NMEA-0183: Two-pair shielded cable with proper wire insulation colors

Electrical and electronic topics for small boats
jimh
Posts: 11672
Joined: Fri Oct 09, 2015 12:25 pm
Location: Michigan, Lower Peninsula
Contact:

NMEA-0183: Two-pair shielded cable with proper wire insulation colors

Postby jimh » Fri Dec 13, 2019 1:39 pm

The recommended method of wiring an interconnection between two NMEA-0183 devices that use the recommended balanced ports and the recommended wire insulation colors and the recommended signal names is shown in Figure 1.

Image
Fig. 1. Schematic diagram of NMEA-0183 interconnection

The recommended wire insulation colors are now starting to be used by manufacturers of VHF Marine Band radios and recreation-grade electronic chart plotters in their provision of a NMEA-0183 interface on loose conductors that are part of a multi-conductor cable. Typically the length of the cable with the NMEA-0183 signals is not very long, perhaps only one-foot, and if two devices have to be interconnected, their integral NMEA-0183 cables may not be long enough to reach each other. In that situation it would be useful to extend the NMEA-0183 cable from one of the devices to reach the other device. The problem with that solution is finding suitable two-pair shielded cable with the corret wire insulation colors.

The topic of finding an off-the-shelf cable that complies with the NMEA-0183 recommendation for wire insulation colors to identify the signals in a four-conductor cable has been mentioned before in several threads discussing NMEA-0183 interfaces, but this thread will concentrate on the wire insulation color, how it came to be chosen, and where to find suitable wire.

As mentioned in one of the other threads, long-time forum participant FNO sent me about a 12-foot length of (very nearly) appropriate cable that he uses in his professional work in industrial control networks. The cable is made by a German wire manufacturer. The sample I received was marked

    LAPP KABEL STUTTGART UNITRONIC® BUS LD 2 x 2 x 0.22 ROHS ART. 2170204

I interpret this as a cable made by a firm LAPP KABEL of Stuttgart, Germany, with the registered product name Unitronic, and having two pairs of two-conductors with wire cross section area 0.22 mm2, compliant with the Reduction of Hazardous Substances regulations, company part number or article number 2170204. And many thanks to Frank for sending this to me. I am very appreciative of his kindness.

On the manufacturer's website I found the cable specifications:

https://products.lappgroup.com/fileadmi ... 0203EN.pdf

The specifications do not mention the wire insulation colors explicitly, but instead refer to "Core identification code acc. to DIN 47100."

Searching for DIN 47100, I found that DIN 47100 (Deutsches Institut für Normung or in English “German Institute for Standardization” number 47100) is a wire insulation color code for identification of conductors in the telecommunications industry in use prior to November 1998, but afterward said to have been withdrawn without a replacement. (Cf. : https://en.wikipedia.org/wiki/DIN_47100 )

The standard identifies wire insulation colors to be used in sequence in multi-conductor cables. The first four conductors are to use these colors:
    First conductor = white
    Second conductor = brown
    Third conductor = green
    Fourth conductor = yellow

These are, indeed, the wire insulation colors specified by NMEA.

Regarding the size of the conductors, in this particular cable they are specified by their cross-section area in millimeters. A cross section of 0.22-mm2 is equivalent to AWG-24. The outer diameter of the cable including the outer insulating jacket is slightly greater than 7-mm or 0.28-inches.

LAPPKABEL2170204.jpg
Fig. 2. A cable made by LAPP KABEL in their Unitronic line of communication or data cables. This cable has the proper NMEA-0183 wire insulation colors. I included a penny to give a sense of the size of the cable.
LAPPKABEL2170204.jpg (57.75 KiB) Viewed 7143 times


In its literature, LAPP KABEL describes the typical use of this cable:

    UNITRONIC® BUS LD
    Flexible bus cable with PVC outer sheath, for use in different bus systems
    For bus systems like e.g. Modbus, SUCOnet, Modulink, VariNet.

There is no mention of use in NMEA-0183, which is NOT a bus system but rather a serial point-to-point interface system.

Based on this cable, the first actual sample I have seen of a shielded two-pair twisted-pair cable with proper wire insulation colors that almost conforms to the NMEA-0183 recommendation--in the recommendation the pairs are supposed to be individually shielded--it appears that NMEA has chosen a standard for their interconnecting cables that is in reality and practicality IMPOSSIBLE to comply with. It is thus little wonder than I have never been able to find any off-the-shelf cable that can be easily purchased in small lengths at reasonable prices which complies with NMEA-0183. It is also further confusing that NMEA--primarily an American marine standards organization--instead of using common Electronics Industry of America (EIA) wire insulation color codes chose to use a German wire color standard and a standard which has not been in effect for about 20-years.

