NMEA-0183 Interface: GX2150, a Difficult Example

Electrical and electronic topics for small boats
jimh
Posts: 11670
Joined: Fri Oct 09, 2015 12:25 pm
Location: Michigan, Lower Peninsula
Contact:

NMEA-0183 Interface: GX2150, a Difficult Example

Postby jimh » Thu Oct 25, 2018 11:52 am

In this article I will examine the interfacing of a NMEA-0183 device that presents several problems, and explain how they were solved. (I am writing this article primarily for my own reading. This interconnection was so complicated I need to write down how it was done so I don't forget how it was accomplished.)

The device under discussion is a Standard-Horizon VHF Marine Band radio with DSC and AIS features, their model GX2150. This device has NMEA-0183 ports, but the electrical signals are not the recommended balanced signals as provided in NMEA-0183 v.2.0 and later specifications. The signals are single-ended signals referenced to a common ground. The signal names are also not given in proper NMEA form. And the wire colors used are not the recommended wire colors. In terms of conforming to NMEA-0183, this product is a trainwreck in its implementation, nomenclature, and wire color, but it does actually work and interface with other devices--if you can overcome the obstacles the manufacturer has put in your way.

In the GX2150 the NMEA signals are provided on only one signal wire, and the voltage on the wire is referenced to a "NMEA common" wire. Using this method was allowed in the very first version of NMEA-0183, i.e., the version that came out in 1983. But by year 2000 or so when the GX2150 was made, NMEA had moved on and their NMEA-0182 v2.0 specification called for balanced signals. It is possible to make an interface between a balanced signal and an unbalanced signal, but it should not be necessary with modern devices. Further, the electrical interface of the signals is generally compromised when you have to directly connect balanced and unbalanced signals.

Adding to the problem for the user to accomplish the interface is the nomenclature used by Standard-Horizon. They refer to the signals with names of "input" and "output" and assign a polarity designator (+) to those signals. The NMEA recommendation uses simpler names, TALKER and LISTENER, and assigns no polarity to the signals. The two signals for each TALKER or LISTENER pair are just called A and B. The user must accomplish a translation of the nomenclature from non-standard to standard names. In this case the translation looks like this:

NMEA            Standard-Horizon
TALKER A = NMEA OUT (+)
TALKER B = [not provided]
(no equivalent)= NMEA common
LISTENER B = [not provided]
LISTENER A = NMEA IN (+)


To identify the individual wires in a multi-conductor cable that has these signals, the wire insulation color is used. We can then add this to our table:

NMEA                Standard-Horizon  Wire Color
TALKER A = NMEA OUT (+) = GRAY
TALKER B = [not provided] = [none]
(no equivalent) = NMEA common = GREEN
LISTENER B = [not provided] = [none]
LISTENER A = NMEA IN (+) = BLUE


Now we have a table with NMEA standard signal names and wire colors. Let's eliminate the Standard-Horizon name column to keep things simpler:

NMEA                 Wire Color
TALKER A = GRAY
TALKER B = [none]
(no equivalent) = GREEN
LISTENER B = [none]
LISTENER A = BLUE


But now the real fun starts. The cable from this device is only a few inches long. It is much too short to be able to reach another device for interfacing, so it has to be extended. To extend the cable, it would be nice to maintain the wire colors, but where do you get a multi-conductor cable with wire in gray, green, and blue conductors? That's a rhetorical question, but the answer is you cannot find a multi-conductor cable with those wire colors anywhere. When the cable is extended the wire colors will have to change to whatever wire insulation is available in a multi-conductor cable.

In my case, I used a two-pair shield cable to extend the signals. The wire colors are blue and white for one pair, and blue-white and white-blue for the second pair. The cable also has a shield.

I make ALL my interconnections between NMEA-devices using the same wiring convention, a system I created myself and have written about, which allows for a universal interface connection. All devices are wired to five-pole connectors. These five pole connectors plug into a backplane, and the backplane has a consistent wiring pattern. By the artful design of the system, all devices interconnect in this universal manner. (For details, see my article describing the system, "Universal NMEA-0183 Interface," which explains my design and concept in detail.)

