece of essential gear is duplicated side by side using the exact model with adequate cabling in the bundle to allow swapping out the plugs. So each GPS, Radar, plotter, sounder, sonar, SSB, VHF, Autopilot, etc can be swapped with its mates antenna/display/sensor mechanically in a minute. This is a system I adopted in the early 1970’s and it has served us well. I even include a spare power plug connected to an independent power source from outside the engine space. While I’m on the subject I want to endorse Furuno equipment, this stuff is tough, dependable, and quite tolerant of rookie technicians. I’ve seen it all while trying to modernize; Mathers electronic engine controllers that shifted gears when a infrared mouse was introduced to the wheelhouse by one captain. Simrad wind instruments that develop a slight error after a period of light wind. This slight error is in the wind angle and really raises havoc with all the devices daisy chained into the circuit on my sailboat. Repeatedly flipping the power on/off will sort out the error once discovered. Simrad failed to sort this out during years of mergers choosing rather to sunset the model. Everything has been replaced at least once so the issue is still a mystery. Cross talking acoustical gear can be tamed by judicious use of multiple frequencies and shielded cabling. I won’t run a cable today that isn’t shielded and that goes for power cables if they connect to electronics. All this stuff is asked to operate in close proximity to similar stuff often in a humid air not likely to be at room temperature. My good friend and electronic technician Dick Ross ( deceased) carried a heat gun and cold spray as his primary diagnostic tools. Most issues could be replicated with either/or treatment. Today’s devices are indeed more climate tolerant but hard drives don’t like falling off seas still and although the voltage tolerances are much better things run hot when starving for amperage. I plan to be the last adopter if my only choice be network cabling, I hope all you brave souls have things sorted out by then. Bob T.
All good advice, as always. And some frightening stories too: mouse shifted gears, now there’s a scary picture.
Really good advice about shielding cables as well. That is a good thing about the new NMEA 2000 system—much better shielding and waterproofing.
What are the advantages of NMEA 2000 that justify the problems outlined in this article?
My stuff is old, so I use the NMEA 0183 standard in order to link my GPS (the heart of the system), two radios, AIS, and computer, which acts as chart-plotter (the AIS boosts the signal to the computer). I wire everything to and from one eight stud terminal bus, so it is not daisy-chained. Since I did the wiring myself, I know where everything goes and what color means what.
I also have a backup GPS and sextant, by the way.
1. No vendor lock-in. In theory, all N2K devices are compatible with all other N2K devices. (In practice, some vendors are going out of their way to implement proprietary PGNs that only their own devices understand, and two have built their own incompatible physical layers to run it over.)
2. One hell of a lot more robust than NMEA0183, which is somewhat less than fun to configure as anything more than a point-to-point serial link. CANbus can handle an awful lot of abuse and wiring faults before it actually fails to deliver the data.
3. Much less wiring to pull. One central backbone providing power and data, with a single cable T’d off it at each sensor, is much faster and easier to install than running individual power and data lines from each sensor to the helm.
4. Useful new capabilities- like having a display unit monitor dozens of sensors and fire off an alert to the skipper if anything goes out of spec (“BZZT BZZT Engine #2 cooling water flow is lower than normal, temperature high, but not enough to kill it- yet.”)
None of that nullifies the problem of having a shared means of failure, though. While I will happily specify NMEA2000 for all the convenience (non-safety-critical) instruments, I believe that- at a minimum- the engine bus should be bridged (never shared) to the instrument bus, and that the plotter, DSC VHF and AIS should use their own dedicated GPS receivers on direct point-to-point links. A set of non-networked engine instruments (eg. Murphy Gauges) are of course a given if you depend on the engine. And, although I am part of the “digital native” generation and am constantly surrounded by computers (and spend half the day programming them), I’ll always prefer to trust my paper charts and magnetic compass.
Can you run a redundant system along with the primary?
NMEA-over-bluetooth would be fantastic for new-installation in a metal boat. Just be sure to have a repeater or signal booster within signal range of all the devices. No more troubleshooting network wiring, at least.
Why not in a wood or composite boat? For the same reason metal provides interior lightning protection – the hull acts as a faraday cage to help block signals. It’s why all your comms gear needs externally mounted antennas. The reinforcement embedded in a ferrocement hull (sh)would provide similar. You probably don’t need or want to be sharing your network device-chatter with your neighbors in the marina if/when you park there.
Of course, multiple-redundancy for your most-important items is good.
those of you with the sextants: do you use them for anything other than the noon sighting when the sun is visible ? or do you also carry all the tables and other paraphenalia to do the non-noon sightings ? if the latter, then how do you find the energy and time to do all that ? my view is unless you do these exercises regularly then proficiency, thus accuracy, goes south rapidly…even now i’m sure i would have to brush up on doing the noon sight before i could do it again it has been so long, and then you need the sun to be visible, which is iffy at just the right moment…i’ll stick with my base station plotter with handheld backup along with the deduced reckoning, which seems to work automatically with me…also, the sextant needs so much babying not to mention the space it requires in its cabinet or carrying case…bleah
what am i missing here ?
richard s., tampa bay, s/v lakota
For us the sextant is just backup and something to use once in a while to take out and use on a long passage for fun and see how close we are to our GPS readings. We only do noon sites now as the rest is just too much work. I trust in an emergency and we really have to use it with noon sites we can find our way to a safe port.
