N2K Gremlins

[sailingharry]

Our new-to-us boat (coming up on a year) has a fairly robust instrumentation system. The chart plotter is a Furuno NavNet, with position information provided by an N2K GPS unit. Other instrumentation is Ray ST60, which eventually is turned to 0183 and comes into the chartplotter. There is also a Nauticast AIS. Presumably, all of this worked at one point.



I have some experience with N2K, having designed and installed an N2K network on my last boat. But I'm hardly an expert.


Well, I recently realized that the Furuno chartplotter wasn't getting the GPS signal from the N2K unit. It was getting it as a data element from the AIS. Hm... Interesting.


So, some digging.


The N2K network is stupid simple. 3 Tees in a row, power comes in one end, 2 resisters on the other end, the GPS and the chartplotter in the middle. Ultimate bare-bones N2K network. But the chartplotter doesn't see the GPS.


I have an old B&G Triton around, so I made up a patch cable to go investigate. Here's the weirdness.


* Remove the chartplotter, plug in the Triton, GPS not visible.
* Remove the GPS, plug in the Triton, chartplotter not visible.
? How can both be bad?
* On a lark, removed one resistor (theoretically a bad network now), plug in the Triton (so I have 3 devices on the network), and BOOM, the Triton can see the GPS and Chartplotter, the Chartplotter can see the Triton and GPS, fix is found, all is good.
* Remove the Triton, fix is lost (and chartplotter doesn't recognize the network). Reinstall Triton, fix is found. Remove Triton and reinstall resistor, still no fix.


How can removing a resistor and adding a Triton instrument display fix the network? Or perhaps I should say, how come no 2 devices make a network, but all three do? And no, I don't want to leave the Triton rolling around in the lazaret as a solution!

[tanglewood]

Maybe you have termination enabled in one or more of your devices, causing you to be over-terminated in some configurations? Pull off your external terminators, then measure the resistance across the two data lines on the bus. It should be more or less open circuit. If you get something around 60 ohms, then something, somewhere is still trying to terminate. Or go device by device and measure, looking for a rogue 60 ohm terminator.

[sailingharry]

Tanglewood,


Thanks for the thought. I can do that easily, and will try it tomorrow. Two reasons to doubt it:
* The system was "professionally" installed, so presumably worked at one time.
* I pulled one resistor to patch in my B&G (so down to only one), and it worked with the B&G plugged in -- but stopped as soon as I unplugged the B&G (still with only one resistor).


But discounting "thoughts" is a good way to never find the problem! I'll give it a go tomorrow night and report back.


Harry

[SVHarmonie]

Originally Posted by sailingharry

The N2K network is stupid simple. 3 Tees in a row, power comes in one end, 2 resisters on the other end, the GPS and the chartplotter in the middle. Ultimate bare-bones N2K network. But the chartplotter doesn't see the GPS.

This description doesn’t make sense... Ideally, the 12 Volt supply feed should be in the middle of the network, although in your simple case that doesn’t matter much. BUT.... You say there is 12 volt at “one end” and “2 resisters on the other end.”

You need a resistor on EACH end. The 12 volt feed is surely not designed as a terminator.

[sailingharry]

Yes, the resistors are normally at each end. In this instance, the physical network is about 5" long. Or, it could be looked at as 2.5" long, with the drops being the rest of it. It is still solidly compliant with the n2k specs.

But let's assume it is wrong. Why then does it solidly work if I remove a resistor and insert a third drop? And it presumably worked before.

So far, the power drop hasn't been cooperative about unscrewing. But I could swap it and the resistor. I'll give it another try when I'm at the boat tomorrow.

[tanglewood]

I think a few quick measurements with a meter will confirm or deny a handful of possible issues, and let you move forward with them eliminated.


Perhaps also meter power, and try juggling cables at the same time? It's not common, but I have encountered a number of cases where cables were bad, or flaky.

[Dsanduril]

NMEA 2000 is built around the DeviceNet standard, in my day job I find this DeviceNet troubleshooting guide useful. It references 24V systems (industrial controls) but most of the resistance values hold true, as do the common mode voltage test even on 12V systems.

Don't know if it will help your situation but gives a good, general guide for targeted testing with a DMM.

