Electrical bypass.

Since mentioning it in my last two posts, several people have reached out for more information about the electrical "upgrade" that I've been working on. Rather than send the fairly detailed answer to everyone individually, I am just going to write it up here. This post will contain nothing else, so anyone not interested in electrical details can just skip it; I will return to our regular travelogue in the next post.

As a background, Vector has an isolation transformer, added by her last owner whilst upgrading the shore power system to 50-amp, 240-volt from its original dual 30-amp, 120-volt inlets. Vestiges of the original dual-30 configuration persist in the layout of the main panel at the helm, which is now actually a subpanel.

New breaker "panel" for the secondary inlet. The J-box below it will get a non-metallic cover; the stainless one was all I had in my parts bin.

The isolation transformer, which turned out to have been incorrectly wired and thus not doing its job, is principally to protect the boat from galvanic corrosion when connected to shore power. The way that corrosion happens is that the boat's underwater metals, including the anodes for the cathodic protection system, the propeller and running gear, and the hull itself, once connected to the shore grounding system, can suddenly become the "protective anode" for anything on shore, such as dock pilings, and anything else connected to it, such as other boats with poor cathodic protection.

The isolation transformer solves this problem by severing the ground connection to the shore altogether. The boat serves as its own ground for the onboard electrical system, which is isolated from the shore ground, hence the name. The other way to solve this problem is to use a device called a galvanic isolator, which goes in between the boat's ground and the shore ground and prevents the flow of current unless the voltage difference between the two rises above about 1.4 volts. That does happen sometimes, in what is often referred to as a "hot marina" — the isolation transformer does not have this problem.

Gratuitous photo of the transfer switch that connects the main panel to the isolation transformer, which only works on 240vac input. I wrote this project up in a Facebook post and not here on the blog.

One consequence of an isolation transformer on a 50-amp boat is that not only does the ground not connect to the shore, but neither does the neutral. Only the two hot legs are connected; the isolation transformer secondary has a center tap and the boat's own neutral is generated there, where it is also bonded to the boat's ground. This makes for a very safe system, but it means that the boat's 120-volt (nominal) system is always exactly half of the input voltage.

This means that we must connect our 50-amp shore cable to a 240-volt (nominal) supply. Nothing will work at all if we tried to use one of those adapters that are especially common in the RV world, but can also be found in the boat world, which connects both hot legs together to a single 120-volt hot and passes the neutral through on a 30-amp plug. The isolation transformer sees that as zero volts. We do have an adapter that lets us use two 30-amp receptacles together, so long as they are on two different legs of power.

Another consequence of this arrangement is that when we find a marina that uses commercial three-phase power instead of split-phase, we have low voltage throughout the boat. While the hot-to-neutral voltage on such as system is 120 volts, and boats with conventional split-phase power input arrangements will see that on all their 120-volt circuits, the phase-to-phase voltage is just 208 volts, rather than the 240 of split-phase, and thus all our circuits now see just half that, or 104 volts. It takes a really, really long time to toast a bagel, and all our lights are dim.

A few years back, to cope with the inability to use any shore outlet with only 120 volts, I resurrected parts of the old dual 30-amp shore system. I reinstalled a 30-amp inlet on the aft deck, where there was a an abandoned 30-amp line down to the engine room, and I set this up as an alternate input directly to our inverter-charger, which is also a 30-amp device. To deal with the galvanic issue I used a galvanic isolator, two of which were also abandoned under the helm. And to make it seamless, I installed a 30-amp, 3PDT relay that automatically switches the inverter over to this alternate 30-amp input whenever it is live.

The enclosure for the 3PDT transfer relay. This has been in place for nearly a dozen years.

We've used this numerous times and it has mostly worked well when we've needed it. But the way the inverter is wired to not just one, but two hot legs on its normal feed (from the generator or the 50-amp shore cord) meant that the "max shore amps" setting on the inverter control could not account for the loads to back the charger off correctly. Let me take a moment to explain that. The explanation is lengthy, but it is fundamental to the changes I've just made.

The galvanic isolator is under this mass of wiring for the voltage converter. Two exposed screw terminals are on the left side.

Our Magnum MS4024 inverter-charger, while basically a single-phase, 30-amp device, nevertheless actually has a split-phase input and output, which means there are two hots in and two hots out. One of those feeds the charger and any loads on Hot 1, and the other, Hot 2, is just a pass-through, until the input power goes away and the inverter starts inverting. Then both Hot 1 and Hot 2 are connected to the single 30-amp inverter output.

This is the wiring diagram for the transfer relay. The light blue lines are the neutral, which must always be switched along with the hots.

We have both inputs connected, but only Hot 2 is connected to the output. When shore or generator power is available, we thus have a full 30 amps available to the charger on one leg, and a full 30 amps for loads on the output on the other leg. In order to make the inverter work on a single 30-amp leg of shore power, the relay that switches the input simply connects both Hot 1 and Hot 2 to the single available leg.

Everything works, but now the loads are connected directly to the input power, bypassing the inverter logic, and thus the inverter control has no way to see how much load there is and back off the charge rate to fit it all within a given number of amps. In order to do that, the output would have to be wired to the Hot 1 output. I had to stare at the schematics for power-input mode and inverting mode for a very long time to understand this, so don't feel bad if I've lost you along the way. You can see those diagrams for yourself here; the relevant diagrams are figures 3-1 and 3-2 on pages 42 and 43.

I want to take a moment here to say that I really, really like this dual-input single-output feature, so much so that when the inverter crapped out a year ago, I bought the exact same model to replace it, after first evaluating everything else on the market including Victron models. So changing this arrangement by, for example, switching to single-input single-output wiring was not a desired option.

Connection diagram for the new setup, with arrows to show flow. Innards of 3PDT box are detailed in a separate diagram above. The SPDT is actually inside the inverter enclosure and is straightforward.

With the output bypassing the inverter logic this way when single-leg shore power was connected, I could not set the "max shore amps" to 30 when on 30-amp power, or else we would trip the shore breaker as the charger came in on top of the loads. I had to take a guess at our total load, and set the max amps to the difference between that and 30, less some safety factor. Often I was setting input to as little as five amps, which is just enough to keep the batteries from depleting but not enough to charge them.

We actually have a second battery charger connected to the house batteries, a non-programmable charger good for about 30 amps DC and drawing maybe 8 amps at 120vac. It's there as a backup for the inverter/charger if needed, and normally it runs as supplemental to it when were are running the generator. To avoid the problems with tripping the shore breaker on 20- or 15-amp shore outlets, instead of using the arrangement I have already described to connect the inverter/charger directly to this shore power, we've taken to just running an extension cord out of the boat and plugging this auxiliary charger into it directly, and letting the inverter just run the 120vac loads from the batteries.

With all that as background, I am working on two separate projects to improve things. The first, now completed and pictured above, was to simply split that 30-amp shore inlet line, which previously went directly to the relay ahead of the inverter, into two circuits — one for the inverter relay as before, and a second for a power outlet to run the auxiliary charger. No fancy relay here; I will just move that charger's standard plug from one outlet to the other when needed.

