NMEA 2000 certification, the elephants in the room
It’s great that the NMEA’s magazine Marine Electronics Journal is available online these days, but I screen captured that two page spread above for reasons that may not please the organization. The “NMEA 2000, Doing it right” cover story of that May/June issue is a series of interviews with several dealer/installers, and while it’s pleasing to hear that the Standard is really happening — it’s part of most new installs, and seen as a good thing for all parties — I found the interviewer’s emphasis on avoiding non-certified N2K products a bit odd. And in the “Tech Update” seen above, my friend and NMEA Technical Director Steve Spitzer doubles down on this theme, which — sorry, Steve — deserves some debate, I think…
If you read the column you’ll find that Steve goes on quite a rant about manufacturers who market their products as NMEA 2000 “compatible” and suggests that products which haven’t been certified “probably won’t work on a NMEA 2000 network” and “may even disrupt the network.” And at the end he suggests that the industry needs to “work together to ensure that ‘rogue’ products do not disrupt the marketplace” and even urges readers to report such “rogue” non-certified gear.
Well, holy cow, that’s not hard! For instance, how about Raymarine, the big electronics company with the full page ad facing Steve’s column? If you go to NMEA’s handy database of N2K certified products and search on Raymarine you will find that the E Wides shown in the ad are certified but no C Wides, no autopilots, no instruments, not even the SeaTalk to SeaTalkNG/N2k translator that I’ve seen work with numerous displays. And you’ll find only six certified devices for Furuno, which is no surprise given our previous discussions, but I was quite surprised when I looked up Simrad and Lowrance. I have no idea why but neither the NSE/NSO/NSS nor the HDS Series are certified (even if I thought I saw “good N2K citizenship” in the underlying OS years ago).
In short, there are many non-certified NMEA 2000 devices out there, and a lot of them are made by three of the Big Four NMEA manufacturers. And note that the 412 certifications seen in the database screenshot below include numerous repeats, like different lengths of the same cable. Note too the newly added warning stating that non-certified gear “will not interoperate properly with other NMEA 2000 certified products.” Sorry but I just don’t buy that certitude. Sure, I’d rather have a certified device but many of the non-certified ones seem to be interoperating fine on my boat and many others. And guess what? Certification is not a guarantee of interoperability. For instance, though the Navico NAIS 300 is a certified N2K transponder, I’m pretty sure that it is not yet sending the standard Class B static data message as discussed recently.
I just don’t understand why NMEA is taking such a harsh stand on certification, especially when the “rogues” are in fact some of the main companies driving innovation in marine electronics. Steve Spitzer says the goal is provide consumers with “safe and valid market choices” but I don’t see how scaring them away from many decent products achieves that. I get the value of certification, for us and for NMEA, but I also get that while N2K has evolved in a somewhat messy fashion, it has gotten to a pretty good place, and is getting better every day. I think NMEA might do better with a carrot than a stick.
What do you think? But, please, there’s no need to spout about how NMEA and/or N2K are failures. Neither notion is remotely true. Check, for instance, the new Master Dealer Program seen in the sidebar below; seems like a great idea to me. And I can tell you that NMEA 2000 is even better than NMEA says it is! I just completed installation of five MFDs of different brands, two PCs, four radars, and many sensors all networked to two N2K networks, which were by far the easiest part of the job. Tests at my mooring looked good — despite the presence of several ‘rogue’ devices — but tomorrow should be an interesting real world test. That’s when I’ll try to show a WoodenBoat School navigation class what modern marine electronics are all about, while also cruising to Brooklin in what may be fog and rain. Wish me luck.
From the NMEA “Mission” statement, “NMEA believes that the long term profitable growth of the association is a worthwhile objective…….”
It would stand to reason that since they don’t make any money from “NMEA Compatible Devices”, they would bash them. If you give them enough money, they will no doubt certify you. Their goal is the fiscal gain derived from the promulgation of the N2K standards.
The standards are important, I just wish access wasn’t so expensive, at least for me.
While this isn’t consumer friendly, it is consistent with the core dynamic of the (non-indie) IT marketplace.
Consider Microsoft’s effort to block apps not compatible with Vista rather than fix Vista.
Consider Apple’s perspective on Apps vs Android and Amazon, etc.
What you can’t control, you try to freeze out without crossing the “restraint of trade” line. Don’t know what would make Marine Electronics immune.
And when the targets are long term promoters of proprietary standards and vendor lock-in, I’m not sure whether the feeling is schadenfreude or not.