I always get a kick out of folks on the different sailing sites who say they are only going to use a sextant, no new stuff for me. I remember long ago how stressful that was, always worrying if you were really in the right place.
I’m sure the NMEA 2000 situation will be somewhat stressful also especially if it fails. But with backups like a small hand held GPS, charts and the sextant anyone with good common sense should be fine. We always will carry a spare GPS antenna and if need be I should be able to hook up one of the chart plotters direct to use. If I can’t use the radar or AIS I will make it just fine, we almost always did long ago.
The MNEA 2000 is complex and can be a problem I’m sure but when you are having a new boat built it would not be cost effective for the contractor to find and install NMEA 0183 devices, I remember when we thought that was stressful. So we hope that by keeping the chart plotters, AIS and radar to one product type, (Garmin) we will have some luck. Our NKE auto pilot will be programed to be simple no contact with chart plotters we will just set a compass course on it and sail like we always have. We will also have a wind vane and use that when under sail.
I guess that NMEA 2000 is new enough that I won’t have to worry about the next break through in electronics for a while.
Don’t get me wrong, no one was happier than I when the GPS was invented and I could retire my sextant. However, I’m also aware that the electro magnetic pulse from a lighting strike can fry every piece of electronics on the boat, right down to the spare GPS and the watch on my wrist.
If that were to happen the Walker log and a daily noon site (no tables required) will get us to port. And I still carry my battered and salt stained copy of Mary Blewitt’s celestial bible which will remind even this aging mind of how to do the sums.
John, Lots of good food for thought. I put in a NMEA 2000 backbone with new instruments (mainly Furuno) last year. I have just the basics and have unconnected redundancy, but I must say, the simplicity of installation, the robustness of the cabling and connections impressed me. I have assumed, because it is so often the case, that NMEA systems were a spin-off of either aviation or automobile, both areas which tolerate few problems. Do we know of many system wide failures? There is room for installation problems as N2K systems are not all plug & play. And boy, it sure beats NMEA 0183 by a landslide.
My best, Dick Stevenson, s/v Alchemy
Thanks for the real world experience, always the best kind.
I agree, NMEA 0183 was an absolute bastard of a system cobbled together from the decades old “current loop” system, if memory serves, and way overdue for replacement.
My post was not intended in any way to discourage people from using NMEA 2000, but only to point out that a system with this level of integration needs to be very carefully designed to be fault tolerant and not so that we inadvertently put all our eggs in one basket.
For example, the schematic that starts this post shows a single GPS receiver providing all position information to all devices, dito the compass—not a good idea, in my view. I would want at least one navigation device available that was completely free standing and separate from the network and powered up at all times (not a hand held GPS stuffed in a draw somewhere).
Just to make everyone aware, I am with Maretron, a company designing and selling NMEA 2000 products, so this post will obviously be swayed in one direction. Nevertheless, we take our products very seriously and your concerns are legitimate, so I would like to address some of the reliability issues. First, NMEA 2000 is based on CAN bus technology (NMEA 2000 is simply the protocol or formatting of marine specific data within a frame of CAN data) and there are over one billion CAN nodes in the world, a true testament to the technology. Cars, buses, trucks, agricultural equipment, planes, and boats are just a few of the places that CAN nodes can be found. One of the many, many fail safe mechanisms is that each node monitors its own transmissions and if the messages aren’t being interpreted correctly at the other end, the node will automatically disconnect itself from the network. Of course the worst possible failure mode is a short circuit, however if only one of the two data lines is shorted to power or ground, the system will continue to operate albeit in a fragile state that needs to be repaired. A catastrophic failure would be if both data lines were shorted, although there are some preventive measures that can be taken to minimize the impact. First, and probably not very practical from a cost standpoint is a completely redundant network. Two GPSs, two compasses, and so on, each connected to separate networks but bridged together. If one GPS fails, then all of the consumers simply switch over to the GPS that is working. If one of the networks gets shorted out, the bridge isolates the networks such that the other network continues to operate. An alternative to the full redundancy system is to break up the network into segments that are logically part of the same network, but physically isolated in case the cabling system is compromised. As an example, a bridge can be used to create an engine room network segment, a main network segment, and mast segment so if one of the segments is compromised, the other segments continue to operate. Here is a link to a bridge product that allows full redundant networks or segmenting of networks.
As previously stated, and the point of all this is to be safe and competent. One way to improve the safety and competency is through a situational awareness about the boat through a reliable NMEA 2000 network. Robust networks are a reality and can be designed as attested by hundreds of millions of CAN bus networks working reliably around the world.
Hi Maretron, (A person name would be nice, if for no other reason that it feels strange to address a company like this.)
Thanks very much for the well reasoned analysis. I could not agree more that NMEA 2000 is a fundamentally good protocol that is a huge step up from the rat’s nest of NMEA 0183. In fact, if I were doing a boat from scratch I would be buying your gear.
My concern is that no one is dealing with the systemic vulnerabilities of these network systems or coming up with backup navigation solutions that can be quickly isolated from a malfunctioning network by a bleary eyed, non-techy, who is slightly seasick, at 3:00 am, on a dark night, when its blowing like blazes.