ANOTHER "KABEL"

Browsing the LAPP KABEL catalogue, I found another cable which is similar to their article 2170204 cable: article 2170261. (See the LAPP literature for details.) The product description:

LAPP KABEL wrote:UNITRONIC BUS CAN A is a data cable with UL and cUL approval, for CAN (Controller Area Network) fieldbus system according to ISO11898 as well as for high performance data networks with 120 Ohms nominal impedance. The second pair can be used for electrical power supply for the logical bus units.


Pair 1 is White/Brown and Pair 2 is Yellow/Green.

This is again interesting because the cable is for a bus network, but NMEA-0183 seems to have adopted these color schemes for a point-to-point wired serial data connection. This cable would be appropriate for NMEA-2000, a bus system with a single pair data bus and a second pair for network interface power. But for NMEA-2000, NMEA uses a different color coding, White/Blue (for data) and Red/Black (for power).

It is often mentioned that NMEA-2000 uses CANBus, but from what I can tell (from some research on-line) the ISO 11898 standard does NOT get around to such mundane topics are wire insulation color. There is an SAE (Society of Automotive Engineers) standard called J1919-11, "Physical Layer, 250 Kbps, Twisted Shielded Pair" that might specify wire color, but at $81 to get a copy, I won't be able to tell you what it says. By inference, I suspect it calls for the data pair to be Green/Yellow, based on a wire vendor advertising a single shielded pair cable with those conductors.

jimh
Posts: 11672
Joined: Fri Oct 09, 2015 12:25 pm
Location: Michigan, Lower Peninsula
Contact:

Re: Analogies for determining pair H/L from wire insulation color

Postby jimh » Fri Dec 13, 2019 4:11 pm

With regard to identifying signals in the data pairs, the "Data H" and "Data L" signals may be inferred from the wire insulation color by using an analogy to earth-and-sky or hot-and-cold.

For earth-and-sky the color analogies are
    DATA H / DATA L
    Yellow (sun) / Green (grass)
    White (clouds) / Blue (ocean)
    White (clouds) / Green (grass)

For hot-and-cold the color analogies are
    DATA H / DATA L
    White (white-hot) / Blue (ice-cold)
    Red (red-hot) / Black (stone-cold)

jimh
Posts: 11672
Joined: Fri Oct 09, 2015 12:25 pm
Location: Michigan, Lower Peninsula
Contact:

Re: NMEA-0183: Two-pair shielded cable with proper wire insulation colors

Postby jimh » Sat Dec 14, 2019 2:17 pm

Inasmuch as there does not appear to be any convenient, quick, and inexpensive source for getting a two-pair cable with the NMEA wire insulation colors, I have turned to a simpler, faster, and less expensive solution.

A close reading of the NMEA-0183 standard—at least the portions of it that I can find published by NMEA that are not encumbered with non-disclosure—reveals advice on how to extend a cable between two devices if their integral cables are not long enough to reach from one device to the other. NMEA recommends that only the TALKER pair be extended, not the LISTENER pair. This suggests that if two devices were to be interconnected bi-directionally, then the TALKER pair from each device would be extended to the other device, and this would be done separately with two single-pair shielded cables. The wire insulation colors needed for a TALKER are WHITE/BROWN.

NMEA0183NewOnlyTalkerExtended.png
Fig. 3. NMEA-0183 interconnection with only the TALKER pairs extended. This requires only finding a single-pair cable with White and Brown conductors.
NMEA0183NewOnlyTalkerExtended.png (26.23 KiB) Viewed 7094 times


Unfortunately, cannot I find any shielded single-twisted-pair cable with WHITE/BROWN, I can find cables with WHITE/BLACK or WHITE/BLUE insulation colors. With such a cable the wiring could almost conform to the standard, and if the BLACK or BLUE conductor were re-marked at each end of the cable as BROWN (by using, for example, some brown heat shrink tubing), then the cable could be modified to conform with the recommended wire colors.

Finding a shielded twisted pair cable with wire insulation colors of BLACK and WHITE and available in small quantity is also not very simple. I found that MOUSER.COM sells a 24-AWG shielded twisted pair cable for about $0.44/foot in small lengths. More details at

https://www.mouser.com/ProductDetail/65 ... 21-24-0-99

The only drawback to using two separate cables to extend the NMEA-0183 interface TALKER signals is you need twice as much cable. If the two devices are ten feet apart, 20-feet of single-pair cable will be needed.

There are other manufacturers of similar cable, but generally those cables are only sold in a minimum of 100-foot spools. Several types have insulation rated for use in plenum space and are very expensive.

To mark the black insulation conductor to BROWN in a simple manner, brown nail polish can be used. Inexpensive brown nail polish is about $5, and one bottle could probably mark hundreds of conductors. Brown heat shrink tubing is another option, but it may be harder to acquire and more expensive.

This approach seems to be the easiest way to conform to the NMEA wire color insulation coding for NMEA-0183 wiring.