It is so hard to reconcile NMEA’s approach to the standard with other successful standards.
Nevertheless, it works pretty darn well. My confidence level is very high that I can mount and wire (the nmea-2000 side) of a new device on my boat and go sailing the same morning if the backbone cable (or a “rogue” “compatible” daisy chainable NMEA-2000 product) is within reach.
Several curious points in this article:
– The title. What intellectual property is being protected? NMEA 2K is a market specific application of CANBus. It is not a breakthrough in networking technology.
– The major vendors themselves, who I assume are NMEA members, often are ignoring certification requirements?
Bill Bishop is correct. The is all about $$$ for NMEA. And that’s NMEA’s decision to make. But it has consequences. The relatively closed nature of the NMEA 2K club has resulted in a 10 – 15 year start-to-deployment cycle. That’s an incredibly long technology cycle versus other tech markets.
The cost of the 2K specs strikes me as high versus other ANSI, IEEE, and certainly IETF specifications. If NMEA is truly interested in rapid growth of a large 2K ecosystem, reduce the price of the specs and certification.
Ben is also correct that certification and interoperability are two different animals. How is 2K interoperability testing done today?
Sadly, the current mess is solely the result of initial lax enforcement of the physical layer “Standard”. Proprietary schemes, such as proprietary connectors and daisy-chaining were once tolerated and even condoned. Early Simrad and Raymarine products with proprietary connectors and embedded NMEA2000 “T”s for daisy chaining were indeed welcomed and fully “NMEA2000 Certified”!!
I remember the product manager at Furuno bitching about being denied full NMEA2000 certification with their new instruments because Furuno simply copied Simrad and Raymarine with an embedded “T” design in their FI-50 instruments even though Furuno used standard NMEA2000 Device net male/female connectors on their products.
After that incident, the big three were given told to comply or be forced out of certification.
I believe they all said SCREW NMEA because this is where we are today and it doesn’t look like it will change.
@Insider. I would not lay all the fault of the physical layer incompatibilities @ the feet of NMEA. Strong enforcement by NMEA would have helped, but the manufacturers need to assume some responsibility here. If the manufacturers made a good faith effort to standardize the PHY, it would have happened.
I have always been troubled by the marketing side of NMEA dictating policy for the Standards side. In fact, it seems inevitable that Marketing (and Politics in the larger arena) run headlong into conflict with standards meant to rein in the market hype and hostage-keeping techniques of IBM, Apple and other large consumer product manufacturers. In a better world, NMEA would divorce itself completely from all marketing, and survive exclusively on Standards promulgation and maintenance.
In the past I accused NMEA of pooh-poohing products that stressed ease of installation, by-passing the services of its installer members. That didn’t float then, but I suggest it again.
In fairness to the Association, I think Steve’s Tech Update article was focused on the distinction between those manufacturers who “play by the rules”, and those that try to go for a free ride on other’s investments/dues. Clever marketing and advertising can deceive a lot of folks that they are buying something they’re not. The companies that buy the standard, participate in the process, and adhere to design protocols have earned the right to market their products with full NMEA 2000 Certification, and the presumed sales benefits that flow from that. Those that can’t be bothered with that trouble and expense, and yet attempt to share in the benefits by advertising “compatibility” and hoping the buying public equates that with “certified”, are the target of the article – and I would agree that they do not deserve our support. I think that’s all Steve was saying on that page.
Of course, there are plenty of great products that are non-compliant, mainly because manufacturers’ desire to innovate was constrained by the current spec – daisy-chaining for example – but that’s a separate issue and points both to a failure of NMEA to update the protocol, and the member manufacturer’s to figure out a way to innovate within its boundaries.
As far as your observation that “certification is not a guarantee of interoperability”, I don’t agree. Just because a given certified device does not support a specific sentence or function, does not (in my mind) mean it lacks interoperability. There are many devices out there, both 0183 and N2K, that support multiple sentences/messages, but maybe not the specific one that is needed to communicate with another device on the network. This is the job of the customer/installer to select devices that support the features desired, and does not represent a failure to “interoperate”.
Hello Grant, you kindred soul! Part of my shameful checkered career was stone-walling people wanting my Agency to change its mind about something, and I recognize a master of that dark art! Like me, you have avoided the issue by retreating to generalities (rather nicely at that!) and painted over the nasty details with a broad brush.
Of course, someone is going to object with those pesky penetrating questions, but you and I know how to grind them down with breathy rhetoric.