See Bob T’s comment above for the kind of thinking we need to make sure that these systems are safe in the hands of the non-technical in adverse circumstances. Maybe we don’t need to go as far in that direction as Bob does on a fishing boat that stays on Georges Bank winter and summer, but we do need to establish some standards for backup and isolation.
My name is Rich and I’m sorry for not giving my name earlier, I just wanted to make sure everyone knew where the post originated due to my obvious prejudice on the subject matter.
With regards to systemic vulnerabilities, the key is to understand the vulnerabilities or potential failure modes and then design the system as robust as practical to either prevent failures, or keep the system up and running given a particular failure. Take for example cable and connectors, which from my experience is where a lot of issues arise. The Maretron cable is triple shielded, one shield around the data pair of wires, another around the power pair of wires, and another shield outside the two internal pairs of wires. This ensures that the NMEA 2000 signals stay in the cable (and not corrupting your VHF transmission/reception) as well as protects the NMEA 2000 signals from being corrupted by high power RF like the SSB and radar. The cable also has over molded cable ends that are waterproof (IP68) with built in strain relief. The connector pins and sockets are gold platted and the connections are o-ring sealed for high reliability connections.
Nonetheless, stuff happens and cables, connectors, and product can be compromised and this is where one needs to decide just how important the information is because its going to cost to design in some redundancy (using bridges and isolators as previously discussed). But it is possible and Maretron would be more than glad to help anyone understand the tradeoffs and assist anyone designing an NMEA 2000 network. We offer a free software program (N2KBuilder) that allows one to design an NMEA 2000 network.
If anyone wants to discuss their network design or wants Maretron to review their N2KBuilder design, just send an email to the following email address:
Thanks very much for sharing your knowledge of NMEA 2000.
Going back to my orginal worry about having too many eggs in one basket, I would be really interested in your take on the correct network architecture and backups to make as sure as possible that the scenario I sketch out in the post does not happen. And further, if it does, a network design that would allow our slightly seasick non-tecky, to immediately make adjustments that would preserve or return the base functionality that I specify at the end of my post.
Assuming you can come up with something that meets those goals, preferably with a diagram, we would be happy to publish it as a separate post.
The very comforting thing about the comments to this post is that clearly all of you have thought about this issue and planned for a failure, and that’s great.
The other takeaway for me, that is very cool, is that somewhere in a region of a million miles of experience has spoken up here and come to fundamentally the same conclusion.
Thanks to all, and please keep it coming as you think of relevant additions.
All those years, all that money and we reinvented a string of Christmas lights?
Chris, Not even close to compare to Christmas light. I would confirm Rich’s description of the robustness of the Maretron cabling. It is their products I used for my backbone. They also make the easiest to install multi-wire (5 prong) connectors I have ever used, both straight and right angle. They were so well shielded that I used them for a 5 wire connection for my ham radio over other products. I suspect that the very spiffy diagram that is at the beginning of this article was done by Maretron design software. If not then it is very close and in any case, the design software is a free download from Maretron which can be a huge help in designing a system as it trouble shoots design errors and checks things at the same time.
Dick Stevenson, s/v Alchemy
My comment was about the network architecture, not the quality of components. And the diagram I am looking at, and the NMEA 2000 caveat quoted by John looks like Christmas lights — albeit with spiffy bulb holders.
When one component can bring a whole network down and one needs software to figure out which component it was, I don’t believe it is ready for prime time unless a Sysad is on the crew manifest.
A ring bus architecture might help with single point component failures, but I would rather have is self healing network that can multipath around a failure, localize, isolate and report the offending component. (hmmm, sounds like ARPANET).
These topology control networks exist in the internet, aviation and nuclear power world, and dumbed down for boats could be quite useful.
A simple change is to make sure there is a separate power supply to wind/speed/depth instruments. I had a very hard to find short in the wire from the GPS, which kept blowing the fuse. I also lost my depth gauge. At least wiring a separate power line to these instruments helped a lot. And it’s easy and cheap to do. You may not know where you are but you know the depth without he lead line.
Hi Jim, good point.
Before we get too far down the line of discussion about the quality and reliability of NMEA 2000, I need to remind everyone that that is not the issue that I brought up in the post. I’m quite willing to stipulate (as the lawyers would say) that NMEA 2000 is a reliable robust network that is a vast improvement over NMEA 0183.
What concerns me is that a network architecture, like that displayed in the diagram above, of using a single backbone to provide all sensor to display communications for all mission critical equipment is, in my opinion, fundamentally dangerous.
So what I think we need to do in equipping our boats is think about how we will make sure that mission critical displays—most notably plotters and radar—will stay working even after a network failure, or can be quickly returned to function by a seasick non-techy on a dark night.
I will be very interested to see what Rich, of Maretron, comes up with, should he decide to take on the challenge of coming up with a network layout that solves the fundamental problem outlined in my post.
Slightly expanding the topic here…
The risk of integrated systems was brought home to me by a crew in the ARC 2011. They had engine problems that meant they needed to request other participants to ‘drop-by’ mid-atlantic and pass them food. However, they didn’t require food for the obvious ‘going slower than we planned, so need more food’ reason. What had happened was their engine problems meant that the alternator wasn’t charging the batteries, so they had no electricity, so they had no instruments and the solenoid on the gas system meant they couldn’t use their cooker and most of their food required cooking to be edible!
