The NMEA Mystery

[Sandero]

Just thought I would relay a tale about my experiences with the NMEA 0183 standard.

This was invented about 20 yrs ago more or less to establish a language for different manufacturer's instruments to "talk" and "listen" with other instruments.

There are many bits of data, but they developed a bunch of sentences which contain the data and the listener can "extract" it and "use it" to display, plot or drive an autopilot for example. You probably know all this.

Back in the 90s I put this to the test and managed to get three manufacturer's instruments to have successful "conversations" and there was kind of a "gee whiz.. this sh*t works!"

Fast forward to the new mellenium and I decided to upgrade instruments... principally a chart plotter MFD by Raymarine (C80). The other devices remain in the system... KVH sailcomp 103AC which is also a "cockpit repeater" of XTE, BTW and DTW which sends heading info TO the C80 and a B&G cockpit repeater for position fixers which can supposedly display up to 12 different bits of data.

The main ones on the B&G being COG, SOG and XTE, BTW and DTW and then you can program one button to display some other data .. I used to leave it on TTG (time to go to waypoint) I found this data up in the cockpit very handy... especially once a waypoint was programmed.... because I did not have to go up and down to see if we were "on course" etc... The autopilot is set from the cockpit so I can observe the data as I steer. Makes sense! Today people are installing "pods" with MFDs in the cockpit these days.... but I have a "dash" type instrument panel above my companionway with a row of displays. It works and it is readable from anywhere in the cockit.

My first shock was when the B&G would not report COG or BTW accurately. It displayed 2 digits and a decimal place. So 105.6° read as 05.6° which made little sense as it could mean 5.6, 105.6. 205.6 or 305.6° The old display dumped the decimal and rounded off... so it would read 106°. Weird ha? The TTG did not display at all and the KVH would not show XTE data!

I was told that the equipment was too old to chat nicely... so I turned to a Brookhouse MUX which is a programmable NMEA multiplexer. This device allows you to do lots of stuff, most of which I do not understand, but you do it by connected it to a computer and "programming" the NMEA data stream. Brookhouse programmed my decimal place problem away, but not the TTG or the XTE on the KVH. They said if I could "read" the sentences from the devices with a PC then it WAS possible to get that data back with some reporgramming.

When I decided to add AIS to the mix... I lost the required baud rate of 4800 for my NMEA network. Without some fix I had to choose between the AIS and the NMEA repeaters. Raymarine had a device which offered promised - a NEMA Seatalk bridge.

Seatalk is the "daisy chain" propriatary network for Raymaine... but it is NOT NEMA.... The Bridge will pop out and allow you to pop in NEMA data essentially "through" the Seatalk network. So for about $150 I had a possible fix.

The AIS got the C80 NMEA port at 38400 baud and the KVH heading went to the NMEA on the Bridge and the NMEA output of the Bridge went to the MUX mulitplexer and on to the cockpit repeaters. Mind you this involves lots of tiny little wires connect to "eruostyle" busses/terminal blocks. What a mess!

So I do all the little wires and fire everything up... and lo and behold... the heading data DOES get into the C80 (as promised) and the 4800 baud data DOES go to the MUX... but the B&G now does NOT display BTW or DTW but it DOES display TTG. The KVH seems to be working fine as a repeater. WOW this never ends...

Brookhouse tells me I need to "find the BTW and DTW" data in one of the sentences of the Bridge... because it has decided to select one of many ways to communicate this sentence... which the KVH CAN read but the B&G cannot! Is this insane or what?

So now I have to get a laptop out to the boat and the software to "read" the sentences from the Bridge and compare that with the sentences from the C80 at 4800 baud without the Bridge and then reprogram the MUX to do the BTW and DTW sentence that the B&G likes. Is this possible? Dammed if I know.

I wonder how many other sailors find themselves sucked into the NMEA 0183 magical mystery tour? Oh... BTW the AIS is not working with the C80 yet... still have to figure that one out!

Break.... more to follow....

Jef
sv Shiva
Contest 36s

[markpj23]

Wouldn't it be nice if the marine electronics manufacturers would just embrace the IEEE Ethernet Standards? Imagine being able to connect all devices using common hubs / switches.....

But it's far more profitable to add proprietary extensions to ethernet and sell it under your own name... or is it really? Imagine how many more units they could sell if it would just communicate across a standards-based network...

[hellosailor]

Jef, you've been misled. There are several different versions of the NMEA protocols, and the key to understanding the chaos is to find out EXACTLY which levels each piece is set to use, and is capable of using.

http://www.nmea.org/pub/0183/ will give you some idea how many levels there now are.

After that, the only question is whether the brain boxes you buy are capable of translating the strings to what the other boxes need.

Frustrating, complex, but really, just a matter of digging up the exact specs and then finding out which are compatible.

[Sandero]

Hello,
I hear ya, but in my case I knew I was trying to get instruments to chat which are from different generations and I was aware that the NMEA has advanced in its "versions" and now there is a NMEA2000.

