Moderated Discussion Areas
ContinuousWave: Small Boat Electrical
NMEA-0183 Interface: Four-pole Genderless Connectors
|Author||Topic: NMEA-0183 Interface: Four-pole Genderless Connectors|
posted 12-21-2011 08:32 PM ET (US)
[This discussion was originally a sidebar to another thread. I have separated it from the other thread and created a new thread for this topic.--jimh]
posted 11-13-2011 04:35 PM ET (US)
[Described in a narrative an idea an NMEA-0183 serial data interface using genderless connectors. After several days of careful review of the idea, I have concluded the method will not work, and I have deleted it. I will be convinced of its value when Chuck implement his method with actual connectors, tests it with some real devices, show us in photographs how he did it, and give us links to where we can purchase the components. --jimh]
posted 11-24-2011 11:19 AM ET (US)
It's a real pain that Jim chose to delete my earlier post rather than
just pointing out the problem with it. Now I have to recreate it from
[Editor's note: Actually it was a benefit for Chuck that I deleted his initial method. It did not work. In recreating the method, Chuck revised the method and turned it into something that does work--jimh]
It wasn't just theoretical. I wired my Montauk in a similar manner,
The female contacts will be NMEA out, the male contacts will be NMEA in.
The top flat-two will be NMEA +, the bottom flat-two will be NMEA - or ground
if there's no NMEA -.
Or, to put it in Jim-speak:
Top Left: A = TX+
Bottom Left: B = TX-. If not available, E = Ground
Top Right: C = RX +
Bottom Right: D = RX-. If not available, E = Ground
posted 11-25-2011 10:27 AM ET (US)
Chuck has proposed (now twice proposed) an wiring scheme for interconnecting NMEA-0183 serial devices as an alternative to the method I described as a Universal NMEA-0183 interface method. I have given Chuck's method careful review and find it lacking. I explain its faults below.
The first concept in Chuck's idea is the use of genderless connectors. At first this seems like an interesting idea: a connector that can mate with itself. There several suitable connectors that have this property. A very good example is the Anderson Power Pole connector. They are genderless connectors and mate with themselves. You can create multi-pole connectors with them. The use of a genderless connector is an interesting notion, but it is not the real basis of Chuck's idea. One could, if they wished, use genderless connectors for my wiring devices.
The actual concept of Chuck's idea that is the essential element of his scheme is that the connectors, when mated to each other, flip the transmit signals to the receive signals, and vice versa. It is this concept which is important. In this way, Chuck's idea eliminates the second device, the back plane connectors and their cross-over wiring wiring. If this were possible, it would be a good idea.
As Chuck has described his proposed connector, it is indeed possible to connect it with a similar connector and have the signals cross. However, there is a dreadful side effect of this property--the connector can be mated to its twin in more than one way. The desired cross-over property is only obtained in one of the possible mating configurations. In the other mating configuration, the cross over connection is not obtain. This destroys the fundamental requirement of the solution—it is not a universal solution. It only works when you are smart enough to know which way to mate the connectors.
In addition to not meeting the fundamental requirement of the solution, Chuck's idea is a kludge. (Kludge is an electronic term for sloppy or inelegant solution, notwithstanding its actual functionality) I judge Chuck's idea to be a kludge on this basis:
--the connectors are only available as two-pole devices, and must be assembled into four-pole devices by taping them together;
--the connectors are only available as pre-assembled, pre-molded devices with wiring already attached, and existing devices must be spliced onto the connector's wires, an ugly and hard to work solution;
--the typical pre-wired connector is provided with conductors of 12-AWG or 14-AWG, which are much larger than needed for the currents involved, and the larger wire size makes it more difficult to splice to the usual 22-AWG conductors from the devices.
In the larger world of electronic serial data communication, the problem of connector selection and wiring for electronic data has been faced by designers for many decades. If there were a good workable solution for interconnecting two serial data devices with a genderless universal connector, I suspect that a solution would have already been invented, but none has. Nor has Chuck done it, here, from what I can tell.
posted 11-25-2011 11:27 AM ET (US)
Jim, this is no more kludgey than yours, and I don't
have to find a dry, accessible, place to mount the circuit boards.
It only mates one way.
