Vector's DC electrical system

I mentioned here recently that I am making some changes to our DC powerplant. These have been on the plate for a long time, but they've moved up the list, in part, because of recent capacity problems with the house batteries. Also, I have a lot of project time on my hands. The rest of this post will concern nothing else, so if you have no interest in such things, feel free to skip it; I will return to our regular travelogue in the next post.

As background, when we purchased Vector back in 2013, she had a pretty conventional 12-volt DC marine electrical system. I say "conventional," but Vector is at the very large end of the scale for 12 volt yachts; most boats her size and larger use a 24-volt electrical system, and for good reason. Above a certain size, perhaps 85' or so, batteries are not used at all, except for emergency backup, with house power being provided by a set of various size generators, one of which is always running unless connected to shore power.

In any event, Vector's builders chose to supply a 12-volt system for primary house power. A smaller, completely separate 24-volt system was installed in the bow of the boat to run only the bow thruster and the windlass, which are 24-volt items. This system is not interconnected to anything else at all, and its pair of size 8D gel batteries are charged by a small charger located with the batteries and connected to a 120-volt receptacle fed by the house inverter.

Only one other item on the boat mandated 24-volt input power, and that was the Naiad stabilizer control system, located under the helm and installed by the last owner. Rather than run a pair of wires up from the 24-volt system deep in the bow, which admittedly would have been a lot of work, the installer chose to mount a small 12-to-24 volt converter under the helm for this application, and added a breaker to supply it with power from the 12-volt house system already feeding the helm area.

The main house bank consisted of five size 8D AGM batteries in parallel, mounted in racks along the sides of the engine room. Two more 8D batteries served as dedicated starting batteries for the main engine and generator, respectively. So all told, there were nine enormous batteries on board when we purchased the boat, in four different systems.


Vector's battery system at purchase time.

The system worked well enough for typical part-time cruising, but its limitations for full-time life aboard, with significant time away from dockside power, soon became apparent. For one, the Heart Freedom 30 inverter/charger at the center of the system was barely adequate in either mode, with just 3,000 watts of output power when inverting -- not enough to run major appliances -- and a 140-amp max charge rate, or a little less than 2,000 watts.

Under way, house DC power and battery charging was supplied by the stock main engine alternator, a 130-amp model with an internal regulator, again less than 2,000 watts in total. And the two enormous starting batteries were underutilized -- overkill for the application, and with the main engine start battery being charged only by a small "echo charge" circuit on the Heart inverter.

Upgrading this electrical system to something more appropriate to a full-time, off-grid live-aboard vessel was an early project on my list, and although I could not manage to crank it out during our first yard visit in 2013, I was able to get it done while we spent three months in Stuart at the beginning of 2014, while our friends were having their new Nordhavn commissioned there. I did not write it up at the time, because I did not have a good way to make the necessary drawings.

Now that I can do the drawings is a good time to fill in the story. Not shown on any drawings is the separate 24-volt thruster/windlass system, because really it's completely separate and not connected on the DC side to the other systems. I will note this has worked flawlessly since we bought the boat and is still on the same gel batteries that came with it. The only things I've done in there were to clean up the wiring, and add a 24-volt anchor washdown pump, which is located in the same bilge.

On the drawing of the original system you can see the five house batteries in parallel, the alternator and inverter/charger, the two separate start batteries, the interconnections among the three systems, and the main fuse. The loads are not shown, but include a handful of stand-alone breakers and fuses in the engine room powering local high-current items such as the water pump, water maker, and davit, and a large 4/0 cable running up to the helm console, where the main DC breaker panel and subpanels are located.


The 24-volt upgrade as originally installed in 2014, mostly unchanged until last week.

You can't make a 24-volt system with five 12-volt batteries, so the first order of business was to re-purpose one of the two starting batteries for house use. (I actually replaced all the batteries as part of the project as well.) The main engine starting battery, located in a compartment under the engine room sole near the engine, was physically closest to the other five batteries, so it was incorporated into the house bank.

The main engine starting and electrical system was instead connected to the same battery that starts the generator. There is really no need for two systems here; the engines are never started simultaneously, and a single battery is plenty for both. The enormous 8D was replaced by a pair of commonly available group 65 automotive batteries, which fit together in the same compartment.

The 1-2-Both battery selector for the main engine was discarded and the one for the generator was left in place, now serving the combined engine/generator battery. This selector allows the main engine or generator to be started by using the 12-volt side of the house bank in the event the start batteries have failed. It also allows the start battery to be charged by the house charger or main engine alternator if needed, and, lastly, by bridging the batteries with the "both" setting, the house 12-volt system can use power from the start batteries in an emergency.


New 24-volt distribution at left, showing disconnect switch. main fuse, and equalizer fuse. I chose orange tape to indicate 24 volts. The two breakers just right of the switch are for the heads and the new feed to the helm. Far right is the smart regulator for the 24v alternator, and in between is an extra charger to get more current from the generator.

In this arrangement, the start batteries are charged only by the 60-amp automotive alternator on the generator engine. In practice, this has been sufficient to keep the start batteries charged even through multiple main-engine starts and multi-day offshore passages, where the engine instruments run on batteries alone. But if needed the selector can be operated to the bridged position to provide charge to the engine batteries.