My deal was simply to try to get this group to speak the same language... even if they spoke different dialects. I did not want to rip out what I had and so I am trying to get this to work. The solution may be in the programmable MUX which may be able to sort of the dialects and spew out one readable language for my "old timers".

Jef
sv Shiva
Contest 36s

[Starbuck]

Let's see:
Compass. Knotmeter. Depth sounder. VHF radio. Paper chart. Hand-held/bulkhead-mounted GPS will come this summer.

I'm pretty happy.

I'm not unsympathetic to your struggles. I just don't want to get that involved with electronics.

Good luck on your project, Jef.

[Pblais]

Jef,

A great post. Actually when you said "about 20 years ago" It was far longer than that. NMEA is actually a bunch of vendors that paid to set this standard up. They all pay for the standards docs (you can buy them too) to fund this thing and guys get junket trips on their companies to meet and argue this all out. Back in 1970's this was a popular thing to do. Sort of a not quite the IEEE but more a matter to keep an eye on the competition.

You've really found all the rat holes - there really are not many more just more sentence varients and non statndard almost a standard interface. The standard supports 3 listeners and one talker. MUXes are not in the standard but you can get them - but they can't be standard. Good news is at 4800 baud you can buffer a lot of data. When you get to modern stuff it gets harder.

Now we get to NMEA 2000. Sorry folks this one is far worse and might not actually really get off the ground at all. The standard is better but darn few heavy weights are in as far as delivering stuff. Its' 2006 and there isn't a lot right now. RayMarine is 100% SeaTalk2 for it's hot new gear. If that does not sound like NMEA 2000 then there is a reason - it's not.

They really would rather put the other guys out of business rather than use a common inteface. They bought a whole lot of the companies that used to be in NMEA 0183 and some of those bought others before that. It's a smaller playground now.

No one is going to use it that is a major player. They really want to beat each other up and own the whole market. Inter-operating with different vendors really won't be that possible once bus speeds go above 4800 baud. The other trend is the number of devices is shrinking. Things are being built into each other. A small computer with lots of inputs on a high speed data bus with a bright sunlight capeable LCD display. They all are doing it but not together.

[Rick]

Hey Paul,
The basis of the Nema 2000 protocol is not only hardware as well as software and firmware it is a system that has been proven in other industries to be very bulletproof, so do not denigrate the system design as a whole. The hardware basis is extremely robust which is why it is being implemented in the automotive industry (the beginning) as well as other industries.

The "problem" with Nema is that there has not been a universal adoption amongst the manufacturers due to specific application regarding to bandwidth requirements. Within the automitive markets the bandwidth applications have been very stringntly defined and catered to. Not so with the wishy-washy marine industry which is less than 1 percent of the automotive industry gross possible income.

So, do not condem the Nema 2000 as being "far worse" than what you did not define because the basis is quite robust.

[jstevens]

Jef,
I can save you a little work on your NMEA magical mystery tour. Don't mess with AIS on the C80. The Raymarine AIS capability is not ready for prime time. If you do get it to work you will regret it. Raymarine has not provided anyway for you to control the alarms the AIS processing generates. I tried using it cruising from Almerimar to Cartagena, Spain. I spent most of the time clearing bogus alarms from the screen so I could at least see the chart. If you do go ahead and connect it up and get it working, when you see the "AIS Connection Lost" alarm come up (about once every 5 minutes if there is not a lot of traffic in your area) don't go trouble shooting your wiring. What this really means is no traffic has been received from the last active target for 2 minutes. It is not an alarm condition. When i asked Raymarine how to disable this alarm they said disconnect the C-plotter from the NMEA network!

So my advice is put the AIS work on the shelf for another year or select another means to display the data than your c80.

John

[Sandero]

John,

When I did the first wiring of the AIS and fired it up I was in an area where there was likely no AIS signals. (Dering Harbor, Shelter Island, NY)

Since there are no idiot lights I have no way of knowing whether the antenna was OK (it was an old metz VHF), the antenna connections.. or my data connections from the AIS to the C80. All I know is the C80 said NO AIS in the little icon at the top of the data bar. I suspect at the very least my data cable connections were at fault at the very least.

I didn't bother to move into this more deeply.... had to leave the boat for a social engagement. But I would like to know that the NASA works and that it IS exporting signal to the C80. I can wait for upgrades and leave it at that. I will see what happens when I get a new data cable for the NMEA input.

Then I will pester Raymarine for the upgrade.

Jef
sv Shiva
Contest 36s

[DeepFrz]

Rick, I think you have confused CAN and NMEA 2000. CAN works, has some standards and is being adopted by some of the large power boat manufacturers. NMEA is pretty much a farse.

There was a good discussion of NMEA 2000 on SSCA, but you will have to look at the old info. It was called "NMEA 2000, flim or flam". I did some pricing and found out that a 3 instrument (or was it 4) system using NMEA 2000 had over $600. US worth of cable and connectors. And these were the Mini cables and connectors. To find out if you were running out of capacity on the Mini cables (or Maxi cables (sic)) you had to spend more $$$. I think like 800 ot them. This $$$ bought you a system tester/analyzer. You can't see the NMEA data on a regular computer any more, like you could with 0183.

