Troubleshooting NMEA 2000 and the troubles with NMEA 200

[Suijin]

Well I did something else. The network faulted and refused to transmit data despite cycling it half a dozen times; I removed a cluster of T’/drops (plotter, GPS, rudder); the network booted fine; put the T’s back in and network booted fine and has been fine for a couple of hours.

This is part of its intermittent behavior. The connections were tight before taking them apart, so nothing changed. But we’ll see.

[transmitterdan]

I suspect the power draw from the devices exceeds the ratings. Have you done the calculations of all loads to verify that the system is not power starved? This is quite common.

[Paul L]

Are you confident that you have both ends of the backbone terminated properly?
What brand of mfd display do you have? Some have a nmea 2000 diagnostic page that shows the number and type of transmission errors while running.

[CarinaPDX]

It seems odd to me that you are powering 2 mfds off the network. You need to start getting quantitative about this: add up all of the loads on the network and see if they are within spec for the micro- or mini- cable that you are using for the backbone. Is the power being inserted in the middle with a power tap of the right rating? Of course this should have been done before installation; if so, then try some measurements.

I would use a spare t-tap and check voltage at various locations in the network. N2K cables have 5 conductors: 12VDC power (+/-), data (+/-) and shield. The terminating (pull-up) resistors at each end connect the 12VDC+ to the data+, while the various devices pull the data+ down to signal. The 12VDC+ line at the resistors would be a good place to check for power adequacy, as it would include the voltage drop through the length of the backbone. It would be more difficult to test the signal line, probably with a scope, so let's hope that is never necessary.

If we are just guessing, I would guess that the problem is a power one, either due to too much load, inadequate source or cabling, or a contact that is too resistive. The thermal sensitivity hints at a resistive (dirty, loose, or corroded) contact but if so, it may just be the last straw in a marginal setup.

Good luck,

Greg

[Reefmagnet]

Originally Posted by Suijin

OK I'm now suspecting the masthead connection. The stick was down for a year with the simnet cable dangling exposed. I did not clean the female sockets before reconnecting in the rush to get everything ready to step the mast.

Would result in bad termination, which i think would explain everything getting power on the network but the plotter and mfd's not seeing any data from other sensors/devices. Yes?

Up the mast I go. Any advice on cleaning the sockets for the pins? They be tiny....

And thanks everyone for the input and advice. I appreciate it very much.

I have a mixed network comprised mostly of Simnet (which I think is just about all but obsolete now) and NMEA 2000 (and NMEA 0183, too!). I vaguely recall that although Simnet allows daisy chaining of devices (akin to having multiple devices attached to a single NMEA 2000 drop) it only works for some devices and some care is required in connecting everything to ensure that the same "backbone and dropper" structure of a standard NMEA 2000 network is maintained.



For cleaning the Simnet connector at the masthead, I'd suggest not using any abrasives, but just giving the connection a good flush out with contact cleaner. Apply silicon grease to the plug and socket before reconnecting to protect against moisture and corrosion.

[Sardean]

Originally Posted by Reefmagnet

I have a mixed network comprised mostly of Simnet (which I think is just about all but obsolete now) and NMEA 2000 (and NMEA 0183, too!). I vaguely recall that although Simnet allows daisy chaining of devices (akin to having multiple devices attached to a single NMEA 2000 drop) it only works for some devices and some care is required in connecting everything to ensure that the same "backbone and dropper" structure of a standard NMEA 2000 network is maintained.



For cleaning the Simnet connector at the masthead, I'd suggest not using any abrasives, but just giving the connection a good flush out with contact cleaner. Apply silicon grease to the plug and socket before reconnecting to protect against moisture and corrosion.

I agree with ruling out power problems first, is it possible that a fridge, or water pump cycles on and momentarily drops the voltage. Check common negative connections for corrosion or loose wires. Intall a 5 to 10 watt globe on the same fuse as the instruments and see if it fluctuates or flickers

[smac999]

get a heat gun or hair dryer and point at stuff. see if you can figure out where the issue is. or narrow it down.

I have never seen a heat issue.

[macman4190]

Originally Posted by transmitterdan

Can you temporarily disconnect the mast cable and replace that with a termination? Then run the network for a while to be sure the problem really is up the mast.

I lost my network several times due to issues with the mast connection and had to remove it and plug in a terminator.

[Quietta]

Originally Posted by Suijin

...and then after another couple of hours motoring at the end the network again went down, of course just when i was entering port with shallows...and no depth sounder. Has anyone heard of or experienced heat-related issues with NMEA 2000 cabling/junctions?

Just on another angle, have you considered it might be voltage related?? After a couple of hours of motoring, your batteries will be fully charged and be around 13.8 Volts, possibly higher. Is one of your instruments reacting to the higher voltage and not working, possibly overheating internally? Can you switch the instruments onto the house bank so that the engine is not chargeing this bank and see if the problem still occurs?

[tanglewood]

Chances are about 5% that it's a cabling problem, and 95% that it's a device problem. You have ruled out heat already, which is good.


What are the exact devices on the network, and in what topology? Can you post a sketch showing it all, including approximate cable lengths?


Since the problem reproduces when not underway, I think you just need to bite the bullet and start segmenting parts out of the network and see when the problem stops occurring.


Another test would be to see if the AP can see any other devices when the failure occurs. Most have pages where you can see what's on the network, and most will alarm if they lose things like rudder, compass, etc. It may be that the MFDs are losing communications, but the devices are still working fine and the AP can see them. The point it that just because the MFDs can't see any of the devices doesn't mean that all those devices are inoperative.