Now that sentence contains an incredible number of poor decisions and lack of problem-solving expertise, but it happened and demonstrates how one issue (with the engine) can affect even food preparation!
Once we heard about this boat, we did a mental exercise to see how we could cope without engine or electricity or both.
I ran into the exact problem that John wrote about in the original post. While at anchor in the Bahamas during a major thunderstorm, my entire N2K network went down. I did not think that there had been a direct hit because everything appeared to work with the exception that there was no communication between any components including the plotter with the GPS antenna. I suspected that the nearby lightning had damaged the masthead unit with wind direction and speed. Since the masthead unit has a built in 120 ohm resistor, you can’t disconnect it without replacing it with a resistor. Luckily I was able to get a terminating resistor and when I disconnected the cable to the masthead, the rest of the system came up and works fine. I think that one lesson is to carry some spare parts, especially terminating resistors and maybe a spare power tap. If you have the resistors, you can disconnect various sensors , transducers and instruments from the backbone until you have the basics that you need working. For instance if you need to connect the plotter to the GPS antenna and nothing else, you only need the section of backbone cable that goes between the two with a power tap and a terminating resistor on each end. Every boat should have the wiring diagram for their NEAMA 2000 network and a somebody with a basic understanding of the network rules.
Thanks for the real world story, always the best kind.
What I find most disturbing about your experience is that clearly you are a technical kind of guy who was able to make a well informed guess and get the important stuff working quickly. But realistically, would most people without a technical background have figured out that the problem lay at the top of the mast, and even if they had, would they have had and understood how to use a spare terminator? I fear not.
Great recommendation on having a wiring diagram. All too rare though.
I think some kind of a master wiring diagram is essential to have whether you are explicitly networked or not. The value of networking, or the degree to which networking has a downside, is a different discussion, of course.
It’s because I’m *not* a electrical contractor that I keep track of which bus or sub-panel powers what, and where the physical runs go. I can’t always assume that I will be the guy available and coherent enough to fix whatever needs fixing in a crisis or even in a situation where I can’t easily comment. I have developed wiring run diagrams, make extensive use of a labeller, and continue to develop various “maps of the boat” showing locations of safety gear, flares, PFDs, crash/ditch bags, charged VHFs, etc. In addition (and I saw this on a delivery of a Bristol 45.5, which has a truly daunting number of cabinets and stowing places), I am starting to number and code various lockers and cabinets, and via the same “map” method, I list what spare or provision may be found where. In what quantities, and purchased at what date (which may or may not be relevant depending on “best-before” factors of gear and food), I leave to an associated Excel file. Light items, seldom used, go to the ends, naturally.
If all this is evidence of OCD, I’ll take that rap. Step aboard some boats and it is immediately evident who has a system of stowage and gear and systems, and who has not. My ideal (along with assurance some degree of safety via coherent and updated stowage records) is to have any reasonably competent crew take command of the boat and to find everything needed without tearing the place apart.
I would recommend that anybody who needs a drawing of their NMEA2000 network download Maretron’sN2K builder. It has complete instructions how to build a network and pretty much anybody that can envision the electronics that are in the network could probably trace it out once you understand the rules. You start with the masthead unit. If you have wind direction and velocity this is probably one end of the backbone. find the first “T’ back from the masthead cable and work your way to the other end of the backbone cable. Some networks may be more complicated than others, but for the most part you have one backbone cable with a bunch of “T”s so it is not difficult to trace with some degree of logic. If you can figure the layout, you can print a drawing of the sytem from the N2K builder. I am not an electronics guy, but installing the N2K system was a lot easier than dealing with NMEA183 and trying to figure which wire goes where, etc.
All good tips, thanks. However, I would still recommend breaking the network up with gateways for the reasons I state in the post.
An interesting post – I am in the processing of planning an instrument setup for a boat I have just purchased with no equipment on it (other than some early 80’s Decca kit that I don’t think I’m going to recommission…). I’m planning a Raymarine SeaTalkNG setup, with a few N2K devices. I qualified in electronic engineering so I’d hope I could troubleshoot most problems, BUT finding a solution to them offshore without a whole load of technical spares is another question altogether. I’m disappointed with myself for not considering enough “what if” scenarios … having read this I think I’m going to design the network such that it can be run as a number of “sub networks” in the event of a failure somewhere. By carrying a few spare power lines, terminators and junctions I should be able to at least get the “cockpit gear” running independently of the rest of the boat. Autopilot and GPS for the help, and speed / depth running on the “forward half” at the chart table. Great food for thought, John.
I think that’s exactly the right way to design a N2K network and just the way I would do it. Sadly, I will bet that less than 5% of the networks out there have that kind of fault tolerance designed into them.