This duplex receptacle has one outlet on the regular shore/gen system, and the other on the aft deck feed. There is no danger or real downside to "forgetting" to move the plug so no need for automation.

This lets us do what we've already been doing, but without having to prop the back door of the engine room open to run the extension cord out of the boat. There's no way to lock up the boat with that arrangement, which has given us pause to use that method when we have to leave the boat someplace. Also, it bypasses the galvanic isolator. Of course, you can't just parallel a 15-amp outlet to a 30-amp circuit, so using the single 30-amp inlet line for this meant adding a small electrical panel with separate breakers for the 30-amp circuit to the inverter and the 15-amp circuit to the charger outlet. They are not intended to ever be used simultaneously.

New MCBs in their enclosure. 16a at left goes to charger outlet, 32a at right goes to the inverter transfer relay.

I did not have the room for full-size NEMA electrical panels or even some of the Blue Sea stuff, spendy as it is. I opted to go with the European-style "MCB" items, which meant a 16a and a 32a because those are the closest available ratings. They make a miniature enclosure for a pair of these, really intended for a single RCD or two-pole breaker. Bridging the input side between the two breakers was a tight squeeze, and there was not enough room at the other end of the enclosure to make all the neutral and ground connections, so I had to add another J-box below it.

J-box for making the ground an neutral connections. Ground comes in from the isolator. You can see the two hots going up to the breaker enclosure. The Wago connectors accept up to 10 AWG.

You may recall I said I used a galvanic isolator on this bypass arrangement, and previously the ground wires to and from the isolator ran from the J-box where the bypass relay is located. With this new arrangement I had to move those wires over to the new breaker box.

This new arrangement facilitates using 15-amp circuits when that's all that is available, but it does not solve the 30-amp problem, and for that I have sourced a 30-amp DPDT relay that will be mounted inside the inverter enclosure in the wiring junction area. This relay will switch the output from Hot 2 to Hot 1 whenever the bypass relay that switches the inputs is active. The big 30-amp, 3PDT power relay that switches the inputs actually has a 12VDC coil, due simply to that configuration being the only one readily available when I built it. A small transformer supplies the 12v when the input is hot, and I've run that same 12v signal over to the new relay, which also has a 12vdc coil.

This is the relay that will move the loads from the bypass side to the controlled side of the inverter output. It will go inside the wiring box to the bottom right. 12vdc control wires are hanging loose. If you look carefully you can see there is nothing at all connected to Hot 2 Out right now.

This final piece of the project is waiting on, of all things, more AWG 12-10 quick-connect crimp terminals (the ones with the yellow barrels). I'm out of plain ones and the fully-insulated ones don't fit the wells on the relay. Previously I was delayed by getting all the way as far as making neutral and ground connections, only to find my giant supply of Wago lever-lock connectors only went as large as 12 AWG. It's always something.

Once this is complete, with charger management working correctly on a single 30-amp circuit, we will most likely opt for 30-amp power instead of 50-amp in marinas that use three-phase power when we do not need either the big air conditioners or the clothes dryer. And we have a marina stay coming up where 50-amp power is $50 per night, whereas 30-amp is just $25 per night, so we will take advantage of it then, too.

One final storm in Maine

We are underway southbound in the Atlantic Ocean, more specifically the Gulf of Maine. As I projected, I am finally getting to post as we depart Portland for points south, after a week of hunkering down for weather and knocking out some projects.

One of those "projects," by the way, turned out to be catching up on two months' worth of blog comments. Sometimes it takes all the wherewithal that I have left just to get to updating the blog, and so I am notoriously bad at keeping up with the comments. Louise has informed me that this would never fly in the quilting world. In any event, I think I am mostly caught up now.

Squirrel Point Light Station on the Kennebec. "Tilted" shed in background, left, is the boathouse, with launch tracks down to the river.

When last I posted, just over a week ago, we had just tied up in Bath, on a lovely day. I did some more walking around town, stopping at J.R. Maxwell's to reserve a table outside on the sidewalk for an early dinner. On my walk I circled around the old train station, now the visitor center, and past the Carlton Bridge, which still carries the rail line that goes all the way to Rockland. The old highway deck above the rail line ends abruptly, having been replaced by the Sagadahoc Bridge right next to it.

We returned to town at dinner time and had a nice meal at Maxwell's. We were fortunate to be under an awning during a light sprinkle. While we were sitting there the weekly BLM peaceful protest marched past us chanting. On the way home I again stopped at the very convenient IGA grocery to replenish the beer supply.

Bath station, looking unchanged from the days of passenger service.

Friday was our one good outside window to return to Casco Bay, and we dropped lines with the ebb and ran downriver to the ocean. As soon as we left the mostly pot-free Kennebec I was right back to steering around floats for the rest of the cruise. We rounded Cape Small and turned shoreward at Turnip Island, making our way to a familiar spot in Potts Harbor, where we dropped the hook. We arrived early enough in the day for me to get the main engine oil changed after a brief cool-down. We opted to remain aboard rather than brave the chill to go to the Dolphin Marina for dinner, as we had on our first visit.

In the morning we weighed anchor and made our way back to Portland, dropping the hook more or less in the same spot we had vacated five weeks earlier (map). We drove through light fog the whole cruise, even getting a late start due to visibility, but we drove out of it just as we got to Portland. It turned out to be a fairly nice afternoon there, and we made arrangements to meet our friends Stacey and Dave for dinner.

The road deck ends here abruptly on the historic Carlton bridge. It still connects to ground level on the other side.

We were able to snag a nice outside table at the Old Port Sea Grill just across the street from the tender landing, and had a great time catching up on cruising Maine. After dinner they drove over to the pier to drop off our loot, which included another pair of lithium batteries, the critical replacement pump for the master head, and our accumulated mail sent up from Green Cove Springs.

Having the pump in hand meant my fate for Sunday was sealed, and I spent the day working on the head. I will spare you the gory details, other than to say that the cause of the pump jam was a portmanteau whose second half derives from "concrete," which has accumulated over several years. I spent most of the time removing those deposits before I could install the new pump.

On our way out of Bath I snapped this pic of the USS Lyndon B. Johnson, DDG-1002, the third and final Zumwalt-class stealth destroyer, being fitted out at Bath Iron Works.

The pump came with a kit of installation parts, including the gasket that fixes it to the china, as well as the weird proprietary rubber discharge tube that connects it to the waste line. Of course, the new tube is the one designed for the tall version of the head, which we have in the guest stateroom, but it does not fit the low-profile model we have in the master. The replacement pump kit is the same part number for both.

That meant I had to reuse the old discharge tube, which was not a big deal because it's in good shape and there is not much that can go wrong with basically a rubber hose. But while the included replacement tube came with a stainless hose clamp to fit it to the new pump, the original tube was affixed with a crimp-on band similar to the ones I use to install PEX fittings. These are meant as throw-away items, including whatever they are clamping. I had to carefully cut it off using a Dremel tool, taking great pains not to damage the rubber tube.