Using a combination of existing cables and new cables, the six house batteries were connected in a series-parallel arrangement as shown in the second drawing. The main engine alternator was swapped out for a 24-volt model using the same frame, and the inverter/charger was replaced with a 4,000-watt 24-volt model. Both the alternator and the charger are rated at 110 amps, upgrading charge capacity from under 2,000 watts to over 3,000 watts, and the 4kW inverter is enough to run our appliances.

Other than the inverter, which admittedly is the largest single consumer of power, the rest of the house loads remained at 12 volts. To operate them with the new system, they were connected to the center-tap of the 24 volt series-parallel bank. To keep the bank from being loaded unevenly, a device known as a battery equalizer was connected to both the 24 volt and 12 volt battery terminals. In simple terms, this device consumes power on the 24 volt side and supplies power on the 12 volt side to keep the batteries in balance.


The battery equalizer, upper right, shoehorned into an unused corner of the inverter shelf. At left is the inverter, and to the right is a small transfer switch to allow use of a smaller shore cable to power the inverter directly.

Equalizers are sized for the highest average sustained load, not for the maximum load. The idea being that a large load might, for a short time, draw more power than the equalizer can supply, and this excess power will come from the 12-volt side of the battery bank. Once the load drops below the equalizer's capacity, the equalizer will continue to draw charge from the 24 volt side and replace the depleted capacity on the 12 volt side until the bank is again in balance. Our equalizer is a 60-amp model, whereas our 12-volt davit can draw more than that under load, as can some combinations of other loads.

This has all worked fairly well for the past six years. The six house batteries I installed new in 2014 served until 2018, with two individual batteries giving out and being replaced over that time. Equalizers are not perfect, and the lower half of the bank inevitably gets more abuse; both failures happened in the lower half. Probably I should have sucked it up and swapped halves periodically, but that involves moving, quite literally, over a half ton of batteries.

A more annoying problem has been that our state-of-charge (SOC) meter, also known as a coulomb counter, is seldom correct. The way these meters work is by constantly measuring current flow into and out of the batteries, by means of a shunt, shown on the drawings, located between the negative terminal of the bank and the ground bus where all the loads and charge sources are connected. The meter also sees the battery voltage, and uses the product of current, voltage, and time to add to or subtract from a running total of watt-hours stored in the batteries.


Our State-of-Charge meter, left, showing percent of battery capacity remaining. At right is the control for the inverter/charger.

A problem arises because the shunt only detects current, and can not distinguish between current returned by the 24 volt system, and current returned by the 12 volt system. It assumes all current is at the system voltage as measured on the 24 volt terminal. The activity of the equalizer corrects for some of the resulting error, but not all, and as time goes on, the meter gets further and further from the correct total.

Long-time readers may remember we had this same type of system on the bus, where it worked flawlessly. But we designed the bus from the beginning to have mostly 24-volt loads. When we could not find a 24-volt lighting fixture, for example, we installed 12-volt ones in series pairs. When all was said and done, only a small handful of items were 12 volt, such as the vent fans and the satellite dish. On the boat, it's just the opposite, with the exception of the inverter. When we replaced the heads a few years ago, we got 24-volt models, and that's been the only exception, until now.

My approach to eliminating the center-tap for good was two-pronged. I wrote about the first part a few posts ago: lots of helm equipment is dual-voltage, some of it even preferring 24 volts, and there's that silly converter just to run the stabilizers. I ran a new feed from the helm down to the 24-volt main fuse in the engine room, moved some breakers around, and created a 24-volt distribution in the pilothouse. Some of this equipment is powered up 24/7, so it was an important step.


System as it now stands, with battery equalizer re-purposed to a straight voltage converter, and center tap eliminated altogether.

The second part involved changing the function of the battery equalizer into that of a voltage converter. Instead of the 12-volt output of this device being connected to the center-tap of the batteries (along with the loads), it is now connected only to the loads. A small automotive starting battery has been added in parallel to the converter output to absorb any surge loads such as the davit. As before, the converter will continue to charge that battery after use by drawing power from the 24-volt house system.

With the center tap gone, the SOC meter should now have a more accurate picture of the battery status at all times. Effective capacity of the system, and thus discharge and charge times, will be very slightly higher due to the small added capacity of the 12-volt "buffer" battery, but this should not affect accuracy. And the lower batteries should no longer be seeing any extra abuse over what the uppers see.


Rack I cobbled together for the inexpensive buffer battery. I need more parts to secure this properly. This type of battery also needs to be in a box; I'll either add the box or replace it with a type that does not require one, once I'm certain this setup is working as expected.

Since the three upper batteries are in slightly better condition than the lower ones, I will likely swap the worst lower with the best upper when next we have shore power, in the hopes that this will even out the discharge profile slightly. And I need to do some more aggressive charging, to try to recover more of the lost capacity.

This will most likely involve "cutting out" four batteries at a time from the bank, leaving a single upper and single lower battery as a pair. I can then supply all 110 amps of charge at this single pair, instead of having the charge rate max out at just 1/3 the capacity of the charger. Heavier charge rate, in this case around .43C, can "shake loose" some of the sulfation on the plates. Two or three full discharge / full charge / equalization cycles in succession might bring back as much as half of what we have lost; I'll have to do the whole sequence three times, for three pairs of batteries.


I added this cheap energy counter from Amazon to the new "buffer" battery after I finished, only to learn it only measures current (and thus energy) in one direction. Oh well; it was only $17.