I understand Johns question and see his point completely. I have put to work numerous vessels with a very high number of NMEA 2000 devises which could cause a loss of a big part of vessel systems i.e, Empirbus electrical system, Fuel, Water, Black Water system monitoring, Navigation equipment etc. The first point is that yes a single failure could bring a lot of things down, but more importantly we need to look at this from an angle that some just don’t think about. If you want a complex piece of equipment you must be prepared to maintain it. If you need to maintain it yourself you need to ensure you can through training and studying. Airline pilots have to understand their systems and airliners have very complex systems. If you buy any modern car these days you wont be able to fix it yourself, unless you have the tools and the required training. A modern yacht will be the same and if the owner wants the sophistication and all the bells and whistles they will have to be like the airline pilot, not like the car owner i.e. trained to operate there vehicle up to the point that they understand their systems inside out. Or they may need someone on board you can. The sad truth is that as boats get more complex there will have to be an acceptance that more knowledge will be needed to operate them. But there is an other issue here. No boat should be built in a way that an essential service is not available due to a general failure. Thus as people are quite rightly pointing out here you should have VHF, GPS, Bilge Pump, Fire Fighting System etc. that works even if you main electrical system has failed. Lets remember that a chart plotter, AIS, echo sounder, electric lights are all nice to have but not essential. Equally you could as mentioned here have an alternator failure and lose everything. If your gas valve works of the electrics why was there not a bypass valve? Cheap to do as a back up? Or why was there not a member of crew who could not disconnect the hoses and dump the valve? The two worst things that could bring a NMEA 2000 network down is loss of power in which case nothing works anyway or its a lose connection, as in a cable has parted, been crushed, chopped in half etc. But your alternator another single point of failure could cause same result. Therefore you could wire your systems with back up after back up, but we know that no yacht from any boat builder comes with any and that most of the extras get fitted by the owner. So it’s up to the owner to ensure that they are comfy in varying scenarios and can cope with them. Sometimes its easy to blame technology although it’s working in our favour really. I remember in 1990 when a yachtsman with a GPS… Read more »
Actisense makes some great products and has excellent tech people for backup. Is not the QNB basically a junction box for N2K (NMEA 2000) distribution? If I am correct Furuno has a similar product (J5002) which I have used and does make set up of the N2K backbone easier. At quick perusal, the Furuno also has internal dip switches for the resisters which it looks like the QNB does not. This also contributes to easy installation. There may be a price differential that explains this.
Checking for a balanced N2K system is easy where any junction box is used as it allows access to the blue & white wires of the N2K cable. They should show 60 ohms as you report. I am not sure what kind of leeway there is, but mine show 60 +/- 0.1. It would be harder to get readings if the N2K network was just terminal fittings & T connections. I believe N2K Maratron has some kind of tester, but I have no first hand knowledge.
Dick Stevenson, s/v Alchemy
Yes Actisense do make some good products. I used to use the QNB with out plugs, but this meant a lot more making up connectors. I used to do this so that I could cut all drops to neat lengths and used field installable plugs. This looked cheaper on paper than buying premade cables but to be honest meant a lot of work fitting plugs and terminating. It also led to issues with reliability. So now its QNBs with plugs and premade drop cables even if they end up being a bit too long. Thus quality is improved which is the name of the game for that reliability.
Yes Furuno J5002 junction boxes do have jumpers that allow you to turn on or off end of line resistors. But your back to more hard wiring and good quality 120 Ohm resistors are cheep as chips to buy in packs of 10 and you need two at most. If you have a NMEA2000 weather instrument or GPS with resistor built in at one end and a Flux gate compass with one other end you don’t need to fit any.
Yes Maretron do have a meter and we have one. It is very good for checking a system though and helps with fault finding. We use lots of Maretron products and they are also good. Though AC Monitors use what is a J1939 derived PGN and works well with Maretron displays like the DSM250 but will give you trouble if your trying to display AC data on other NMEA2000 instruments which expect to see a NMEA2000 PGN.
I have now worked a lot with FW Murphy Powerview displays https://www.fwmurphy.com
which I can program and have worked around Maretron ACM PGNs. But for those hoping to use with say Raymarine Multifunction displays I am not sure.
A friend of mine has just purchased a motor cruiser and asked me to help him install all new electronics for navigation, engines and tank levels.
After reading your article I suggested to him that he splits the NMEA 2000 network and join them with a optically isolated bridge. Any problems with cables, connections etc. on non essential equipment will not affect the essential stuff.
You can buy everything off the shelf and with a little bit of coding you’re done. All you need is a couple of Arduino’s with the CAN bus interface and an isolated connection between them. All the code has to do is listen to to any communication and pass it on in both directions.
It shouldn’t cost any more than $100.
Sounds like a good plan and just the thought of thinking that I’m advocating in the post.
I recently helped deliver a brand new 55′ sailboat with N2K systems. We quickly discovered that none of the electronics worked. The chartplotters, depth sounder, etc didnt work. We discovered through a phone call to the manufacturer that this was a known issue. Faulty data from the windspeed indicator at the top of the mast was shutting down every other system on the boat. We were able to get things working again by unplugging the windspeed indicator from the main line. With that one change everything started working. Definitely something to worry about!
A very sobering story, and even more so when you think about the poor SOB that had to trouble shoot the problem before it was a known issue!
I switched the instruments on Cats Meow from no longer supported NMEA 0183 to N2k from a variety of manufacturers. What I appreciated the most was the ability to choose from a variety of vendors. Nonetheless, there were hard lessons learned before all was said and done.
My biggest unhappy event was multiple “near miss” lightning induced failures….simultaneous wind sensor and AIS induced network failure along with damage to additional components along the network.
Subsequently we further divided the backbone into separately powered sections each with its own isolated power supply. We further added DeviceNet “lightning protection” to drops from individual sensors as well as segments of the backbone.