At the Maine Maritime Museum, just downriver of the Iron Works, these five flagpoles are arranged as the masts of a full-rigged ship, framed by sculptures of a bow and stern. There are docks here for visitors arriving by boat.

It was a long, drawn-out project, but once I got the new pump installed (the old one can probably be salvaged, but that was beyond my ability in the moment) everything was working again, better than it has in quite a while. Sadly, within a couple of days we realized that the obstructions have just moved further down the sanitation hose, and we're back to using the forward head until I can replace the hose entirely.

Monday I set right to work getting the new batteries installed, since they were taking up considerable room in the saloon in their packing boxes. That meant fully charging the existing bank, then cutting back over to the old AGM batteries, which are still in place. That allowed me to disconnect and remove the lithiums so I could remove more of the compartment floor with my oscillating saw, relocating a stringer for the settee in the process.

On our way back from dinner in Portland we passed this wedding limo. Mindful of the horrific outbreak from a wedding in Millinocket, now at 175+ cases and 7 dead, we wondered just how much "social distance" this conveyance affords.

The six batteries all fit quite nicely in the compartment thus modified, although I did have to connect the third pair with some wire I had lying around. The copper bar I had ordered to make new bus bars had not arrived at our friends' house by the time we met for dinner. I did have enough bar left over to at least make the series connection for the new pair. I had all the batteries in place and cut back over by the end of the day.

I had ordered a number of items on Amazon for delivery to the Portland locker. They all had delivery scheduled for Sunday, but one item somehow did not arrive until Monday. As soon as I got the notice, just as I had wrapped up with the batteries, I tendered ashore with my backpack and made the trek to the locker. The deliveries included a battery monitor for the inverter, a remote microphone for one of our radios, replacements for the carabiner we use as a chain hook, which gave up its life here in Maine, and oil filters and test kits to replace those I'd consumed in the last week.

Looking very much like a turbo for a small engine, this is actually the new macerator pump for our head. I think a turbo is cheaper.

We tendered back ashore not long after I returned with my locker packages and enjoyed dinner at Flatbreads right by the dock. They've set up a shelter on their deck and added some propane heaters and we were quite comfortable; we've been joking that we need to invest in whoever makes those heaters.

My copper bar arrived at our friends' place Monday afternoon, and we made arrangements to have a final lunch with them Tuesday. We all underestimated how busy places would be just after lunch hour mid-week, but we found an outside table at local small chain eatery Elevation Burger and had a nice time. No idea when we will run into them again, but we look forward to it. After lunch we returned home and decked the tender.

I had to carefully cut this clamp off to remove the tube from the old pump. I used a cutting wheel on a Dremel to cut through the "crimp" at right so I would not nick the rubber.

The forecast for Tuesday evening and through Wednesday was bleak: 40-50 mph winds out of the south, with 7-9' waves in the ocean, which would send considerable swell into the harbor. And so around mid-day Tuesday we weighed anchor and moved over to an anchorage called Seal Cove, off Great Diamond Island (map), just a couple of miles from where we were. We set the anchor in deep mud and paid out a lot of chain.

That committed us to eating aboard for a couple of nights, but it was calm, quiet, and reasonably scenic there. We did ultimately see winds well in excess of 40 on Wednesday, but we were mostly protected and comfortable. I continued working on projects, and, in particular, with the copper bar now in hand I could finish up the batteries.

Weighing anchor after a full day of 40+ wind -- we were well-set in thick clay.

That meant once again shutting them down and running on the old AGMs while I installed new bus bars and the new battery monitor. For the former, I basically made two more long bars identical to the first pair. It can be argued that using longer bars to bridge all three pairs would be neater and more efficient, but that would have meant buying at least another foot of bar, plus would complicate installation and future maintenance. Having two bars to cross among three batteries meant I also had to make spacers with small pieces of the same bar so that it would all lie flat.

The battery monitor installation involved installing a "shunt" (a large, high-current, calibrated resistance) and a small electronic module which communicates with the inverter-charger. This lets the charger make decisions about charge rates based on what is actually flowing into the batteries, rather than guessing based on what's leaving the charger. Some of what leaves the charger goes to the loads instead of the batteries, so the guesswork is problematic.

Finished installation. Total space is about 21" square.

Of course, there was already a shunt in there, which supplies usage data to our State-of-Charge (SOC) meter at the helm. In theory, the new shunt and module obviate that need, because the usage information can be seen on the inverter display. That said, I wanted to keep the SOC meter, because it provides more detailed readings, and also has an alarm circuit which we have connected to a very bright warning light to tell us we need to charge.

That meant having two shunts, and I needed to move some things around to accommodate both and wire them together in series. When I initially designed the compartment and installed the SOC meter shunt, I had not counted on the kit for the inverter. Only after getting it all working did I realize that the charger was working at a disadvantage without this input.

Two shunts in series. New one, left, is for inverter/charger monitor; older one on right drives the SOC meter.

Fortunately I had enough terminal lugs and battery cable on hand to put it all together. I did have to drill out a 5/16" lug to 3/8" to make it all work; the new shunt has larger terminals. I was able to get it all back together and working by dinner time on Wednesday; I am only missing a couple of terminal boots, which were on backorder. The new system is now 100% operational, with 7.8kWh of capacity, and we are very happy with how it is working.

Yesterday morning was calm and serene, and we moved back to our old spot in Portland at mid-day, hoping to go back ashore for one last dinner and walk around town. But well before dinner time, the swell had moved back in from the ocean, and the rolling became intolerable. We ended up moving back across the harbor to a slightly different spot near Great Diamond Island (map) and eating aboard instead.

Box at right, with green LED, is the battery monitor module. It connects to the inverter with a network cable.

Today we weighed anchor at the turn of the tide and headed out to sea. Seas have been three feet on seven seconds, which is brushing the edge of comfort, but it's been tolerable and we've had a good cruise. The lobster floats are unending, and it's been a challenge to try to type here while having to steer manually every few minutes to avoid them.

Update: We are anchored in a small natural harbor near Cape Porpoise, called Stage Island Harbor (map). There is but a single other boat here, a cruising sailboat that came in just ahead of us. My route originally had us going to Cape Porpoise Harbor, just around the corner, where we might have gone ashore at the small dock and eaten at the waterfront pub there. But it was 55 and raining as we approached the cape, and with no appeal to go ashore, this spot was closer to our route and a more pleasant anchorage.

We had a lovely sunset at dinner last night in our anchorage at Great Diamond Island. A fitting finish to our Portland visit.

In the morning we will weigh anchor to have favorable tide on the Piscataqua on arrival. We have reservations at the Prescott Park docks in Portsmouth for the night. The nicer, newer concrete docks were unavailable, so we'll be on the older wooden dock with no power, but the off-season rate is favorable. We should be tied up by lunch time, and my cousins will come up from Chester to meet us in the warmth of the afternoon.