[Jerry Woodward]

I think another potential weak point in the N2K network are field-installable connectors. When I installed my network, I used 5 field installable connectors from Maretron. They were really tricky to attach the cable and inspite of being really careful, I still managed to screw it up. My network did not work right. Rather than trouble shoot, I just bought new cable with connectors already installed and everything worked fine. I don't know if you have any field installable connectors, but if you do, I would take a close look at those.

[transmitterdan]

Jerry,

That is a good point. NMEA 2K is a CAN bus under the covers. CAN is pretty reliable when the cables are good and all devices behave on the network. A proper device will not hose up the rest of the network no matter what goes wrong. A device that stops the network is not designed right.

But a bad cable causes lots of problems. Devices operate assuming the cable is intact and not intermittent or shorted. CAN does not handle flaky cables well. Few networks do.

[Cloroxbottle]

Perhaps this will help, I'm currently trying to troubleshoot a similar problem.

NMEA-2000: Method for Locating Intermittent Network Wiring Applicances - Moderated Discussion Areas

Here are some other resources...

http://www.cnerd.com/interesting/boa...ndamentals.pdf

http://www.nmea.org/Assets/2012%20ib...stallation.pdf

https://www.maretron.com/products/pd...on%20Guide.pdf

[geekoman]

I have a pretty extensive Simnet setup on our L450, and just went through a learning curve on troubleshooting, probably due to a near miss on a lighting strike that brought the network to its knees.

The Simnet bus is pretty straight forward; terminator, devices, power, terminator. There is usually one terminator in the MHU due to the cable length. The other is at the end of the chain.

I found that having a couple of terminators available is handy for troubleshooting because it allows you to narrow down your network. I started out small and grew. For instance terminator+ MFD + power + terminator. If that works, add something else until it fails. By works I mean shows no errors as described below.

The terminator built into the mast head unit is a challenge. So buying an extra terminator is a good idea. For two reasons, debugging as noted above, and getting your system back up when you are out and about and your MHU fails. The second terminator enables you to disconnect the failed MHU and at least you get the rest of your system back.

Getting the layout of your simnet is very handy. If you call your product's technical support, they might send it to you. B&G received one from Lagoon for my boat model, and provided it or me, I attached it. (Note the terminators in the lower left and upper right (MHU) of the schematic). Otherwise, build one. They are not complicated - the biggest challenge is following the wires in your boat. But, it starts at one terminator, and follows a single path of devices, a power point and to the terminator at the other end. Daisy chained devices are just part of the path, not tee'd off.

Checking voltage is a challenge since you need exposed wires. I found them in my Autopilot Computer, where the simnet was connected to a terminal block.

The terminators are 120 ohms across the data lines. Two terminators on the bus are in parallel and result in a 60 ohm resistance on the data lines for a properly terminated bus. This is measured with the power off. If yours read 60 ohms, that is a good sign. You can find out which pins are voltage and which are data on the internet. Connector pins are in order, power, power, ground, data, data. Looking at the end of a cable connector, with pin number 3 at 12 o'clock, the data pins are at 4 and 5 o'clock. 7 and 8 o'clock are power pins.

You can check your MHU by disconnecting the connector at deck level (i.e. not have to climb the mast) and check the data resistance. If it is not 120 ohms (since it is just a single terminator), you found a problem. Seeing 120 does not say it works, just not seeing it say it does not work.

Finally, you might be able to check your bus properties via you MFD. On my B&G (Navico supplied devices), I am able to select Network, and then Diagnostics. It shows me a count of receive and transmit packets, and the count of errors for receives and transmits. The B&G guys said the Rx and Tx errors should be zero. They have big packet errors and other errors, but they are not as important. Minimize your system, check to make sure Rx/Tx errors are zero, then add to it. No Rx/Tx errors is the rule. In my case even a single MFD (a Triton 2) gave errors. I disconnected it, added another MFD and moved on.

The B&G technical support guys were GREAT! Use the option to let the system call you back, and keep your phone with you. My wait times were 2 hours, and if you miss the call back, another 2 hours. Nonetheless they were very helpful.

BTW, this knowledge was gained over a week of non-stop troubleshooting my network since we were trying to get out of the hurricane box. Hopefully yours will go faster.

[Suijin]

Ok this is still a problem. Here's what I have on the network. All B&G/Simrad/Navico, from end to end:

Masthead anemometer (network powered)
Speed/depth transducer (network powered)
Electronic compass (network powered)
VHF w/AIS receiver
Power injection T, with fuse
AP computer
GoFree wifi module
Radar junction
GPS antenna (network powered)
Rudder feedback (network powered)
MFD chart plotter
2 Triton instrument displays, daisy chained with termination (network powered)

I have two field plugs in the install but they were carefully done with shield wire heat shrinked and I also tried two different cables and the system still faulted.

Yesterday after @ 4 hours of motoring he network went down. MFD cannot see anything and the Tritons cannot see anything. Cycled the network a few times with no joy. Then turned the MFD off and the network came up. Did this a few times and worked, so figured the MFD (a remanufactured unit) was bad. Left it off for an hour or two, network and MFD come up fine. Then the network faults after 30 minutes and even leaving the MFD off, it won't come back up.

It always seems to be fine for a few hours when everything is cold and then it starts faulting, which lead me to believe it was a heat issue. But maybe it's a voltage issue, when the batteries are charged up after hours of motoring. I'm flummoxed.

I was going to make the run from Cape May to Block Island but that obviously is not going to happen with no AIS or radar or AP. Can't risk it.

Oh, and my multimeter just went south so I don't have that to use either :/