Just for my personal delection, how are you going to respond to something like this: Many raymarine products are not certified, but some are. Does that mean that part of Raymarine is a slimeball imitator trying to lure a gullible public into assuming their stuff is interoperable, possibly to the same extent that certifed products are? Remember, you get points for boring the pants off you new penpals.
Not quite sure how to respond to your comments. I can assure you I have no background in stonewalling anyone, at a corporate level or otherwise, and my post was an attempt to clarify the issue, not avoid it. I’ll try again. 1) I think the purpose of Steve Spritzer’s article was to differentiate between those who support (i.e. purchase) NMEA 2000 certification, and those that attempt to associate themselves with the standard without supporting it. I think it’s a worthwhile distinction and I agree with its premise – protocol development and certification costs time and money, and companies ought to pay their fair share if they want to market their products as certified.
2) Your question about RayMarine is what I would consider an example of the other point raised in my post, a company who contributes and supports NMEA but offers some products that are non-compliant. The reasons may be a desire to innovate in ways that run afoul of the current specification, or to maintain backwards-compatibility with their own product lines, which cannot be done strictly in accordance with current NMEA 2000 protocols.
As long as different product lines from a single manufacturer are clearly marketed as being “compatible” as opposed to “certified”, I don’t see a huge issue with that. I don’t believe this type of situation is what Steve’s article was addressing.
Hope I haven’t bored you too much –
Sorry, I came down on you pretty hard. Rhetorical questions need not be answered. But you did.
So if Raymarine, and by extension, all the name brand major players are actually the good guys even though they sell products that are not certified, and other products that are certified but not compatible in some cases, WHO ARE THE BAD GUYS? Is it impossible for a customer to buy a perfectly functional N2K product that is completely compatible even if they never pay NMEA a penny?
Who is Steve pointing a finger at? Is there such a huge threat to justify this rant? I don’t see it. If it looks like a Duck, and quacks loudly at all intruders, it’s protecting its nest.
By the way, I am a member, just to support the standards with my dues.
Perhaps there needs to be a marine track at the IEEE?
I’d buy something that was IEEE standards branded over NMEA.
No apologies necessary. From your previous posts, I suspect you have a much closer and more informed relationship to the marine electronics industry than I do. In reference to your questions, however, I would respond as follows: 1) The “bad guys”, if you choose to use that term, would be any company that attempts to knowingly portray their products as NMEA certified when they are not, and thereby potentially undermine the standards while at the same time refusing to support the Association that makes the standard possible. Steve doesn’t name names – I certainly won’t either, mainly because it’s not my battle to fight. But the principle is nevertheless worth upholding, in my view. Whether it’s a “huge threat” or deserving of a “rant” would probably be better addressed by the author.
2) “Is it impossible for a customer to buy a perfectly functional N2K product… even if they never pay NMEA a penny”? I assume you are referring to the manufacturer, rather than the customer, regards not paying NMEA a penny. I’m quite certain it is very possible – just not very ethical. I applaud you for maintaining a NMEA membership, and I can only assume you feel your dues serve a worthwhile purpose. Shouldn’t ALL those that benefit from NMEA’s efforts share in the costs? That’s all this discussion boils down to, as near as I can tell. If that’s protecting the nest, so be it.
Yes, some non-certified N2K equipment works just fine. This is true in most fields. (My lab has some custom, non-certified software to interact with a particular vendor’s X-ray imaging equipment. Does it work? Yes. Does it crash? Very rarely. Would I trust it in a production environment? No.) And yes, the standard is expensive- but it’s a pretty tiny fraction of the cost of developing an N2K device. The certification cost, spread out over an entire production run, is just pennies per unit shipped.
I agree with Paul- there’s not much the NMEA could have done when certain major players split off and built proprietary, more-or-less compatible networks on the same underlying protocols. It was a stupid move on the part of those OEMs, who made a conscious choice to actively undermine an industry-wide standards process so that they could force vendor lock-in. Had the NMEA said “no, you can’t do that”, they’d have done it anyway. The most we can do about it as customers is to tell those companies “you effed up” and stick with gear that is certified (or properly compatible 😉 ) until they come around.
I find it odd that a standards body feels it must grow and be profitable- they’re not supposed to be companies in the usual sense. They only need to cover their costs and produce good, relevant work so that members feel they’re getting value for their dues. I can’t think of many other groups with similar purposes who think this way.