That will put us in New Hampshire tomorrow, and we will be out of Maine with three days to spare on our 60-day clock. When next you hear from me, we'll likely be in Massachusetts.

Vector gets LiFePO4 batteries

As I have mentioned in several previous posts, I just replaced our under-performing AGM batteries with lithium ion batteries, more specifically lithium iron phosphate (LiFePO₄) chemistry "drop in" models. This post is the project write-up thus far, and will contain nothing else, so feel free to skip it if you are uninterested in such matters. I will return to our regular travelogue in my next post.

New batteries installed.

A brief history.

When we purchased Vector back in January of 2013, it came to us ready-to-cruise. There was a complete and functional 12-volt house electrical system, powered by five size 8D VRLA batteries, Lifeline brand AGM specifically. While everything was more or less working, the batteries were already damaged, with a couple having burst open their cases (really), likely due to overcharging by an internally regulated automotive alternator.

This burst case was likely due to overcharging, Photo: Steve D'Antonio

I knew when we bought the boat that the electrical system was insufficient for how we planned to use it, and intended from the start to upgrade it. We were spoiled, coming from a bus conversion with a 24-volt system, eight 8D batteries, and a 7.5 kW main engine alternator, a system which could literally run air conditioning overnight after being charged by a day's driving.

Nevertheless, it had been sufficient for the boat's first ten years, and, with many other pressing projects, we opted to live with it as-is until we had the time and wherewithal to address it. I used the time to good effect to identify all the shortcomings and the most cost-effective ways to overcome them with an upgraded system.

Two of the six current Lifelines in their rack. I'm trying to sell them now.

That time ultimately came a full year later, when we settled in to a three-month stay in Stuart, Florida to join our good friends Stephanie and Martin as their new Nordhavn was being commissioned there. And while a year may sound like a long time to live with cracked battery cases and other issues, in fact we were only out cruising for less than half that time, with the remainder having been spent at marinas during our move-in and training, or in a boatyard getting all the other work done. In fact, we mostly lived in the bus during our nearly four-month boatyard stay, as the boat was too torn up to use.

The system upgrade consisted of replacing the 12-volt main engine alternator with a 24-volt model in an identical frame, increasing production from 1,820 watts up to 3,080 watts (still a far cry from the bus, with more than double that), and replacing the 12-volt inverter-charger with a 24-volt model, increasing power capacity from 2,500 watts to 4,000 watts while at the same time increasing battery charging, from 1,900 watts to 3,000 watts.

The change in voltage meant going to an even number of batteries, so we repurposed one of the two 8D starting batteries, and we bought six nice new Deka AGM batteries to replace the six old ones. I added a battery equalizer to allow us to draw all the 12-volt loads off the center-tap of the new battery bank, and re-wired the whole shebang, mostly re-using the 4/0 cables from the old system. I wrote about the project, complete with photos of the battery installation in this post, wherein I promised that an actual project write-up was forthcoming. In reality, it was another six years before I got around to doing that.

Inside each heavy corrugated UN3480-labeled shipping box, was this thinner retail package, replete with marketing. Do they sell these in showrooms? 

LiFePO₄ battery systems were already in consumer deployment at this time, with commercial systems available off-the-shelf. Also, our good friends and lithium trailblazers Cherie and Chris had already been running their roll-your-own bank on their classic GM bus conversion for over two years. So you might be wondering why did we not just include lithium as part of this major and laborious upgrade to begin with.

Two factors conspired against lithium at that time. The first was that our battery racks are in the engine room, not easily relocated, and the temperature near the racks often exceeds 120° under way. Lithiums cannot be charged at this temperature, and it shortens their lifetime even to live in this kind of environment. Even if we could work around that problem, the second was expense: roll-your-own systems will not pass an insurance survey, and the only rated system on the market at that time was prohibitively expensive on a cost-per-kWh basis. I should note here that weight savings is not a factor for us.

The batteries were well-packed in this closed-cell foam. But on one of them, they inserted the foam the wrong way and the posts missed the intended holes. This could easily have cracked the case if the box had been dropped.

Those Deka AGM batteries worked out very well. I failed to record the number of cycles, max discharge, or total number of kWh we ran through them, but they lasted four and a half years, with just a single battery being replaced due to a bad cell. They took a good deal of abuse, being undercharged regularly, over-discharged more than once, and never fully balanced due to the vagaries of a center-tapped dual-voltage system. When another of the six failed a couple of years later, we decided to replace them all, and fully expected to get another four years or so.

Once again we considered lithium. By 2018, prices had come down, making them more competitive with AGM. But we still had the issue of heat in the ER, and we were not doing a complete change-out of chargers, alternators, and other components as we had done four years earlier, so a packaged system was once again not cost-effective. "Drop-in" replacement batteries were on the market by this time, but at a significant price premium, a limited track record, and with only a single brand offering a series configuration. So once again, AGM won out, and we replaced the six Dekas with six Lifelines in September of 2018.

The Lifelines proved to be a bad choice. Unlike the less expensive Dekas they replaced, or the Trojans (and, before them, the no-name AGMs) that we had on the bus, these turned out to be overly sensitive to undercharging. In just 18 months of use, with little abuse, they had sulfated to the point of holding just 80% of their original capacity. By the time we were leaving the Bahamas the situation had become untenable, having to run the generator immediately before bed time in the wee hours of the morning, and again before starting the coffeemaker when we awoke.

The project began with cutting out the bottom of the settee. These were the straightest cuts I could make with a handheld tool in a cramped space. Note what's left of a wood block at lower right that I had to cut out. Also note that the walls under the hole are closer together than those above it.

After returning to the US and heading up the eastern seaboard we spent a few days at a dock with power in an effort to "recondition" them to perhaps 90% of their capacity. Unfortunately, we are simply not well-equipped to do this, and the minor improvement made by two reconditioning cycles was short-lived. And thus we were again in the market for batteries.

The LiFePO₄ decision.

Faced with the prospect of dropping another $3,500 on a replacement set of AGM batteries, it once again made sense to reconsider lithium. The cost per kWh has finally broken through the cost of AGM (our cost per kWh for the Lifelines was, of course, much higher due to premature failure), and there are more choices for "drop-in" replacements that can be wired in series and should pass a marine survey. Importantly, the nature of lithium batteries means we should never experience the same kind of premature failure due to chronic undercharging.

Looking forward in the compartment. Holes were drilled for the cables and a small fan, all high enough to clear the top of the batteries. 20mm fan runs when the charger is on. Behind this panel is an interstitial space between the settee and the galley cabinets.

Standing in the way of making the change was still the issue of the heat in the engine room, where the existing battery racks are located. For this to work, any new lithium batteries would have to be located somewhere else. Importantly, that new location would have to be close to the existing charge sources in the engine room: relocating inverter-chargers, battery equalizers, and other items out of the engine room would create difficult problems, not the least of which is they would no longer be protected by the automatic fire suppression system. 