There are four ways you can try to put two connectors together.
posted 11-25-2011 11:49 AM ET (US)
Chuck--I think you revised your plan upon its re-introduction. You original proposal permitted mating the connectors in two ways because the genders were staggered diagonally. I will now study your revised plan, and reply after careful consideration.
posted 11-25-2011 12:34 PM ET (US)
To study Chuck's proposed plan, which I will call the four-pole genderless connector plan, I actually implemented it on my bench, at least so far as to make up the connectors as Chuck described. Here is a connector created according to the plan:
I used color-coded conductors to identify the four poles, using Brown, Red, Orange, and Yellow (the resistor color codes for 1,2,3, and 4).
I made a second connector, identical to the first:
Now I mate them. I agree, they only mate in one way with this new configuration:
Referring to the conductors by their color codes, we see the mating provides
BROWN --> ORANGE
RED --> YELLOW
(and, of course, by definition then)
ORANGE --> BROWN
YELLOW --> RED
This is useful and functional interconnection which could be wired for NMEA-0183. However, I think there are several drawbacks to it. I will explain them in a follow up.
posted 11-25-2011 12:57 PM ET (US)
The four-pole genderless connector method (or Chuck's method or perhaps, as Chuck disclosed in an e-mail to me, a method used by IBM back in the 1980's and perhaps patented) is a very useful method to interconnect two serial data devices, such as two NMEA-0183 devices. I appreciate Chuck introducing this idea. I think is has merit and I will follow up on an implementation using different genderless connectors.
In my opinion, the implementation of the method using off-the-shelf SAE two-pole power connectors is difficult. My basis for that is my experience in trying to assemble the connector I show above in my photographs.
Taping two separately molded two-pole assemblies together is not particularly easy. The connectors have small ridges on them which prevent them from laying again each other. You might want to use a razor-knife to cut off the ridges to enhance the assembly.
The typical pre-molded connectors come with attached wires using red and white insulation in short lengths. The wiring color code is flipped at one of the connectors, making these pre-made devices have two schemes, which I will call R and S. In the R scheme the red wire is connected to the male pin, and in the S scheme the red wire is connected to the female socket. When making up a set of two four-pole genderless connectors there will be opportunity for a lot of ambiguity in the wiring to the connector if one were to refer to the original color of the wire insulation of the pre-made connector. (This is why I used color tape in my test implementation. It was very difficult to keep track of connector designation without a unique color identifier.)
The mating force of the connectors is high. Since the home-brew four-pole connector is made with adhesive tape, it is extremely difficult to mate two of these together. (That is why I show them not completely mated in my photograph, above.) I never did try fully engaging the home-made connectors. I suspect that disengaging them would be even more difficult.
My objections remain regarding the splicing of existing cables from marine devices to a home-made four-pole genderless connector made from the SAE two-pole power connectors taped together. The wire sizes are much different. In addition, the ambiguity of the color codes (that I mention above) will also make errors in wiring more likely.
In summary, the four-pole genderless connector method could be useful, but the suggested implementation with two-pole SAE power connectors is not very workable in my opinion.
posted 11-25-2011 01:03 PM ET (US)
The next step in this investigation (for me) is to find a more suitable four-pole genderless connector. I will investigate the Anderson Power Pole connector to see if it is suitable. The Anderson Power Pole connector appeals to me because:
--you assemble it from individual poles into a multi-pole connector
--it is available with color coded housings so there will be no ambiguity about wiring
--it is not a pre-molded connector with captive wiring, so you can assemble it to an existing cable
The drawback of the Anderson Power Pole is that it is not at all waterproof.
If anyone has a suggestion of another four-pole genderless connector, please let me know by joining the discussion.
posted 11-25-2011 01:16 PM ET (US)
Yes, I revised the assembly. I was working from memory 40
miles from the boat.
I'd love to find a less kludgey connector that's small (so
posted 11-26-2011 12:05 AM ET (US)
I assembled a four-pole genderless connector using Anderson Power Pole components into a square two-by-two assembly. It mated with itself just as the home-brew SAE power connector did, that is, there was a cross connection. This verifies that one could use an Anderson Power Pole connector to make a four-pole genderless connector that would be useful for interconnecting NMEA-0183 devices according to a universal wiring scheme. However, I don't think I would recommend this connector because it is really not very appropriate at all for marine use.
posted 12-21-2011 08:47 PM ET (US)
[At this point in the discussion Chuck introduced another method of interconnecting NMEA-0183 devices which did not employ genderless connectors, but instead employed a pair of two-pole connectors of opposite gender. That discussion has been removed from this thread and is now its own topic in a separate thread.]
Purchase our Licensed Version- which adds many more features!
© Infopop Corporation (formerly Madrona Park, Inc.), 1998 - 2000.