The wind sensor has an especially robust “lightning protection” circuit to almost instantaneously tie both the signal and power lines to ground when voltage exceeds 18V on any line. Admittedly its unnerving to see parts of the network drop out during “near” lightning strikes while the excess charge bleeds off, but everything comes back so far.
I don’t know whether our belt and suspenders approach actually did anything to further protect the network or just gave me something to do, but I suspect it is more than luck since anchorage neighbors repeatedly have been less “lucky” since.
I’ve also added “lightning protection” to all incoming antenna cables after seeing the solder evaporated from the inductive coil connections on our SSB tuner.
Note, I have no expectation that the “lightning protection” would do anything to protect against a direct strike.
I think that’s great advice, thank you. Just the sort of thinking I was hoping to inspire with the post, although I think you have taken it to the next level, over what I was thinking about.
By the way, what lightning protection are you using?
I’ll detail the lightning protection when I get back to the boat from Colorado at the end of the month. The protection for the base of the mast is made up of components from MGC Surge Protection which I got through Ken Thomas at Sentinal Power in PA. Other points were protected by Litton DeviceNet lightning arrestors I got on eBay. The network backbone was divided via Actisense QDP-1 power taps, and QNB-1 network blocks. The isolated power supplies were from Victron, although they simply rebrand them. Coax cables were protected by Polyphaser gas discharge devices.
Finally, even the control cable between the ICOM 802 and AT140 has a Litton Devicenet lightning protector inline.
Sigh, I wonder if I am jinxing myself……
That’s great information, thank you very much.
I admit that I haven’t done lightning protection for anything that isn’t on the mast, and if I were doing more than a little weekending when work allows that is a worthwhile undertaking. As far as NMEA 2K goes I like to see each sensor hooked to its own display (wind transducer to wind display, depth transducer to depth display, anything than needs GPS to a GPS antenna, etc…) and then I really like the ease of NMEA 2K to hook the different displays together. Even when I’ve seen the NMEA network fail in this configuration all of the displays still had their primary data.
Granted this may result in more displays than the other way around but multifunction displays with sensor ports seperate from the NMEA 2K ports are not that expensive or hard to find places to mount and you still have the ability to bring all of the data to one device where you want it.
I think that’s a great recommendation for a way to think about NMEA 2K systems.
I recently retired and am a less-experienced sailor on bigger boats. I’m a bit OCD and could really use some help. We bought a Freedom 39 pilothouse schooner because it seems to be easy to sail short-handed and it has a couple of features that may keep my wife onboard longer. All of the electronics were outdated or inoperable. I debated long about an integrated navigation system or independent components. I knew independent components made more sense but just being retired, cost was a greater consideration. I purchased Simrad NSS8 chart plotter with 4G radar (one each in pilothouse and cockpit), Simrad components to reestablsh the autopilot, Simrad depth finder transducer, and Marethon weather station with the thought of complete integration with NMEA2000 architecture. These components are installed but not sea-tested. There are already conflicts between radar and depth finder.
Since I know less about electronics than I do sailing, I am seriously reconsidering the integrated set-up. I would really appreciate advice on how to isolate some of the components without completing starting over. I think at a minimum I want depth finder, weather, and autopilot independent from the rest of the system. I know dead reckoning navigation and plan to have paper charts.
My plans are 95% coastal cruising with short open water passages. The boat is currently in south Florida and I wsh to sail it to NC after storm season.
This site is an incredible resource for neophyte and experienced cruisers. Thanks much for any advice.
Gary, two thoughts: if everything is Simrad, should not conflicts be the problem (under warranty) of either Simrad or the marine electronics installer? Secondly, is there a requirement for these displays to be networked? Can they not act as stand-alones? I like seeing raday overlays on plotter screens, but that’s about it, and that can be done on a relatively cheap laptop or netbook.
Thanks Marc. I don’t think every component must be networked -at least I hope not. Depth finder is Airmar – not Simrad – but spec’d for Simrad, so I guess I can get a separate monitor for that?
Any other thoughts about simplifying this seemingly complex arrangement?
The specifics of this are going to be hard to help with from afar, but what I can do is share some troubleshooting techniques from my days as an electronics technician.
The first thing I would do is isolate the radar-plotter from the rest of the system (you may need to buy additional network terminators) and get that working reliably. If you have trouble you will have to work with Simrad to get that sorted. Their support used to be great, but I have heard that it has slipped a lot in recent years. Still, if you keep at it, you will get there in the end. The key is to make sure that you have no non-Simrad gear connected to the system when you call, otherwise the tech will almost certainly cop-out by blaming the third part kit.
Next, isolate the autopilot from the rest of the system. This may require buying an additional fluxgate compass, but maybe worth it to get things sorted out. Also I don’t believe in connecting the autopilot to GPS.
Only after you have those two systems working independently and reliably, consider adding in stuff like the depth and weather station.
Assuming that the depth transducer is NEMA2000 you can build a freestanding system for that and maybe the weather station by buying a generic NEMA 2000 multi-function display.
Once you have all of that working right then you can consider connecting the independent networks together using a box like this. Or simply don’t bother and leave them separate to make troubleshooting easier in the future.
Hope that helps. Let us know.