The first couple of times we evaluated lithiums, the space outside of the ER that I had looked at was the vestibule between the master stateroom and the ER. Racking batteries in there in such a way that the removable sole could still be removed for maintenance would be a big project, and holes would need to be drilled through 1/4" steel plate from the ER, and re-sealed later to maintain the waterproof bulkhead. All in all, a lot of work.

Jumpers made, breaker installed. Grounds at left waiting for the meter shunt to be moved up. Small wire is fan ground.

With more drop-ins available on the market today in different form factors, we had a new option now. Taking my cue again from our friends over at Technomadia, I looked into whether any of the products might fit under our saloon settee. This is more or less directly above the main DC hardware in the ER, and there are already two chases between the spaces. Initially for routing the refrigerant lines for the upstairs HVAC units, this pair of 2" chases also carries some TV cables, the signal cables for the HVAC controls, and some miscellaneous wires. I reckoned I could get a pair of 2/0 (but no larger) battery cables upstairs by putting one in each chase.

The Selection.

Thus started a planning process that involved taking detailed measurements of the under-settee spaces, in order to assess exactly which brands of batteries, and how many, would fit in there. It turned out that, if we chose judiciously, we could get four 100aH batteries in the chosen compartment as-is, and I figured I could get six in there with a little surgery.

This enabling decision thus made, I set out to survey the market. Frankly, I was hoping to just go with the industry-leading Battle Born brand. I had already been talking to them about the application, both by phone and in person at a boat show, and I knew of several existing installations with good results, including Technomadia. Our friends up at Infinity Coach were also installing them as upgrades in RVs. They are a bit more expensive than the competition, but seemed to have better support as well.

This sheet of 3mm neoprene rubber will keep the batteries from sliding even the couple of mm permitted by the sides of the hole, and will provide some vibration isolation. The cheapest way to buy this was as "tool box liner"; 3mm neoprene marketed for battery mats was twice the price.

Battle Born makes three different products in two form factors that would provide our required capacity. Unfortunately, neither of the form factors would allow me to get even four batteries into the allotted space, let alone six. If I had been adamant about them, or they had been the only working alternative, I might have fit them in a pair of different compartments, but those were further away, adding more cable and thus voltage drop, and would have eliminated more storage.

I looked at a lot of other brands, many of which were cheaper, or more compact, or both. But most contenders could not be assembled into a 24-volt, 200aH system with an option to expand. Several vendors did not support series wiring of 12-volt batteries, and some others would support series or parallel, but not both together. Native 24-volt options either did not fit the space or could not be paralleled.

 

A note about multiple batteries:

I don't want to repeat here the entire development history of LiFePO₄ batteries, but a few words are in order about the limitations. Suffice it to say that, while the chemistry of these batteries is nothing short of amazing, the charging and discharging reactions must happen within very narrow limits or else the cells will be destroyed -- or worse, they can fail catastrophically and destroy other things, too. The limits are imposed on a per-cell basis; a nominal 12-volt battery has four 3.2-volt cells.

Making the cable ends. This proper crimp in the engine room is for a 3/8" stud on the Class-T fuse holder. A length of 3/4" red heat shrink tubing will finish this off.

Keeping the cells within limits requires that all the cells remain in "balance" (voltage and state-of-charge of each cell nearly identical), that temperature remain between 32° and 120° Fahrenheit, and that current remain below 1.0C for charging and about 1.5C discharging. All lithium battery systems use a "Battery Management System" (BMS) to achieve this. The BMS monitors temperature, voltage, and current and will disconnect the cells if limits are exceeded. The BMS also keeps the cells in balance, usually by "burning off" some charge from higher-charge cells.

The way really large LiFePO₄ battery systems are built, such as the kind that power electric vehicles, is that individual cells are assembled into one giant battery, with one consolidated BMS. Sometimes this is a messy affair, where looking under the hood you will literally see all the individual cells, the bus bars that connect them, and myriad individual small wires running back to the BMS. Some systems are modular, with blocks that look more like traditional battery shapes, but with data transfer cables to connect each module back to the master BMS.

It is quite common for hobbyists, off-grid aficionados, and even some RVers and boaters to "roll their own" LiFePO₄ systems by buying as many individual cells as they need, an off-the-shelf BMS with the proper ratings, and all the required bus bars, temperature sensors, contactors, current shunts, and other bits and bobs to construct a complete system of whatever size and voltage is required. This, incidentally, can cost as little as half what "drop in" batteries cost for a similar capacity, and it has the additional advantage that individual cells can be replaced in the event of a failure.

These lugs had to be cut off in favor of larger ones for a replacement breaker, but you can see the shrink tube and also how compact the wire becomes under the crimp. To get these to fit on the breaker I had to saw off about 1/8" on the tops. I did not have that problem with the replacement.

Drop-in batteries also employ a BMS. But because the concept is to have a completely self-contained battery that looks and acts like the (presumably lead-acid) battery it is going to replace, these batteries have only a positive and negative terminal, and the BMS is sealed inside. It has no way to communicate with other devices such as chargers, meters, or even other drop-in batteries. Though some manufacturers are starting to add Bluetooth status reporting to their drop-ins, and some, such as ours, have built-in SOC displays right on the battery.

It is this lack of communication between batteries that limits the ability to string drop-in batteries together. Many, or perhaps most, manufacturers now permit identical batteries to be wired in parallel. The technical issues of keeping each battery balanced and in limits while working together to carry a load or accept a charge are relatively straightforward in this configuration. Series wiring, however, is a different animal, and thus far, very few drop-in manufacturers will support it under warranty. And one thing is certain: there can be no center-tapped loads or charge sources (fortunately, we had already addressed our own center-tap issues at the last reconfiguration).

And the winner is:

I will spare you all the gory details of batteries I investigated that would not work. Generally the reasons were no series connections allowed, no series-parallel, would not fit the available space, poor construction, or insufficient warranty terms. Roll-your-own was eliminated due to the marine survey issue as well as lack of warranty and the fact that secure, ocean-ready mounting and installation is challenging. When all other drop-ins were eliminated, what was left was the Safari UT 1300 from Lion Energy.

The chaos of the carpentry work to fit the batteries. Louise keeps reminding me that this is actually her seat at the dinner table under which the batteries now live.

These 12-volt (nominal) batteries are rated at 105 amp-hours or around 1,300 watt-hours (hence the model designation). The manufacturer supports series configurations up to 48 volts (nominal) and identical series strings can be paralleled to increase capacity. They have a "lifetime" warranty, where lifetime is defined as 3,500 full discharge cycles with up to 20% reduction in capacity by end-of-life.

The project.

Having decided on a battery, I set out to design the installation. Four of these batteries would fit in our chosen compartment if placed on their sides, but I figured I could fit six if I cut the bottom out of the compartment, allowing them to be placed vertically. I opted for this placement out of the gate, because vertical installation would simplify the bus bars, securement, and wiring, as well as provide easy access to the aforementioned SOC meters, which are operated by pushbutton.