Simrad is using (sort of) NMEA 2000 and a proprietary bus and Furuno uses Ethernet and NMEA 0183. Unfortunately I'm not buying any marine electronics soon, but if I was I would not be buying Raymarine. I have heard many bad things about them..

Deep

[hellosailor]

I'm not sure how critical timing issues are on USB, but since both USB and NMEA2000 are "just" serial data protocols, and both use one data pair and one voltage pair, and USB is way faster than NMEA2000 (although it can't handle anywhere near the max amperages)...

I would expect an "NMEA2000 to USB adapter" would be feasible and some tinkerer might even put one on the market. At that point, all you would need to read NMEA2000 would be a simple terminal program that spooled the USB data stream to the screen.

Maybe one of our high power electronics guys could take a closer look at this and tell me if it could or couldn't ever happen.

[Pblais]

I would expect an "NMEA2000 to USB adapter" would be feasible and some tinkerer might even put one on the market. At that point, all you would need to read NMEA2000 would be a simple terminal program that spooled the USB data stream to the screen.

Don't confuse transport layer with data layer. USB is a non issue at this point. In the bigger picture the issue is just which devices can you connect with? There just arer not many NMEA 2000 devices you can buy. Then after that what can you do with them and who else is interoperable?

In the marine world there is so little desire to be "standard" that there is no standard at all. You won't find and the NMEA 200 standard mentioned in most ads. IT's a 6 year old standard with few products. It's more a matter of who can integrate enough and control distribution enough to squeeze out the smaller players. That strategy has been working quite well. The folks at Raymarine see it that way clearly. Why not beat up every one else and get all the money.

Once you get into high speed data bus technologies the "adapters" get to be very expensive. They have to capture many times faster to buffer the data so they can deal with it and reprocess it and perhaps do more before transmitting it in a protocal some other device can deal with. Now you are talking computer grade bridge routers that can bridge multiple transports and perform protocal translations. The adapter just got very expensive.

Frankly, other than radar overlay the old bandwidth of NMEA 0183 isn't all that bad or at least faster wouldn't make it much better. When you start combining radar and sonar fish finders the bandwidth does get a little bit off the scalle relative to 4800 baud.

The new concept is central unified display with multiple remote sensors. Everything all at one time on one display. It's not really a bad thing but more a matter of you can't stay in the family if you you don't stay in the brand. It's the technical solution that meets the sales goal. Even the CEO can't ask for more than that.

[coot]

Originally Posted by hellosailor

I would expect an "NMEA2000 to USB adapter" would be feasible and some tinkerer might even put one on the market. At that point, all you would need to read NMEA2000 would be a simple terminal program that spooled the USB data stream to the screen.

Such devices already exist, but you have to search for CAN bus adapters instead of NMEA 2000. Some are USB, some are RS232 serial.

NMEA 2000 transmits packets (i.e. groups of bytes) over a Controller Area Network. You can buy off the shelf devices that your PC can use to act as a node on a CAN bus network. Some of them are obviously designed for somebody who wants to monitor the network and figure out what it is doing. Whether you could make any sense of the data you monitored is another question.


p.s. Yes, CAN really does mean "Controller Area Network", as weird as that sounds.

[Sandero]

One of the problems I ran into was that even the NMEA 0183 standard is by no means a guaranty that instruments will be able to communicate. In my case I was interested in some basic bits of data being displayed / communicated between instruments. What I am learning is that those "bits" are embedded in NMEA sentences and several NMEA sentences might contain the data and do it in a different "way".

So you might want to "read" BTW from a position fixer like a GPS... on a cockpit display (my case) and the display looks for this data inside of one sentence but the GPS is sending out another sentence which contains the data... so it is "missed".

The ironic thing about my particular experience with the C80 is that the data I am looking for IS available at the NMEA buss at 4800 baud, but not via their Seatalk interface which claims to make available all data by sending Seatalk and "translating" it or bridging it to a NMEA port. This turns out not to be accurate. Because if may be there, but it is now encoded in a "different" and in my need... unreadable sentence. Weird ha?

Jef
sv Shiva
Contest 36s

[Rick]

Deep, the NEMA 2000 proposed standard (it is still dynamic and not fully adopted) has the hardware protocol based upon CAN hardware protocol which is very robust. What is often confusing is that the automotive industry (where CAN is used) has several different standards relating to bandwidth and priority, etc. to serve needs from high priority like anti-lock brakes and low priority like door locks and the two CAN structures are not tied to each other for integrity and reliability.

The problems associated with NEMA 2000 are largely "political". The economic problem with the the NEMA 2000 is the choice by Teleflex to use very high quality connectivty which cost more than other implementations yet are quite good. Yes, the price will probably come down because industry uses similar connectivity yet you do get what you pay for in this case.

I work with another consultant who specializes in creating portals between various standards such as the NEMA 2000 and USB with the attendant problems. The marine industry is peanuts compared with other industries using similar standards of interfacing electronics and control and we are not likely to see a universal adoption of any one standard any time soon.