Well as an early practitioner of the dark art of EDP (1963 and onward) just let me say this. CANbus and NEMA2000 stand on the shoulders of TCP/IP standards which were developed and published as open standards. (Read the RFC’s) NEMA’s policies have set the development of marine instrumentation back decades and it will take decades more to bring on NEMA2000. The best effort to date is a project ‘gpsd’ which has implemented an open system interface in a clean room environment.
It is headed up by Eric Raymond most famous for his book ‘The Cathedral and the Bazaar’ which is a must read for anyone trying to understand where we are coming from. Remember the only real standard is the width of a role of toilet paper and that doesn’t fly in many countries at that.
Using free NOAA charts on free applications software (well $40 for PolarNavy this season) using free Linux o/s and complying with all copyrights and patents under the GPL. Read your Operating System License, it preclude use in life endangering applications such as medical or navigation.
Doug, a good post. I did read the essay The cathedral and bazaar –
and true to the spirit, it’s free.
Correct me if I’m wrong, but I think you’re saying open source speeds development, and closed source, along with copywrongs (I like the word) impedes. Reading some of the threads ( http://lists.berlios.de/pipermail/gpsd-users/2009-March/003643.html )I can understand the frustrations involved in trying to innovate, and the lack of access to information that should be available to the public. Thanks
A very strange article indeed, is it a ” coded” attack on some of the big four who have blatantly ignored the certification when it suited them. Who Are these ” rogue ” products. Do they mean most of say, Furunos product line.
Interesting you cant even get going without a manufacturers ID which is purchased through NMEA. So NMEA know all the manufacturers anyway.
Then there is the laughable issue of intellect property. Most manufactuers use a standard CAN reference design and a protocol stack live say Kavsers one. What intellectual property are they stealing.
In many cases NMEA brought this on themselves , the standard has been extraordinary slow to evolve. Witness the debacle over cable types, daisy chaining, third party gateways, AIS PGNs , digital switching PGNs etc. All this means that the industry had to go outside the certification to innovate. Then late in the day NMEA start complaining
I generally agree with Doug. Had 2k gone with a more open development process, it probably would have matured and gotten to market faster. For example, we,d likely have 2k drivers available for the major embedded and mobile OS by now. My understanding is that NMEA’s mission is to foster a vibrant marine electronics market. If that is correct, I don’t understand the still relatively closed nature of 2k.
One nit Doug. 2k, unfortunately, is not in any way based on TCP / IP nor any other IETF standard.
I feel INSIDER hit the nail on the head. Its not about stndards, its about greed. ON ALL SIDES.
Whoa Sandy! Greed ?
I have been on the board of a telecommunications research and standards organization funded by a core group of 12 members competing with each other.
These organizations have a director and staff that have aspirations of providing new and better services for their members, have rising costs just like any business, and need to grow their revenue a little just to tread water and much more than a little to grow their services. Whenever they can’t get that revenue, those aspirations must be delayed, which can be very discouraging. So to be in a position to do more for the membership, these folks need to put growing revenue way ahead in their priorities. Is that greed?
How about enforcment of a revenue model / insisting competing members are treated equally? Is that greed also?
Maybe if NMEA had more revenue and more resources from their membership, PGN’s would evolve as fast as the technology. Maybe if they had a different standards approach those resources would come from a voluntary effort outside the core membership. But, for whatever the disappointments that exist with their standards efforts, let’s please not suggest those folks are greedy.
Am I wrong?
Well said, Dan!
Correct my understanding, but I believe PGNs DO come from a core group of volunteers, or they are given time to work on them by their employers. And THEY are not the people I call greedy.
That’s correct, Sandy, but it’s the NMEA staff — particularly the Technical Director — who drive the overall standards-making process. They are not greedy people, and if they were, they’re certainly in the wrong line of work! I don’t even think “greedy” applies to the big manufacturers who are to some extent defying their own trade and standards association. “Competitive” for sure, but greedy, not so much.
This whole situation is pretty complicated (and I’m sure that there’s lots that I don’t know about). For instance, though we’ve heard small developers complain about the cost of NMEA 2000 documentation and certification, Steve Spitzer rightly points out that those costs have been substantially reduced, and, besides, for the big companies who already have some certified products, the cost of certifying additional ones is trivial. I understand how obvious issues like daisy chaining have caused some major products to be non-certified, but I don’t understand why some others lack certification.
But what I really don’t get is why NMEA has gone on the war path about certification when the real world situation is muddled. The idea that non-certified products won’t work with certified products while all certified products will work fine with each other simply isn’t true in my experience. In fact, what I find is that most of it works OK together in most install situations, and these days I regularly hear from installers and DIY boaters who LOVE N2K.