I measured the cable run distance from the new compartment to the existing connection points, and I measured each connection point for proper lug size. I sketched out the installation, updated the battery diagram, and wrote out a checklist of about two dozen items for installation, test, startup, and cutover. I ordered four batteries direct from the manufacturer shipped to my cousins' house, and I made up a bill-of-materials for the rest of the installation.

Fabricating the bus bars. I cut the 1/8" x 3/4" copper stock into two 7.5" sections for the parallel connections and two 3" sections for the series connections using a hacksaw. Here they've already been marked for drilling the terminal holes.

Most of those items were available on Amazon Prime. I ordered 15' each of black and red 2/0 battery cable, packages of 2/0 lugs in stud sizes of 1/4", 5/16", and 3/8" to match the various terminals, red and black heat shrink for the lugs and bus bars, 6mm terminal bolts for the batteries (to substitute for the inappropriate items included with them), terminal lug boots, 3mm neoprene mat to place under the batteries, and a 300-amp push-to-trip DC circuit breaker. I also ordered a small lithium-compatible 12-volt battery charger to top up each battery individually before installation, per manufacturer's guidelines.

I had the Amazon items sent to a locker, for the simple reason that, right now, that's a much faster delivery than to a residential address. The only item I could not source on Amazon was the copper for the bus bars; instead I ordered a two foot long section of 1/8" by 3/4" copper bar from McMaster-Carr, which arrived next-day to my cousins' place. Twenty bucks, shipping included, and McMaster actually emailed me a metallurgic certification for it.

Test certificate for the bus bar. No, I did not buy a ton of copper, but McMaster-Carr did.

I had all the gear in hand, minus the batteries themselves and the copper bar, shortly after we arrived in Portsmouth, and I immediately set about re-working the compartment and pre-running and terminating all the cables. These projects never lack unexpected challenges, and my first came as I was cutting out the "floor" of the settee compartment, a noisy, dusty process involving the oscillating saw, the only tool which could get the job done.

Despite first cutting a 2.5" hole in the bottom, sticking my flashlight and an inspection mirror down there, and carefully measuring the distance from the hole to each of the four "walls" under the floor, when I finished the floor cut I discovered a large block of wood in the very corner that would be right in my way. It was another hour with a 3/4" hole saw and the oscillating tool to cut most of the meat out of that block, leaving just the front and side edges. The corner of the cabinet is rounded, so I had to round the cut as well to avoid breaking through the outside veneer.

Since we're starting out with only four batteries rather than six, I cut the hole just big enough for the four, but all the way to one side, leaving room for another two later. I'll have to cut out the rest of the floor if and when we add those, as well as relocating an upright below it that is in the way. No sense in doing that relocation until we're sure we want those other batteries, and, also, the smaller hole constrains the four batteries from moving laterally, and I will only need to secure them from the top later.

Finished bus bars.

I drilled holes in the "back" of the compartment (which is "forward") for the two battery cables, the monitor cable, and a small 20mm flushing fan, which I then installed. This latter item gets its 12v power from the "charger on" relay that is part of a loss-of-power alarm that I designed and built while we were in the Bahamas. That means it will run whenever the main charger does, although not when the batteries are being charged by the alternator. We'll see if this is any kind of issue down the road. 

I ran the two battery cables from the engine room to the settee and installed the lugs. I have a 15-ton hydraulic crimper for this task, really the only proper tool for the job. By contrast, many of our existing battery cables, original to the boat, appear to have been done with a "hammer-type" crimping tool, which makes inferior connections. I'll be glad to see those go when we get rid of the old batteries; I could not justify making all new cables just to fix that issue.

Hydraulic crimper. For the best crimp on the 4/0 lugs I used one 70mm and one 50mm die.

After cutting the battery cables to length, leaving some slop at both ends for cable re-terminations or minor relocation of items, I used what was left over to make the short jumpers from the batteries to the ground-side meter shunt and the hot-side disconnect breaker. This latter item was required because, for now, the main Class-T fuse is remaining in the engine room, and ABYC requires overcurrent protection closer to the batteries, plus I wanted a separate disconnect at the battery compartment.

I was all ready for the batteries when Hurricane Isaias crept up on us. In an uncharacteristic stroke of good timing, the storm sent us to a dock with power, where it was easy to load the batteries aboard, and I could complete the project without fretting about running the generator for hours while the batteries were disconnected. My cousins delivered the batteries and copper bar to us just before the rain started.

Batteries in place and bus bars fitted. Note the wing bolts, all I had on hand that fit. I did not want to deform the boots on these bolts so I omitted them from the negative bar. Still room at left for two more batteries, after cutting out that section of floor.

While it would have been great to immediately get them in position, I first had to fully charge each battery individually. I bought a small 6-amp lithium-compatible charger just for this purpose, only to find that neither of its two lithium settings would charge these batteries. That went right back into the box for return to Amazon while I came up with Plan B. It was back in UPS's hands the next morning.

We don't have a 12-volt charger anywhere on the boat, but the Vanner equalizer, mentioned here in my DC electrical system write-up, provides exactly half of whatever the 24-volt system is doing. By setting the Magnum charger's Equalize voltage to exactly twice the correct charging voltage for the Safaris, I was able to bring each battery down to the engine room individually to top it up. That took about an hour and a half per battery, as they arrived at perhaps 30% charged, and I did not finish until the wee hours of the morning. I needed to work straight through, because I had disconnected the 12-volt house system in order that nothing corrupt the topping charge, and I had to reconnect it before I turned in.

Original thermal breaker/disconnect installed. Small fused wire provides power to the Link Pro state-of-charge meter. This is also a good view of the serial bus bar sitting atop the post adapters. The boots cover the bars completely.

In the meantime I got to work on the bus bars. I cut the 2' copper bar into a pair of 7.5" sections for the parallel connections, and a pair of 3" sections for the series connections. I made the four cuts by hand using my vise and a hacksaw, which is more precise than any power tools I own. I filed off the burrs and rounded off the sharp corners with a bastard file, and then I carefully lined each bar up against a pair of batteries to mark for the holes.

Making precise holes in bar stock with a handheld drill is not my forté (a drill press is really the proper tool here, just as a bandsaw would have been for the cuts), but I more or less got correctly placed and centered holes, and then I covered each bar with shrink tubing except at the terminal ends. They lack the polished appearance of factory bus bars, but not bad for twenty bucks and an hour or so of time. Plus they are oversized at 1/8" x 3/4".

In the morning I got all the batteries in position, the last one just barely squeezing in past the edge of the compartment opening, and started installing the bus bars, now sporting terminal boots at each end. Of course, this is where I learned that the Lion support team had given me the wrong spec for the terminal bolts, and the ones I had purchased for the job at the nice hardware store in Gloucester were all too long.

I think was pretty clear I was removing the adapters.

The mis-step was owing to a part of the design that is at once clever and problematic. To wit, the batteries are supplied with "post adapters" that thread into the M6 terminals atop the battery, presumably to accommodate old-fashioned lead battery clamps. The post adapter is also drilled and tapped to accept an M6 bolt. The holes on the battery itself are 12mm deep; the ones on the post adapters 20mm. When I said I needed to replace the post adapters altogether, they nevertheless gave me the 20mm measurement instead of the correct 12mm one.