[Bycrick]

The old Simnet power cables came in two versions: one included a built-in terminator, the other did not. They were coded with red or yellow ends, but I can’t remember now which was which. IIRC some N2K cables supplied with, or sometimes the masthead wind indicator itself, had a terminating resistor built in.

[sailingharry]

So I've done some more digging.


* The ground and shield are connected (as they should be).
* The data wires read 0 ohms when all devices are connected (and off), as they should.
* The data wires read 120 ohms when one resister is installed.
* The data wires read 60 ohms when both resistors are installed.
* Suggestions on this forum and another both pointed to my resisters being bunched at one end of the network. On this "network," it's a red herring, but I reconfigured to put a resister at each end.


Still at a loss.


Here's a picture of my entire "network." From left to right
* Terminating resistor
* GPS device
* Chart plotter
* Terminating resistor
* Power injection





Remember, if I remove a resistor (so only one), it still doesn't work. But if I then plug in a B&G Triton into the slot the resistor was in, everything works.

[SVHarmonie]

Try this:

Put the resister (closest to the power point) where the power injection point is. Put the chart plotter where the resister is, and the power inject where the plotter is.

Just to make things “look” like they should.

There are only two possibilities here. You have a device that is acting badly, or a bad connection.

[sailingharry]

SVHarmonie,


Remember, if I remove a resistor and install a B&G Triton, it all works. That kind of rules out tempermental devices, or bad connections, or resister on the wrong side of the power wire (the N2K and Devicenet standards allow the power injection at the end).


The power wire screw doesn't want to move, and I don't want to force it.

[Tehani]

I think the problem is that the FURUNO chartplotter and the GPS receiver are not identifying each other on the network.
It is a bit difficult to explain, but when a NMEA2000 network is turned on, the first phase of linking the devices is to send an "ISO address claim", through this process the logical addresses of the devices sent to the network are resolved.
Your GPS antenna is likely not complying with NMEA2000 certification in that regard.
You can also test whether you physically plug the antenna into NMEA2000 with the PLOTTER already running, and vice versa. Always without connecting the Triton. In this way we will be able to appreciate if the power-on time makes them not identify.
I think it would help us a lot if you said which GPS N2K unit you are using.

[tanglewood]

I think at this point you need to get an analyzer of some sort on the bus to see what's happening. Maretron's N2KView or Actisense NMEAReader would both do the job.

[sailingharry]

I have finally figured this out, and wanted to share.


Cutting to the last chapter, the Garmin power cord had an internal short between the V+ (Pin 2) and SIG-L (center pin). This had the effect of eliminating all data on the bus. Since the SIG-L pin is superfluous on a power cord, I took a pair of needle nose pliers and removed it. Problem solved.


For those who want to know more:


* First, I think the odd behavior I had was not true "behavior" but rather, erratic shorting.
* I replaced the existing "Backbone" (4 tees hooked together with a resistor on each end) with a 6-port multi-tee, in the hopes that the stack of tees had a problem. No such luck. However, this did give me an extra port for further troubleshooting.
* I hooked my B&G Triton on to that spare port, no luck (and no signals). The "Diagnostics" page showed a stack of "TX Errors."
* I removed the two devices on the boat (the chart plotter and the GPS antenna), same situation.
* I added a second B&G Triton, same.
* I took the tee, the patch cord, and the Triton's home.

* I slit the side of the patch cord cable (it was just a home made cable, no big deal) and injected power with a battery charger. SUCCESS! Each Triton could see the other Triton, and no more TX Errors.
* I took it back to the boat, installed it, failure. HMMM, all I did was change from the "side injection" power to the installed power cord.
* I removed the power cord and ohm-ed it out. Found a short between the two pins.
* Hooked up my "working network" (two B&G's and the battery charger) and as soon as I plugged in the power cord, errors.
* The Garmin power cord is a 3-wire cord (V+, V-, Shield) with a molded on 5-pin plug. The data pins are there (it's just a standard plug) but not connected to anything. To have a short would require some sort of internal failure. Since the center pin serves no function in a power cord, I simply removed it. The entire rest of the connection is clean and undamaged, so this did not impact function or water resistance.


Wow. What a journey!