So I don’t think NMEA 2000 is in crisis. But it sure would be nice if issues like daisy chaining were cleared up. I’d love to see the day when most products were certified and NMEA was focused on expanding the standard. I just don’t see how exaggerations about incompatibility will get us there.
Nicely put Ben. Besides, I like N2K, and I really like having standards.
Ben , your comments are generally spot on nor do I think this is as simple as greed. There is and was no doubt a desire in NMEA to not allow the 183 situation to redevelop. That situation was where everyone made money on 183 but NMEA.
That said, this does point to a crises within NMEA. If the technical director feels the need to ” have a go” at what are his major sponsors. ( ie most of the big four) then something is rotten in the state of Denmark. It is self evident that’s who he his talking about because I see no evidence of small rogue manufacturers flouting the system.
We should not loose sight that it is we Consumers who are the chief beneficiary of any universal industry standards. While I certainly can’t dispute that there are problems within NMEA, we Consumers need to avoid wherever possible any equipment that is not N2K certified. Forget the “compatible with” or “as good as” claims. If we continue to buy this stuff, we are ultimately only shooting ourselves in the foot.
I agree, Dave, that much of NMEA 2000 is a reaction to 0183, maybe even an over-reaction. But I don’t think that it’s simply about money to NMEA. 0183 had no certification process, and no or little licensing fees to finance a certification process and an ongoing standards improvment program. 0183 also lacks a well defined physical layer while DeviceNet was obviously seen by some manufacturers as too well defined a physical layer. (Funny, though, how the claims that DeviceNet cabling and connectors were too expensive have not really been borne out when you look at the cost of the alternatives.)
Here’s another factor: NMEA 2000 was conceived during the “best of breed” days when many boats had large devices from multiple manufacturers, which meant that those manufacturers had good reasons to play together better on the narrow-band data level. But during N2K’s long gestation, the MFD evolved rapidly and with it, of course, the “own the helm” concept of marine electronics marketing. MFDs are great for boaters, I think, but the wicked competition does not encourage interoperability.
But I don’t think there’s a crisis here. NMEA 2000 is working pretty well for lots of installers and boaters, and it can take some of the sting out of changing major brands of MFD. And to my knowledge none of the big boys are pulling out of NMEA, even if that dire language about “rogue” products must grate a bit. By the way, there are a number of non-certified N2K products from small companies, but how can NMEA come down hard on them when there’s so many mainstream non-certified products out there? Just seems like a time for diplomacy and flexibility, not
a hard line.
I agree Ben, and as I have averred before, I don’t miss the old 183 technology, and N2K has worked well for me. It seems to me there must be a path here that is more beneficial to all. I like the idea of seeing a more open source environment (or at least a less expensive paywall), that still allows NMEA to manage the standards, and certifications. I have an idea that with more aggressive development, the general technology would be able to migrate into different technological environments. It would be interesting to hear Mr. Spitzer’s, along with the manufacturers comments, about these comments. You know they have to be reading.
The NMEA is basically a voluntary organization. From its inception they have relied on both manufacturers and dealers to develop standards and work tirelessly to improve a relatively small industry. As time went on it became obvious that member dues alone would not cover the costs associated with new standards or in fact any programs designed for the betterment of the business. You can’t just cut costs to operate a business, you must also raise revenues.
While 0183 and N2K are not perfect imagine a world where there was NO standard whatsoever. I can because I lived through it. But NMEA still has a way to go to rectify the inconstancies in their approval process. And they certainly dropped the ball with the concept of standard connectors, which I can assure you were specified when the BOD of NMEA first decided to move forward with the N2K project.
Finally, I have noticed that some of the industry people identified on Panbo do not seem to be members of NMEA, at least not according to the NMEA site. I do not see them under the dealer, trade, or associate listing. Perhaps it is an oversight on NMEA’s part, but if you are b*tching and sell/install marine electronics you should be a member, volunteer your time, and try to improve the organization. Or just be content to walk the docks and accept what comes your way.
Arnie, funny you mention joining NMEA. I just yesterday received the 2011 NMEA membership application in the mail. Here is the deal. For just $325 I can join as a dealer (Electronics Installation), along with the check I need to send a copy of my liability insurance (Okay), the company FEIN (okay, but why?), the number of both full and part time employees (Okay 1.5), and a list containing at least three major companies I am a authorized dealer for. Oops, I am certified by two, and not a dealer for any major manufacturer. So that doesn’t work. I don’t fit into the trade ($325), boat builder ($750), manufacturer ($1395) categories either. But I can pay $145 dollars to be an “Associate” with limited benefits (the site doesn’t tell you what the limited benefits are,) and no voting rights. I’m just one of the several thousand guys who have to install the stuff, make it work, train the clients how to use the exotic electronics, and repair it when it breaks. Not B!&ching, just saying.