The reason why these post adapters are problematic is that they are stainless steel. Anyone who has heard me lecture on RV and marine electrical systems will know that stainless is a very poor conductor and should never be used as part of a high-current path. I have seen more than one switch or fuse holder completely melted from the heat generated by the addition of just a single stainless washer between lugs (stainless washers and nuts are fine on top of a stack, where they provide only clamping force and not current path).

Positive bus. Small fused wire on battery terminal is voltage sensing for the Link Pro. Note the distortion in the boot at left from the wing bolt.

Lion actually supplies stainless M6 hardware for the terminals. Each battery comes with two 12mm bolts and two 20mm bolts, along with flat and spring washers. Also, the post adapters themselves can be used as a 12mm bolt; there are wrench flats on the top of the post for tightening. The problem with the supplied bolts is that they are wing bolts; these are expressly forbidden by ABYC guidelines and will cause a survey flag, and also, there is no way to torque them to any kind of spec, or even properly tighten them, really. And, they did not really fit under my terminal boots.

In any event, in spite of having the wrong bolts on hand, I was able to finish the job by using the supplied wing bolts until I could get proper items. Also, notwithstanding what I just wrote, I actually installed four of the post adapters on the series connections temporarily, so I was able to use the 20mm bolts I already had. The reason for this is so that I can get my clamp-on ammeter under those bus bars to see if both strings are equally sharing the load and charge. Once I have some experience with this, I will remove the post adapters and bolt the bus bars directly to the battery tops. Those connections, BTW, are the hottest spots in the battery installation, owing to the stainless.

Meter shunt relocated from ER. I did not want to shorten the wires, in case it needs to move back, so they are coiled underneath. "Load" side is on right, and the fan and meter grounds are attached here.

Because it was so critical to managing our existing, failing AGM batteries, I could not relocate the LinkPro meter shunt and its wiring from the engine room up to the saloon until I was mostly ready with the new batteries. That was the last step before making the final battery connections, and I tested all functions of the meter, reset its counts, and programmed the new settings before moving on to cutover.

Cutting over to the new lithiums.

Because I am a belt-and-suspenders guy, and the lithiums are an experiment, and our DC power system is absolutely critical (even when it is working at reduced capacity), I left all the old AGM batteries in place. I removed the meter shunt, but then reconnected the battery ground directly. And the old master 24v battery switch is still in place as a disconnect for these batteries alone. The new battery cables tee into the system just before the main Class-T fuse.

Poor planning during design means the pushbutton SOC meter is hard to engage or read when using bus bars. I keep a popsicle stick in the compartment to push the button and I can just see the LEDs. Offsetting this by just a half inch would have prevented it.

This meant that cutover really meant turning off that switch, and turning on the new breaker in the new battery compartment. I bypassed and then shut down the inverter for this task, thankful we were still on dock power. The cutover was completely without drama, and we moved into a partial discharge/recharge test.

This is where I learned that our pre-lithium-era Magnum inverter/charger has some serious issues in this configuration. It's actually in the manual, buried in an obscure description of the charging algorithm. Namely, at any battery voltage over 25.6, the charger will bypass the Bulk and Absorb stages and go directly to Float. Having set the Float voltage properly per Lion's guidance and various other sources, this voltage was not high enough to charge the batteries at all, and thus the charger just sat there doing nothing, while the batteries continued to discharge running the DC loads.

25.6 volts turns out to be where the batteries drop when they have just 12% of their capacity remaining, one of the quirks of lithium technology. Fortunately, by setting the equalization voltage to the same setting as the bulk voltage, I was able to force a charge by putting the charger into Equalize. This is a workaround at best, and so I emailed Magnum for a more permanent fix. We also tested the "normal" charge process by letting the batteries discharge to 25.6 volts, and the charger seems to work normally after that.

Positive bus after replacement with the proper bolts. The boot on left is still deformed from being forced over a wing bolt. I moved the fused power wire here after replacing the breaker, since the ring terminal was then too small.

It was a good thing that Isaias was mostly a non-event, because I literally spent our entire time at the dock working on the battery system. And after testing most of the modes and functions, it seemed all was well, right up until we left the dock. I'm not sure whether it was unplugging shore power or starting the main engine that did it, but the new 300-amp circuit breaker tripped, fortunately before we cast off lines. I was able to just reset it and we carried on, but it left me scratching my head -- 300 amps is a lot of current for this system.

The reason for the trip soon became apparent when we did our first deep cycle of the batteries on our mooring ball back at the yacht club. This nominally 300a thermal breaker was tripping at just 90 amps of continuous current. After two consecutive trips during the charge process, I dialed the charge down to 70 amps or so. Knowing it was temporary, just until the AGMs are gone for good, and also redundant for the main Class-T fuse still in the ER, I had purchased an inexpensive off-brand breaker, and that proved to be a mistake.

After getting the proper bolts I was able to boot the ground bus and clean it up. An exposed ground would have been fine; after all, the shunt is fully exposed. But this is cleaner and makes it consistent with the positive side.

With things working well enough, albeit with some work-arounds, we cast off our mooring in Portsmouth and headed north toward Portland, with an overnight stop at Wood Island. That's a relatively remote place, perfect for continued testing. It's great to once again have batteries that will carry us overnight and then some, without having to jump up first thing and start the generator, even before coffee.

Remaining issues.

Magnum got back to me, and the more permanent fix for the charger turns out to just be a new remote panel, a $210 item, which I immediately ordered to our next stop, Portland, on Amazon. They gave me a week-out delivery estimate, but there was really no other alternative. I also ordered a new circuit breaker, and the proper size M6 flange bolts, both of which were next-day items. And, of course, I started the return process for the crappy breaker that would only carry 90 amps.

I was able to hike up to the Amazon locker the day after we arrived in Portland, picking up my items and dropping off the return, and by mid-afternoon I had the new circuit breaker in, and the proper flange bolts installed. The new breaker, a 200-amp model that was of much beefier construction than the 300a item it replaced, had much larger 3/8 studs, vs. the 6mm ones on the cheap knock-off (and perhaps that should have been my first clue). That meant cutting off the 1/4" lugs and installing new 3/8" ones.

The replacement breaker. This 200a model is beefier than the original 300a model, and had much larger 3/8" studs, compared to the 6mm (1/4") studs on the original item.

By some mysterious internal Amazon process, the delivery of the new inverter/charger remote got moved up from Friday to Monday, and so Tuesday morning I made another pilgrimage to the locker to pick it up. I had it installed in time to run a full charge cycle on the new settings, called "Constant Current to Constant Voltage" mode, and that's all working now, with the charge cycle starting right up once the genny is running.

The Link Pro battery monitor, however, is not coping well with these batteries. Its algorithms are really built around lead-acid chemistry, and there does not seem to be enough leeway in the settings to force it to do the right thing with lithium. I've finally got it dialed in to where the SOC% is close to correct most of the time, but the meter needs to be manually synchronized any time the batteries are fully charged, which, with these batteries, should only be at the dock or after several hours under way.

