NMEA 2000: Based on CAN, but why so flaky?

[HighTemp]

The boat we sail on had been having major issues with its instrumentation. I have an electronics background ( years ago as a technician ) and volunteered to help isolate the problem. From the beginning the only thing that appeared to work was the chart plotter / GPS, everything else on the network was not reporting when you tried to perform an source auto find. During my troubleshooting, I noticed that the VHF head unit would power on erratically. When it would not power up, the instruments did not report, when it did, many were there. So after disconnecting the NMEA 2000 cable between the head unit and the NMEA bus, all instruments came back and appeared to be functioning as expected. The VHF head unit now powers up without fail.

I have some limited experience with automotive CAN bus systems. When something goes bad, the entire network doesn't drop as one typically gets an error code sent back to the ECU. In other words, Automotive CAN bus systems appear to be far more fault tolerant than does the NMEA 2000 system.

What accounts for this difference? I suspect that government regulators would never tolerate an automotive standard that acted this way, especially since most new cars are all "fly by wire" and a system wide drop would result in almost certain accidents.

[smac999]

which radio is it? I think there was one that had a lot of issues on nmea2000. one of the simrads I think?

[Paul L]

Automotive CAN busses are not installed by amatures either. Is the backbone properly terminated and the instrument drops correct?

[250224]

Originally Posted by HighTemp

Automotive CAN bus systems appear to be far more fault tolerant than does the NMEA 2000 system. ... What accounts for this difference? I suspect that government regulators would never tolerate an automotive standard that acted this way, especially since most new cars are all "fly by wire" and a system wide drop would result in almost certain accidents.

NMEA 2000 is 'proprietary crap'. A 'standard' only accessible for the few who pay the hefty price and then are bound to the condition to not betray anything.

Such a standard should be openly accesible on Github! In contrast 'we' have a nmea.org website from the Stone Age and hidden protocols.

As for your question: I think with NMEA 2000 one is supposed to call for a professional if something doesn't work and not to play around by yourself...

NMEA (2000) is maritime cancer :-(

[HighTemp]

Originally Posted by smac999

which radio is it? I think there was one that had a lot of issues on nmea2000. one of the simrads I think?

It's the Simard RS35.

[HighTemp]

Originally Posted by Paul L

Automotive CAN busses are not installed by amatures either. Is the backbone properly terminated and the instrument drops correct?

Yes, the backbone is terminated properly. All of the drops appear to be correct.

The entire network architecture just makes no sense. By design, one of the benefits of a multi-drop system is that when one sensor or instrument goes out, it should, in theory, just drop off the network and quit communicating. It certainly shouldn't drag the rest of the network down with it.

Perhaps the VHF NMEA 2000 port had a driver failure in which either the low side or high side driver shorted. Since they are not isolated from the bus, I suppose it would pull the remainder of the network down as well.

There has to be a way to physically isolate the instrument from the bus to eliminate this possibility, while still passing the data back and forth.

[Lake-Effect]

Originally Posted by HighTemp

There has to be a way to physically isolate the instrument from the bus to eliminate this possibility, while still passing the data back and forth.

You seem to be avoiding the obvious answer: there's something wrong with the VHF head.

You could try:
- borrowing a NMEA2000 bus analyzer or laptop interface and using that to observe the network
- take the VHF to a dealer and try it on their demo NMEA2000 network.

There's some mention of RS35 N2K interface issues here, apparently corrected by a later software update.

[HighTemp]

Originally Posted by Lake-Effect

You seem to be avoiding the obvious answer: there's something wrong with the VHF head.

You could try:
- borrowing a NMEA2000 bus analyzer or laptop interface and using that to observe the network
- take the VHF to a dealer and try it on their demo NMEA2000 network.

There's some mention of RS35 N2K interface issues here, apparently corrected by a later software update.

Quite to the contrary. The VHF head unit definitely has an issue and will be sent for repair, or replaced altogether.

My question is more focused on if there may be a way to mitigate the effect that any of the transducers or other modules connected to the network have on the network if and when they go down. Should they break, they should just stop communicating, fall off the network as it were. They should not disable the entire network.

[Lake-Effect]

Originally Posted by HighTemp

Should they break, they should just stop communicating, fall off the network as it were. They should not disable the entire network.

Agreed, it doesn't make sense that one device failing kills the network. Bad design choices in the radio's N2K interface, maybe.

Do let us know what the radio's issue is/was.

[barnakiel]

I would assign any (if present) differences to the fact that too many manufacturers are pushing too hard to assure that the consumer buys THEIR hardware only. This starts with different plugs, settings, use of proprietary sentences, etc.

If it were my choice, I would buy only from companies that offer N2K, not re-branded versions of N2K. But I am not sure this is doable at all.

b.

[HighTemp]

Originally Posted by barnakiel

I would assign any (if present) differences to the fact that too many manufacturers are pushing too hard to assure that the consumer buys THEIR hardware only. This starts with different plugs, settings, use of proprietary sentences, etc.

If it were my choice, I would buy only from companies that offer N2K, not re-branded versions of N2K. But I am not sure this is doable at all.

b.

Your observations align with mine. An industry standard should be a standard. Like USB or HDMI. If a device is N2K compliant, it should operate on any N2K network regardless of manufacturer, otherwise, can it really be labelled as N2K compliant? The different connectors alone, make things more difficult than they need be.

[OS2Dude]

I've installed NMEA 2K on two boats I've owned now. Not fancy setups, but multiple senders and displays and they worked with no issues. Even with the NMEA 0183 connection to my VHF.

I would do as suggested above and borrow another VHF (Not the same brand/model) and see what happens. I would assume if it is somehow corrupting the data it would not 'drop off' as it might with an actual electrical issue.

I really like my NMEA 2000 setup. Having been in IT, I have seen what happens to 'OPEN SOURCE' protocols. Everyone and their mother bastardizes them to suit their own needs and soon nothing works together anymore.

[Richfind]

This is more of a general answer than a specific response to this problem, though it is a possibility. I have not had any problems with the fairly extensive N2K I have onboard. One of the things that you have to watch for is unexpected terminators. The standard does not allow the devices to have built in terminators. Obvious reason is if there are more than two devices with built in terminators the bus would be incorrectly terminated. Some manufacturers for good technical reasons put a termination in the connecting cable. If this situation exists there may be more terminators than you are aware of, this can cause unexpected results.


Sent from my iPhone using Cruisers Sailing Forum

[rendern]

Don't know if this is related, but I installed a new Garmin 741 this past weekend and connected my other chartplotter (another older Garmin) and AIS transponder (not Garmin) to it and was playing around with it when I found a menu in the 741 for additional NMEA sentences. I enabled one (don't remember which) and other devices started acting up (i.e. rebooting, screen blank type behavior). I disabled the NMEA sentence that I had enabled and everything returned to normal. Don't know if that could be part of your problem, but I learned a lot just noodling around with it for a day. It can be pretty particular about which sentences are enabled across manufacturers and even with same manufacturers. -Norm

[alctel]

Originally Posted by HighTemp

It's the Simard RS35.

Yup, that's the model that causes all the problems (I have the same one).

Update it to the latest firmware, it fixes it.