Tom Hemphill
Posts: 80
Joined: Fri Oct 16, 2015 5:04 pm

Re: NMEA-0183: Two-pair shielded cable with proper wire insulation colors

Postby Tom Hemphill » Sun Dec 15, 2019 11:13 am

Thank you for spelling out and explaining those German acronyms and manufacturing standards. It will be helpful to me as I just bought a Volkswagen.

I'm surprised that shielded cable is recommended. Is it to protect the NMEA-0183 signal from outside interference, or to prevent it from causing interference?

jimh
Posts: 11672
Joined: Fri Oct 09, 2015 12:25 pm
Location: Michigan, Lower Peninsula
Contact:

Re: NMEA-0183: Two-pair shielded cable with proper wire insulation colors

Postby jimh » Sun Dec 15, 2019 1:54 pm

Tom--thanks for reading my article. I know it is a bit out of the mainstream.

Tom Hemphill wrote:Is [use of shielded cable] to protect the NMEA-0183 signal from outside interference, or to prevent it from causing interference?


A shielded cable performs both functions: it keeps the internal signal from leaking out and keeps external signals from leaking in.

In NMEA-0183 the signals are logical-opposites of each other, so they are more like an alternating-current waveform; when one wire is in the positive voltage signal state the other is in the negative voltage signal state. They are (or should be) balanced to ground, and there is no net DC offset in the signals themselves, unless the data creates a drift. (But most encoding protocols try to remedy that.) The cable is a twisted pair, which tends to be self-cancelling in terms of emitting a signal from the cable, and also rejects any common-mode signal interference from external signals.

The argument about the need for shielded twisted pairs was more or less settled by the Bell System in the 1920's. The Bell System has about a billion twisted pair circuits, some of them running in multi-pair cables of 100-pairs which are then twisted into 400-pair cables, and no shielding. The crosstalk between adjacent pairs is very low, probably better than -50 dB isolation. As far as ingress, it is common for unshielded twisted pair telephone cables to run just a few feet away from 120-VAC power lines for miles and miles. There is astonishingly little ingress of the 60-Hz power into the telephone twisted pairs.

Of course, the Bell System signal bandwidth was a lot slower than typical modern digital signals, but that was because the signals might go 15-miles on a twisted pair. That was when the signal on the wire was still at baseband. Later, with digital signals, the Bell System was able to re-use those twisted pairs to delivery 1.5-Megabits/Sec signals, albeit limited to a distance of about a mile. But, again, sending 1.5-MBits/Second on unshielded twisted pairs is routinely done in (what is today the companies created by the break-up of) the Bell System.

Is the shielding absolutely necessary for NMEA-0183 signal at 4800-bits/second that goes a few feet? The specification calls for shielding:

NMEA-0183 Standard wrote:3.1 Interconnecting Wire
Interconnection between devices may be by means of a two-conductor, shielded, twisted-pair wire.


But shielded cable is probably not really necessary, and especially in cables that run only 1-foot to an interface with another device, with the signals being balanced signals, all active gear powered by a common 12-Volt power source, and common ground bus connection within a few inches.

A really good reference for information about NMEA-0183 is not available from the NMEA organization without spending hundreds of dollars, but a vendor that makes NMEA-0183 devices, ACTISENSE, has an excellent document that explains the protocol and interfacing in detail:

ACTISENSE: The NMEA 0183 Information sheet
https://www.actisense.com/wp-content/uploads/2017/07/NMEA-0183-Information-sheet-issue-4-1-1.pdf

In the quest to find cable that complies with the NMEA recommendation, the specification for shielded cable is really nothing hard to comply with. There are many two-pair twisted-pair cables, but usually under a common shield. The element of the NMEA specification that makes no sense to me and causes the biggest problem in compliance is the wire insulation color code.

Exactly why NMEA chose to use wire insulation as specified in an obsolete DIN (German) standard is unknown to me as well as being hard to figure out. As I mentioned, I do not believe that in the CANBus standard there is any specification for wire insulation colors for signal identification. That is left to the user to implement. Why NMEA picked such impossible-to-buy wire colors makes no sense to me.

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

Re: NMEA-0183: Two-pair shielded cable with proper wire insulation colors

Postby fno » Sun Dec 15, 2019 7:07 pm

To all of those interested in obtaining lengths of the Lapp CanBus cable that has white/brown/yellow/green/shield conductors simply drop me an email and I will send you a piece. I do not have a roll of the stuff but come across many feet of this stuff in my workplace. For those that want 100' or so you can find it at Allied Electronics but they may have to order it.



Update: I looked up the Lapp cable# and found that the 217204 cable that I sent to JimH is now discontinued at Allied Electronics. There is an alternative cable that is rated for continuous flex(wave) and has the same characteristics as the former cable. The new cable# is: 2170214

[See this link for Allied Electronics Inc. sources--jimh]