Remote switching of NMEA2000 network power for improved performance.

In this project, I'll describe a typical scenario, the problem, and how to correct it. My boat has a flybridge, and a NMEA2000 network that runs from point C to D as shown by the Blue wire. Heavy-duty wiring to supply DC power runs from point A to B as shown by the Red wire, to power the helm electronics, including multi-function display, radar, and other instrumentation. And finally, the Yellow wire shows how the NMEA2000 network is powered at it's center through a switch located at the helm.

All of that wiring can cause trouble in the form of excessive voltage drop. For the NMEA2000 network itself, there are several devices at each end of each segment, so they will be especially vulnerable to voltage issues. Approx 18ft of wiring is required to get the DC power from the main breaker to the helm distribution panel, another 10ft from this panel through the switch to the network power tap, and two network segments of 10ft each. Therefore to power the devices at the ends of the NMEA2000 network, we'll experience voltage loss along almost 40ft of wire.



While the accepted permissible voltage loss along the wire is stated at 3% or less, the network is only about 1/3rd of the distance, so ideally we should power each leg network so that it presents no more than a 1% loss. The standard NMEA2000 micro cable only uses 22AWG for the power wiring, so using the method presented here, we find that to maintain a 1% voltage loss (0.12VDC), we can only subject the wiring to a 350mA current demand. From our primer on NMEA2000, we know that only 7 LEN (Load Equivalency Number) can be powered on each network. Seven LEN is not sufficient for any more than one or two devices on the network, so we need to improve this situation.

Fortunately, Maretron makes available a "Mid" cable, that has identical connectors and signalling cable, but has a 16AWG power pair. In this situation, limiting the cable to the same 1% voltage loss will still allow 1.3A of current along the cable. This equates to 26 LEN, which is more than enough to power any devices we wish to add in the future. So, we're going to rip out the Micro cable, and replace it with Mid cable.

But that only solves half of the problem - or more precisely, one-third of the problem, as there is still excessive cable being used to get power to the network to begin with. The solution for this problem is to construct a relay box, so that the main power feed to the network (point B) is very short, and the remote switch located at the helm will control the relay. We have eliminated at least 20ft of wire. The side benefit is that we can probably now realize a 2% loss within the network cable while still maintaining less than a 3% overall loss. This will then allow 2.6A of current, or 52 LEN. As each network is limited in design to 50 LEN, we will have a robust network (at least in terms of power).

The relay box we are going to construct will be built within a standard NEMA-4x waterproof enclosure, and can power the two network segments. It will include two 4Amp fuses for the networks, a switching relay, and a voltage display monitor so I can do a quick check to ensure the network is being sufficiently powered. The weak link here is the NMEA2000 power tap, as it contains the 22AWG power cable. However, if this cable is kept short - no more than a foot or so, the voltage loss here should be minimized. At the system maximum of 2.5Amps, the voltage loss will be 0.08VDC, which represents a 0.75% voltage drop. This will have to suffice until manufacturers create a NMEA2000 power tap with larger AWG power wires. It should be cautioned here that the power tap itself may come with a 3 Meter pigtail, but you should realize by now to cut it off as short as possible to minimize the voltage loss.

The schematic of the wiring within the box itself reveals it is a farily simple device, and if desired, the voltage monitor may be skipped. When examining the schematic, notice that the relay's coils are independantly routed to the terminal board on the left. Even though this is a DC relay, there is no polarity requirements of the coil. I routed the coil to the terminal board so that you can power the relay by two different methods. One method - shown above, is to connect the ground of the input power to one lead on the relay's coil. This will then require a switch to supply +12V to the other coil lead. Conversely, you could connect the relay's coil to the positive side at the box, then route a switch to ground. However, if you do, you will need to add a fuse to the relay power. If the first method is used, it is assumed the power from the switch is already fused. At any rate, for safety and compliance with the USCG, you must ensure the relay is fused at some point.

Building the remote switch box.

Enough of the planning, here is where the fun begins. For the voltage monitor, I will be using an Onstate monitoring board. Detailed construction plans for my battery monitoring project can be found here.

For the enclosure, I used a Hammond NEMA-4x, 4.7" x 3.5" x 2.4" P/N: 1554F2GYCL, with holes drilled on both sides as shown for the two terminal boards. Be sure to read my review on precision drilling holes found here.

I also constructed a mounting board out of 1/16" thick G-10 epoxy fiberglass. If you have difficulty in finding this material, you can purchase small quantities from Aerospace Specialty Products. If you prefer, you can purchase a sheet-metal base from Hammond P/N: 1554FPL if you do not want to build a fiberglass one. Regardless of which one you use, the drilling will be the same.

After fabrication, and prior to the actual wiring, I pre-fit the fiberglass base to the box. Notice the cut outs in each side of the fiberglass to allow for clearance of the terminal boards. If you use the metal board, you will have to raise the terminal boards higher, or otherwise accomidate the terminal boards.

Construction is pretty simple if you have any soldering or electronic skills. I started with the fiberglass base, and installed the dual fuse holder to the left, and the relay to the right. The standoffs, 1/2" long x 4-40 threads, are used to mount the voltage display unit.

The base wiring was added, along with the voltage monitor unit. At this point, the system is completed, other than the required calibration of the monitor LEDs.

So how should the LEDs be calibrated? Since we are upgrading to Mid cable, and with our worst-case scenario of 2.5Amps, we know that we should expect a maximum of 0.08VDC loss. Since the 3% point is a loss of 0.36VDC, we need to set the point where the left-most Yellow LED is on at a voltage of 11.64VDC. This represents the loss of voltage we can tolerate before we exceed a 3% loss. Any further voltage loss will turn on the first Red LED. But you can set the monitor to any point you wish.

We can now button up the project by putting the case cover on the enclosure, attaching the relay box to the boat, and wiring the power from the main breaker to the box.

So one question remains; how large of a wire does that take. Again, this is a matter of using the same calculations; however, we first need to determine the total current requirements, and we will again use the worst-case scenario. We know that we will never require more than 2.5Amps on the NEMA2000 network. Since we have two legs, we need to supply up to 5Amps. Next, we have to determine what power the relay box will require. Depending on how the relay is powered, this figure can be 10mA for the display, plus up to 83mA for the relay. Of course, if the relay is powered externally, then only 10mA is needed. But again, we want to design for the worst case, so our current requirement is 5.1Amps.

Although I calculated that 16AWG wire could be used for the 8ft run from the main distribution panel, since I have put so much effort into this project, I used the "overkill" rule of thumb and ran 12AWG. I mounted the completed enclosure behind the "cheek" panel within about one ft. of where the NMEA2000 power Tee is at (yellow wire in the lower right corner of the photo).

After powering the system, the green LED - second from the right - was lit. This equates to 12.64VDC, which means just about zero voltage loss.

For your convenience, the reference section provides a NMEA2000 network voltage drop calculator to make these steps easy.


Parts list (of the more significant parts):

ItemPart Number
NMEA 4x enclosureHammond 1554F2GYCL
Enclosure mounting feetHammond 1554FT
12VDC SPST Automotive RelayTyco T9AP5D52-12
Feed-through style terminal boardMolex 538-38720-3205
Dual panel-mount fuseholderKeystone 3524
Fiberglass panelAerospace Specialty Products G10-062
Onstate Battery Monitor Battery Monitor Project

Calculating voltage drop
NMEA2000 Primer
Circular Mils chart (up to 24 AWG)
Handy NMEA2000 voltage drop calculator

Reader Comments

Home    Return