If you check the NMEA website they list the various benefits to members. But they also state the following:
” Any questions on NMEA member benefits should be directed to the NMEA national office at (410) 975-9425, or you may email David Hayden (President and Executive Director) at [email protected] or Cindy Love (Office Manager) at [email protected].”
I think at the very least you should contact them and let them explain what you may or may not be missing. You might be surprised. I’m not sure you have to be direct with a manufacturer to join as a dealer since there are many members listed that are not. Either way they need to hear from you. You are part of the industry and your opinions matter.
Arnie, I did call NMEA, and all of the above regarding being a dealer is correct. The only benefit that an Associate does not get, is voting rights. I will send them a check for $145.00 to be an Associate. Tnx Bill
From my point-of-view, just like Arnie touched on, Steve should have perhaps started by explaining that the NMEA is a voluntary, non-profit organisation that uses the standards income to simply exist and continue to operate for another year. Non-profit is not the same as Takes-no-money-to-run. 😉
The NMEA only has a handful of full-time and part-time employees that are paid purely from the standards income. No standards income, no NMEA and no industry-wide standards any more.
Actisense, along with many other manufacturer’s have willingly given their engineers time for free to help support the various NMEA standards working groups, and so support the industry as a whole. The payback for us manufacturers comes much later when our products are released. We understand that without industry standards many manufacturers would not be able to innovate in this marketplace and the end user would have a far harder time connecting everything together.
The intellectual property that Steve talks about is the NMEA 2000 protocol: its message structures and communication methods. NMEA 2000 uses both CAN and J1939 as foundation blocks but the NMEA 2000 protocol (Higher Level Protocol) that sits above them is totally separate and unique (defined nowhere else), hence it is the NMEA’s IP.
Both of the issues stopping the big players better embracing NMEA 2000 (built-in terminators and daisy-chaining) are being enforced by the NMEA for safety reasons as they can both lead to bad installations and so damage the NMEA 2000 brand name. For example, the confusing danger of built-in terminators actually reared is head during a manufacturers ‘PlugFest’ meeting when a certain manufacturer’s product started to cause masses of errors on the bus because the engineer who had actually designed it (!) forgot that he had left the ‘jumper’ in so there was an extra terminator resistor in the middle of the bus! If the design engineer can forget about a hidden resistor what hope have installers got when the time is against them and the customer wants to leave port after extending the NMEA 2000 backbone?!
Andy, I don’t think this is the kernel of the issue. The intellectual property Steve mentions was the intellectual property of the participant companies.( as if people were reverse engineering their software ) I fail to see how rogue companies are purloining this given that anyone can buy it the standard and hence acquire NMEAs intellectual rights.
As to daisy chaining etc. Major manufacturers know what they are doing. Companies like Furuno are very unlikely to release products that don’t work well. No I dont accept that this article is about that. It’s more a general rant at thin air, because the situation now is that a large proportion of equipment actually isn’t certified at all ( Ben outlined some of them). What seems to have happened, is the big guys get one or two products certified, which validates the hardware and software of their CAN interface. Then they don’t bother with the rest. It’s a two fingers to NMEA, yet these companies are key contributors to NMEA and the standards development That’s a crises in my view.
Dave, as I see it, buying the standard is just the first step. Only by the final step of NMEA 2000 certification does a company have the right to use the IP.
There are a few small companies buying NMEA 2000 certified products from OEM manufacturers, and then selling them as their own NMEA 2000 products without having them recertified with their own details inside. Perhaps this was the part of the issue?
The NMEA has never said that making a device daisy-chain-able would make that product “not work well”. The fear is that whilst the initial installer of the NMEA 2000 backbone will have carefully mapped the whole network and not misused the daisy-chain devices dual drop connections, future installations (when the backbone routing is not so clearly understood – because of time restraints for example) could misuse these dual drop connections leading to loops being created in the backbone. These loops can create errors/miscommunication and the resulting troubleshooting could damage the NMEA 2000’s name for robustness – something all CAN bus networks are proud of.