Following my universal interconnection, I wire a five pole connector with my two-pair shield cable as follows:

TALKER A = WHITE
TALKER B = BLUE
COMMON = Shield
LISTENER B = BLU/WHI
LISTENER A = WHI/BLU


Now to finally make this work, I have to connect my extension cable with universal connector to the three wires from the Standard-Horizon radio. This is where the real science of the interconnection will occur. I follow my own advice as described in the article on my Universal NMEA-0183 wiring. The most unusual arrangement of the signals: I substitute the "NMEA common" signal from the radio for the missing TALKER B signal. But I cannot wire "NMEA common" as LISTENER B because that would short circuit another device's TALKER B at the Universal interface connection. (See the linked article above for more explanation.) Now I need another table to get those connections straight:

RADIO     INTERFACE
GRAY --> WHITE
GREEN --> BLUE
NC --> Shield
NC --> BLU/WHI
BLUE --> WHI/BLU


Are you totally confused? You should be. Who could ever work through all of these crazy obstacles to interconnecting two devices? Connecting two devices should NEVER be this hard. I have worked with interconnecting serial data signals for decades, and I can barely figure this out. How can NMEA, Standard-Horizon, and the Coast Guard expect the average boater to accomplish this sort of ridiculously difficult interconnect on his own? It is just not going to happen, as the Coast Guard statistics demonstrate: 90-percent of DSC radio calls don't have any position data in them.

Here is a final table, beginning at the radio with its names, translating to NMEA names, then noting the wire color on the radio, then the interconnection to my cable, and finally the connections at my universal interface:

NMEA OUT (+) --> TALKER A   --> GRAY  --> WHITE   --> TALKER A   --> (A) terminal
NMEA common --> TALKER B --> GREEN --> BLUE --> TALKER B --> (B) terminal
SHIELD --> COMMON --> (C) terminal
BLU/WHI --> LISTENER B --> (D) terminal
NMEA IN (+) --> LISTENER A --> BLUE --> WHI/BLU --> LISTENER A --> (E) terminal


And there is one final and further complication. At some point in the production of the GX2150 radio, Standard-Horizon changed the NMEA signals to be balanced signals. If you happened to have one of these later-version radios, then there will be an entirely different interconnection wiring, and most of the above will be incorrect. If you don't have the correct version of the owner's manual, you will not see the proper documentation of these circuits.

jimh
Posts: 11670
Joined: Fri Oct 09, 2015 12:25 pm
Location: Michigan, Lower Peninsula
Contact:

Re: NMEA-0183 Interface: An Example

Postby jimh » Thu Oct 25, 2018 12:32 pm

If manufacturers followed the NMEA guidelines, that is, they provided balanced signals, gave them the proper names, and provided them on the recommended wire colors, then interfacing two devices would be vastly simpler:

Image
Fig. 1. The interconnection of two devices that follow NMEA guidelines.

If the manufacturers provided the signals on a multi-conductor cable that was at least 24-inches long, then the two devices could be about three apart and still be easily interconnected with their original cables, and no extension of the wiring would be necessary.

Note that the four colors used for NMEA are WHITE, BROWN, YELLOW, and GREEN. I don't know of any source of a four-conductor cable with those wire colors. Extending the conductors would probably mean using some sort of existing four-conductor cable and the colors available in that cable. The most common four-conductor cables I can find generally use a color scheme or RED, BLACK, GREEN, and WHITE. For example, BELDEN 9534 cable, available in 100-foot spools for $80. This is a 24-AWG shielded cable, but the conductors are not twisted pairs.

fno
Posts: 170
Joined: Thu Oct 15, 2015 10:01 pm

Re: NMEA-0183 Interface: GX2150, a Difficult Example

Postby fno » Fri Apr 26, 2019 8:56 am

JimH, four conductor shielded cable with white, brown, yellow, and green conductors is quite common in industrial CanBus communication networks. The outer insulation is usually some form of purple. In my humble opinion the talker/ listener nomenclature is crap. I can send you some if you'd like it for future projects, but hopefully NMEA2000 has your interest already.

jimh
Posts: 11670
Joined: Fri Oct 09, 2015 12:25 pm
Location: Michigan, Lower Peninsula
Contact:

Re: NMEA-0183 Interface: GX2150, a Difficult Example

Postby jimh » Sat Dec 14, 2019 2:03 pm

FNO (Frank)--many thanks for promptly sending the cable as you promised. I promptly set it aside for a few months to go boating. I just got around to writing more about that cable. Please see the new topic on off-the-shelf cable useful for NMEA-0183 wiring:

NMEA-0183: Two-pair shielded cable with proper wire insulation colors
http://continuouswave.com/forum/viewtopic.php?f=9&t=5123