Post adapter. These should really be brass, not stainless, which conducts poorly.

Speaking of under way, I've also had to adjust all the charge settings on our Balmar regulator for our main engine alternator. Those are mostly working correctly now, although I will need to install the alternator temperature sensor since the lithiums draw more current at low RPMs, where cooling is inadequate. I have the sensor available, having removed it from the batteries -- lithium batteries do not need, nor should they have, temperature compensation during charging.

Initial evaluation of Lion Safari batteries.

I had some trepidation ordering these after watching this YouTube video from solar and lithium expert Will Prowse comparing their construction to the Battle Born product, and this follow-on video tearing down the battery. Ultimately, though, these fit where the Battle Borns did not, and I am less concerned about environment and vibration in our saloon than many boaters would be in less controlled compartments. The lower price and excellent warranty also helped sway the decision.

That said, there have been a few issues to date:

1. There is no date or version number on the user manual. I found multiple versions online, in addition to the printed version that arrived with each battery. Manuals should be dated so that users know they have the most up-to-date guidance on charge settings and other critical issues.
2. The stainless steel post adapters are a design flaw. If post adapters are an integral part of the design, they should be nickel-plated brass.
3. Wing bolts are not permitted or acceptable fasteners, and should be eliminated. Fasteners should ideally be hex-head, as this is most common for torque wrenches.
4. No torque recommendation is provided for the fasteners.
5. Charge setting recommendations are not clear and consistent, but ought to be. It's not even consistent throughout a single manual.
6. The operating button and SOC display LEDs are in a direct line with the battery post, meaning they will be obscured by lateral bus bars.
7. One battery arrived with its post adapters (which are shipped installed -- why?) dug into the packing foam, instead of positioned at the holes in the foam for this purpose.
8. Answers from tech support were not confident or even correct. There was no acknowledgement or apology when confronted with the fact that they had provided incorrect dimensions, and no explanation for why the manual is inconsistent on settings.
9. There is no SOC table nor any charge/discharge curves provided for these batteries. I had to extrapolate from data provided by other manufacturers or well-known parameters for LiFePO₄ chemistry.

On top of the above, I can confirm the observations that Will made about the size of the terminals and quality of the casing. Not really a problem for us, but it gives me pause in recommending these batteries for most RV or marine installations.

The serial bus bars, atop the stainless post adapters, are the hottest things in the compartment.

On the plus side, the built-in SOC meters are handy, I personally like button terminals for bus-bar applications, the cases are clean and functional, and, so far, performance is as advertised. Customer support, while not on top of their game technically, are at least very responsive.

I will be continuing to tweak charger, alternator, and monitor settings over the next few weeks as I build a charge/discharge profile from history. If all goes well, we will likely add two more batteries when they go on sale. The 210 aH will carry us for 12-14 hours under normal loads in temperate climates, and upping that by 50% will mean we will not have to run the generator at all when we travel every day, a feature we enjoyed with our AGM batteries when they were working properly.

Project Costs.

I bought manufacturer-direct using a discount code (and you can use my discount code, "6q" if you buy the same way, or use this link). That brought the list price of $999 each down to $849 each, shipping included. The best price I could find from dealers was $899. Costco previously sold these for $699, but that deal is gone.

By contrast, the battery cases are much cooler.

On top of the $3,400 for the batteries, I spent another $440 on parts, including cables, lugs, boots, bus bar, shrink tubing, bolts, a circuit breaker, and, of course, the new remote panel for the inverter. That does not count sweat equity or miscellaneous parts I already had lying around, such as the flushing fan. Call it $4k in round numbers.

If we actually get 3,500 cycles from these at 80% DoD each time, that would be 14,560 kWh stored/used over the course of their lifetime, for a cost per kWh of around 26.4 cents (yes, it really does cost twice as much as grid power just to store energy, and that's on top of whatever it cost to produce or buy to begin with). If we add two more batteries, that will decrease very slightly since I won't need more cables or another inverter remote.

New remote, showing the "Constant Voltage" mode charge for lithiums. Actual current into the batteries is shown on the SOC meter at left; the Magnum remote reads about 10% high, and some of the current is running DC loads.

By comparison, our last set of AGM batteries cost $3,690 and delivered just 2,520 kWh, for a whopping $1.46 per kWh. The set before that probably delivered three times as much energy, although I failed to record it, for right around $3k, or about 40 cents a kWh. An optimistic view of the Lifeline ratings suggest they should have delivered at least twice what they did, but that's still expensive on a per-kWh basis.

Even if the lithiums do not deliver the warranted 3,500 cycles (and there is evidence they often deliver far more than that), our cost per kWh store should still go down. But beyond that, because the charge efficiency of these batteries is so much higher, with the charger running at full capacity from start to finish every time, our cost to produce that energy will also be lower. That cost, incidentally, is right around 80 cents per kWh at anchor. Solar power would be about a third of that and I really need to look into getting some panels.

Updated battery layout. I omitted the old batteries, still installed, but included their disconnect switch. The PowerMax charger shown is actually connected to the old bank for maintenance and will be moved as shown shortly.

Appendix: Settings.

Some folks will inevitably ask, so here are the settings I am using right now for the three key devices. I expect to tweak at least some of these in the coming weeks, particularly the LinkPro meter:

Magnum MS4024 Inverter/Charger

Software version 5.5

ME-ARC remote, version 4.0

• LBCO 24.0 VDC
• VAC Dropout 80
• Battery Type CC/CV
• Max Charge Amps 120
• CV Charge Volts 28.2
• CV Charge Done Amps 30
• Max CC/CV Time 2.8 Hours
• Recharge Volts 26.0
• Max Charge Rate 100%
• Battery temperature sensor removed

Balmar MC-624 Regulator

Software version 3.0

• Basic Program P07 Halogen (my unit is too old to have an LFP program)
• Start Delay 45 Seconds
• Compensation Limit 29.2 volts
• Bulk Voltage 28.0
• Bulk Time (minimum) 12 minutes (the lowest that can be set)
• Absorption Voltage 27.8
• Absorption Time (minimum) 12 minutes
• Float Voltage 26.0
• Float Time (minimum) 12 minutes
• Battery temperature sensor removed
• Alternator temperature sensor installed

Xantrex LinkPro SOC meter

Firmware Version 1.06

• F1.0 Float (finishing) voltage 26.0
• F1.1 Charger ending current percent 2.5
• F1.2 Autosync Time 5 seconds
• F1.3 Discharge Floor 15%
• F2.0 Low SOC 15%
• F2.1 Low voltage 25.6
• F2.2 Low SOC clear 80%
• F2.3 Low battery alarm delay 10 seconds
• F5.0 Battery Amp-hours 210
• F5.4 Peukert Exponent 1
• F5.5 Self-discharge rate 0.2%
• F5.6 Charge Efficiency Factor 100%