Andy, although you make some good points why daisy-chaining might create issues, there is one very important aspect not mentioned and that is the controlled impedance aspect of an N2K network. An N2K network is based on a drop line to an individual product for very good reason and the most important reason is that the signals simply end at a product and never travels “through” the product. If NMEA allowed daisy-chaining (i.e. signal travels “through” the product), then NMEA would need to certify the impedance “through” the product so that products down the line received an appropriate signal.
Now this isn’t to say that a small network might work just fine with daisy-chained products, but as signal integrity becomes degraded with larger networks, I for one am glad that NMEA has stood their ground on daisy-chaining and we can rest assured that large networks will work as well as smaller ones.
Anon, the real question is “what ground is NMEA standing on?” More discussion here:
Bottom line: In many ways we’re beyond theoretical discussions about the dangers of daisy chaining. It’s in use on many boats. Three out the four major members of NMEA decided to suffer the consequences of manufacturing and supporting daisy-chained instrument displays. Has their lack of certification done more to protect consumers or confuse and annoy them? It’s not a crisis, we’re muddling through, but this is an issue that should have been resolved within NMEA a long time ago. I don’t see how ramping up the language about non certification will help.
There’s no compelling reason to not certify a limited number of daisy chaining devices. All they are in effect is ” dual drop” devices. Equally internal software selectable terminators are ate good idea. Clearly there ate big holes on the NMEA ad it pertains to real world installation. Nmea obviously never thought about devices on top of a mast etc. This problem is down to names inability to move a 20 year old standard along , to innovate and ensure that they stay current.
I heard similar “it works w/o being certified” comments when I setup wired Ethernet networks decades ago. We rarely saw problems at the outset of the implementation, but regularly saw problems making the network slow and unreliable after connecting additional devices.
Some non-certified devices pumped out 2 packets of info in the time-frame allocated for a single packet. Some cable lengths that exceeded specs were fine until the data highway got so crowded that we documented multiple ‘packet collisions’ during ‘data rush hours’, meaning that everyone’s network speeds and connections degraded and were ocassionally lost. Some of my clients engineers added enough devices, gateways and repeaters to their portion of the corporate net, again exceeding spec. This often extended problems beyond their own subsection.
You might worry about viruses from the outside on your computers today. In 3-7 years, you’ll be worried about what your neighbors boat (on the slip or on the race course) is doing to YOUR boat’s WiFi network.
If you don’t mind the massive irritation you may experience as you blame your vendors, installers, sales reps, nearby boat owners, harbor WiFi, mag writers, me and perhaps yourself … and can afford to trash everything in 2 years, don’t worry about installing certified products. After all, it works now, doesn’t it?
Actually nobody I beleive is supporting non-certification. What I have suggested is that NMEA are way too slow and rigid in their thinking and hence in their specs (for example there’s no spec on gateways or internetworking, unlike J1939).
Nor has NMEA published a robust defence against limited daisy-chaining. This leads to obvious frustration amongst its member companies and hence the problem. What’s a little rich is NMEA crying about it.
OK, after long last I feel compelled to chime in… I am a servicing dealer who has more experience with NMEA2000 backbones and implementation than most. Suffice it to say that we work regularly with systems large enough that we have multiple backbones on the same vessel and have eagerly awaited the bridge product allowing them to share data. We are also (to the best of my ability to discern with manufacturers and the NMEA) the only dealer that has done a large enough installation to require the use of two isolated power supplies with the DC return tied to a common point in order to comply with NMEA’s installation standards.
Additionally, I am intimately familiar with systems comprised of only certified products which did not work well because of improper design and poor installation; systems that worked very well with a mix of certified and non-certified products because of proper design and installation; systems that took down connected CANbus power systems in an unsafe maneuvering scenario because a non-certified device sent an improper polling request NOT from itself, but passed it on from a piece of equipment tied to it’s own internal OEM “compatible” CANbus arrangement. I hope this validates my experience level to participate in this discussion specifically about the marine application of NMEA2000 backbones onboard recreational vessels.