Thanks so much John. I was thinking along those lines but I guess I needed some affirmation. Gotta deal with health issues first so this may take a while for results. Thanks again.
If have just posted a brain picking challenge under another heading, but it could also have fitted here. In order to obtain maxium exposure I post the link under this heading for you to find.
In summary; possibly a set of burnt out NMEA ports on expensive gear, where the burn is due to unwanted electrostatic discharge finding its way to the NMEA backbone – how to protect and avoid this in the future?
For N2K troubleshooting, one could keep in hand a minimum set of N2K components: 2 drop cable, 1 set of male/female terminal and 3 T connectors and a power cable. Less than$100.00
It would be allow you to rebuild the network starting from 2 devices, and expand from there.
It could be use also to setup quickly a basic network to get up and running with critical devices, such as AIS and chart plotter for instance.
There are a few strategies that can be used for troubleshooting, such as adding removing one device at a time.
Thanks Phil for engaging and the useful comments.
I do not find that debugging N2K issues like these ones is truly plain sailing. Have now decided to use a dedicated galvanic ally isolated DC/DC converter to feed the network, and power only the N2K network from this dedicated converter. Next to add a power surge arrestor between the feed from the convertor to the network. Then the bare drain wire (shield) of the network will not connected to battery minus as usual, but the hull at the same place where the power surge arrestor is grounded.
Here is a link to the arrestor http://www.ebay.com/itm/391240120493
Any additional thoughts and ideas are surely welcome.
Sounds good, with one exception. I really don’t like the idea of connecting the N2K shield directly to the aluminium hull. Way too many chances for stray current corrosion, and no way for you to know it’s happening until the damage is done. I would connect the shield through a capacitor, not directly.
To me, possibly compromising the integrity of the hull to make up for fundamental weaknesses and poor design in an a suit of electronics is not right—tail wagging the dog.
Thanks for reminding me about the capacitors, John, you are – as usual – quite right.
Best of wishes for Christmas and a splendid New Year!
You are welcome. In thinking about it, you may also want to rig an (additional) stray current warning device between this setup and the hull. With all these different boxes and many meters of cable running around the boat, together with metal plugs at every joint, truly keeping things isolated is going to be challenging and because of the 12/12 volt converter your normal warning device will not tell you that you have a problem. See this post for a good option:https://www.morganscloud.com/2008/11/01/electrical-systems-on-aluminum-boats/
Neither am I an engineer nor do I know electronics components. So maybe the answer to the question below is obvious, but I will still stick my head out of the cockpit hatch;
When your response about adding capacitors had sunk in, a thought came to mind. The idea behind connecting the NMEA shield/bare wire to the hull is to direct stray current spikes to the hull rather than let it find the NMEA equipment, and blow it to pieces. If a capacitor is added, is not also path of the stray voltage spikes to the hull also blocked? Hence the object of the “game” is not achieved?
Having read myself up a bit on SSB/HAM installations, I understand that it is good practice to connect a DC blocking device between the antenna tuner and the metal hull which acts as a ground plane. This stops DC from flowing, but lets the RF go through.
Me confused? Yes!
If and when you find time to respond, I am listening.
You are right that a capacitor acts as on open circuit to DC, that’s how it protects from stray current corrosion. However said capacitor will pass a spike to the hull, that’s the whole point in using one. The trick is picking the right size capacitor, but I don’t think that sizing would be super critical. Unfortunately I don’t know enough to specify the capacitor size, but any competent electrical engineer should be able to do so. Also any technician working with this stuff who is worth the name should know.
And the two situations you link to are simply capacitors, so contacting them to ask what size you should use may help.
Also here is an explanation: http://www.answers.com/Q/What_is_surge_capacitor_and_how_it_will_works
Very good ideas, thanks…although the very fact that it would be a good idea to do as you suggest further highlights the fundamental weaknesses of the N2K system.
Hi John, Petter and all,
Why capacitors? Do these in some way cap the energy transferred? I just know them as storage devices to smooth out power availability. How does this help?
Also, I have not found voltage spikes and related contamination a problem so why the 12-12v converter?
Thanks, Dick Stevenson, s/v Alchemy
Thank you for your interest in this issue. To answer the easy question, the house bank of the vessel is 24v so the Mastervold DC/DC converter takes this down to the required 12v. In addition it has galvanic isolation built in.
As for the capacitors, your questions give rise to some doubt (lack of understanding on my account), and hence I have posted a follow up question above.
See my answer to Petter. The key thing to understand about capacitors is that they have very high resistance, but low impedance, and that impedance is a measurement of resistance to a voltage that fluctuates rapidly, and that’s what a spike is. Further, the impedance of a capacitor varies with it’s size (measured in farads and the frequency of the voltage it’s being subjected to. More here, if you are interested:http://electronics.howstuffworks.com/capacitor2.htm
Hi Petter, Thanks for the heads up on the equipment. I should have, but did not, known that it was a 24-12v DC conversion. I believe that some use a 12-12v converter to ensure good clean DC power without spikes etc.
My best, Dick Stevenson, s/v Alchemy
No worries, Dick, as they say Down Under.
You appear to be more than knowledgeable, and since it is Christmas time, I forgive you for not knowing every and all details about every floating vessel that exists – including mine. If you pass by Norwegian waters, drop me a hint and a free tour, with beer, is available.