I will not address some of the issues raised above, some of which are valid questions/concerns, many of which are not…
The simple fact is that certification MEANS certification… not “sort of”, not “probably will”, not “should” but WILL… absolutely… everytime… and it is a VOLUNTARY standard, a manufacturer can choose to do it or not… if they want to advertise that a product is certified (obviously of value or we wouldn’t be having this conversation), then they should certify the product…
While CANbus is an open architecture which comprises the physical layer and some of the transport protocols that NMEA2000 utilizes, the NMEA2000 standard by intent and design GUARANTEE that it will do what it says it will do… EVERY TIME… Specifically it guarantees priority messaging gets through (read that as your throttle control tied to the backbone generates a signal which overrides the other non mission critical data which may be clogging the network). Without this GUARANTEE in place, how can we GUARANTEE safe operation? If this was the intent and design of NMEA2000 from inception, understood and supported by the member and non-member manufacturers that graciously supported the development of the standard from their internal engineering resources cooperatively with their competitors, what has changed now? Are we willing to say that we know, unequivocally, that non-certified products at all times and in all situations will not create a situation like the one described above where a ship lost all power due to non-compliant issues? I believe the NMEA IS prepared to guarantee that performance for certified products.
With respect to daisychaining, IN ANY FASHION, until someone can GUARANTEE that it will never create a problem, under any circumstances, then it simply cannot be allowed. In addition to my “experience scars” listed above, I have personally poured about 1/4cup of water out of a “waterproof” display from a major manufacturer that was a dedicated NMEA2000 display powered from the backbone. This unit simply could not work and was damaged beyond repair. However, BECAUSE the unit was connected directly to a drop AND only approved cables were used in the design & installation, the operation of the backbone was not affected in any way. Nor did water wick through the cable to the backbone. Design intent achieved!
Now the simple question is should a standard that:
(1) describes a method of sharing navigation and peripheral data between mission critical and non-mission critical components;
(2) prescribes specific methods of prioritizing that data for safety and reliability;
(3) includes a certification process to guarantee performance of (1) and (2);
(4) references a separate standard for proper design, installation, and deployment of systems;
be changed to accommodate a less expensive product or a product that is easier to install? I think not.
If you think that it should, what do you think the US Coast Guard will say? IMO? DNV? ABS?
Fortunately the standard was not designed to get as many products to as many people for as low a price as possible.
Unfortunately, it appears that some people have lost sight of this fact.
Well it’s a fine post and you have undoutably plenty of experience. But I think you miss the point. NMEA does not guarantee that the devices will interoperate or in fact effective operate at all. What it does is confirm that device is within spec and that it doesn’t cause problems. ( protocol certification). It doesn’t in practice prevent rogue products. You made this observation yourself
Secondly, NMEA 2k is aimed at the non commercial leisure market in the main. This means competitors DO have to produce for the cheaper more mass Market. Hence they have to innovate And produce more feature rich systems then the next guy. Despite what you say 2k isn’t anywhere near mission critical quality. I don’t believe there are any 2k electronic throttles for example. I for one won’t want it either.
Unless standards move with product innovation. ( say like the RFC process on the Internet )then you get non certified innovation and who can blame
I appreciate your post too, A. Insider, but a good example of Dave’s point used to be the Furuno NavNet 3D MFD. It is a Certified NMEA 2000 product but after a while it was noticed that it could mess up large N2K systems because it overused a utility PGN that queried other devices for their particulars. I saw it myself in bench testing: http://goo.gl/3nYir
Furuno has fixed that, and maybe the certification software has been modified to look for this issue, but meanwhile Furuno is selling other products that purportedly pass certification software testing fine (which I believe) but can’t get certified because of daisy chaining or termination issues.
This is messy and is not going to get fixed by a demonstrably false claim that non-certified devices will not interoperate with certified products. I hesitate to suggest a fix because it’s really between NMEA and the manufacturers, but how about something like another slightly less respectable certification level, or a big, bold daisy chain warning sticker — some way to bring more products into the certified family while still honoring those companies that never strayed from the rules. Meanwhile I think consumers and installers are generally muddling through OK with the knowledge that the certification process has not turned out quite as well as hoped.
By the way, I don’t think there will ever be a guarantee that daisy chaining won’t fail, but that’s true of all electronics and other boat systems as best I can tell (I’ve broken plenty ;-). Also, I would never defend a company that falsely claimed NMEA certification; that’s dishonest and perhaps even illegal.
Interesting post, but as a counterbalance I think it is worth noting that the requirements for commercial shipping (large SOLAS vessels, not small fishing boats or river ferries) is good old NMEA0183, repackaged within ISO standards.
NMEA0183 may not have all the certification of NMEA2000 that you believe in, but it is good enough for a supertanker that may be using an ECDIS system and carrying no paper charts.
Yes, equipment needs to work as specified, but certification doesn’t guarantee this unfortunately, and equipment specifiers and installers quickly learn which kit works and is well supported, especially in the relatively slow moving world of marine electronics.