I spent almost 6 months of the last 2 seasons cruising Norway’s waters and leaving those waters induced more regret pangs than almost any other cruising grounds have done. I am sorry not to have the opportunity to link up, but it is always surprising what a small world the cruising community actually is. So I look forward to sharing an anchorage down the line.
Dick Stevenson, s/v Alchemy
Great initial post and comments to be sure. We are in the final stages of an extensive NMEA 2000 install and I thought I could share some of my experience. We have replaced all helm engine gauges by using Actisense EMU-1 analog to digital data converters and tied all navigation and safety equipment into the N2K backbone.
Here are some of the issues we had to overcome to gain confidence in the upgrade.
First, beware of any “NMEA N2K like”, “NMEA N2K compatible”, and “believe me I work with NMEA N2K devices”. Early on I purchased an AIS unit and had a commercial work boat level installer help me plan the unit’s install. That’s when he informed me the AIS unit was not “NMEA 2000 certified” but NMEA 2000 compatible and that using the non-tested device takes on a risk. That led me to the NMEA website (NMEA.org) where I found several articles on how non-certified equipment that has not been tested to the standard can cause problems. Anyone interested in reading up on this can find lots of info at the NMEA site. It is extensive but worth the read.
Next, is a redundant power tee. We installed two power T’s so we can switch from one power source to another if needed. We used an AB like switch so it can be done quickly. We have redundant battery systems configured so that one source cannot draw down any other source and a charging system that charges all sources.
Then, there is the cable. We opted to use all Maretron products. We initially ordered cable specified by a major marine electronics manufacturer. Before we installed the first foot of backbone we ordered cable from Maretron to compare. I will admit that we did not test the two cables, but the heavy structure of the Maretron cable was obvious. All of their connectors, multiport boxes, and field connectors are of very high quality.
Support: Maretron’s N2K Builder software has a bit of a learning curve, but allows you to plan a system from basic to extensive. Couple this with the ability to send your completed N2K Builder design to their factory where an engineer will review your plan and make suggestions or actual changes to improve performance makes this step one of the most critical. Plus, calling the Maretron factory for help is always pleasant and met with an experienced systems engineer.
We have attempted to build as robust a system as possible, but I must add that I have one camera in the engine room trained on the analog engine gauges and a small stand-alone plotter we use with the dink that is ready for action with a quick plug in.
M/V Belle Aurore
Great and very realistic analysis of the issues involved, thanks.
The take away for me is that NMEA 2000, while useful in the right hands, has a lot of caveats and dangers that we all need to be aware of, and every boat that installs it should do as you have done and have stand alone backups for mission critical systems.
at least there’s a huge relief at hand – everyone, even me, can now become his/her own certified NMEA specialist (and more), in only 4 days… you wrote already about that (https://www.morganscloud.com/2016/05/06/the-marine-electronics-industry-has-really-lost-the-plot-now/). So bus failures are now only a piece of cake.
No? Hm. Tell me why I am a bit doubtful about that…
Best wishes, Ernest
I noted the references to the CAM bus being so robust … as a matter of fact, the CAM bus appears to be the single most annoying (and costly to fix) issue on modern cars older than 5 years. For example in our Audi A3 it leads to intermittent stops of the engine or even preventing the engine from starting, when the only problem is one single faulty CAM bus cable somewhere.
That experience lead to me distrusting NMEA 2000. Convenient is certainly is, but as the article suggests, inviting catastrophic failures for minimal causes
That’s interesting given that, as you say, that NMEA 2000 is an outgrowth of CAM. Either way, there is no question that NMEA 2000 is anything but fault tolerant.
On a pedantic note I understand it to be Controller Area Network (CAN).
I’m very familiar with it in it’s industrial form DeviceNet that originated within Rockwell Automation but is managed as an open standard by ODVA. I sold the first DeviceNet system outside of the USA and it was very popular in the 90’s and on. (These days it’s been overtaken by Ethernet/IP which is essentially the same application layer protocol laid over the top of standard TCP/IP or UDP transport layers.)
The big advantage of DeviceNet is that it included a robust power supply in the network and was highly noise resistant. In our experience 99% of the problems related to people not following the physical installation rules. On top of this ODVA provided a strong service to ensure verifiable compliance with the standard and interoperability in multi-vendor installations.
Stick to the rules, and it worked.
By contrast I’ve no exposure to NMEA2K so far, but I’d be 100% clear that the installation quality remains the most important factor. Indeed the physical components are absolutely identical to DeviceNet. What is much less clear to me is how well the NMEA protocol standard is maintained between vendors; if this isn’t managed well then problems are almost certain.
If you are going to bundle multiple devices from different vendors, I’d look very hard at what each one offers in the way of compliance verification and support.
I’d be looking at DC/DC converters dedicated to the power supply on each network and have multiple networks associated with areas like electrical, engine/tanks and sailing instruments. At the same time I’d want to be certain that if a network did go down you could still run the boat in some fall-back mode. And I’d be utterly anal about the quality of all my connections and terminators.
I can understand the reservations around these networks, and clearly some people do run into issues. These are the ones you hear about, while the vast majority of stable, successful installs just fly unnoticed.
Lots’s of interesting points. To me the key take away is that sure NMEA can be made robust, but the problem is that few people are going to go to that level of trouble and expense to do so.