22 October 2023 – New LiFePO4 Battery Installation (Winston)

When I first built SV JOANA in May 2000, I used 14 – 6V golf cart batteries (wet cells) in series/parallel to produce a 12V bank at 1640 Ah (deep cycle storage batteries). This was a common solution in that era, although a bit big. In time, I realized that this was too big, unnecessarily big, and cumbersome to charge. At that time, I had much less solar as well. The next iteration took place in May 2008, when I replaced the original 14 with 6 – 6V floor sweeper (slightly taller than golf cart size) AGM batteries, in series/parallel to produce a 12V bank at 900 Ah (usable capacity of half, or 450Ah). This seemed to be the right size for our boat. However, cooking with electricity was still a challenge and we had low voltage alarms when night sailing, requiring the generator or engine to be running for an hour at 2am. In 2012, we were early adopters of the LiFePO4 technology and bought a 380Ah battery (400Ah with a 5% safety margin) from Lithionics in Florida. Lithium batteries in general are much more suited for house storage banks. They are much more receptive to solar charging (low resistance) and always present a higher voltage than lead batteries – which is better for all 12V motors (pumps and fans) on the boat.

With the purchase of more electrical devices, more solar, and more experience cooking with electricity (we have no gas inside the boat), it was obvious that although 400Ah was sufficient for our needs, it was insufficient during the winter months (low incidence of sun) and seasonal storms (we could only last for 2 days if there was no sun). Therefore, within a few years, we realized that the optimum capacity was going to be 600Ah, not 400Ah – an important figure for “when we next replaced the house bank”.

When I replaced our 380Ah Lithionics battery in India in 2019 (it suffered from over-charging and insufficient monitoring during its 7 year usage), I was determined to “get the size right”. However, because of the extreme difficulty of importing items into India (where we were docked in 2019), I was again forced to adopt the less than desired capacity and had an Indian defence contractor build another 400Ah battery. It was just not possible to put a larger battery into that defined space, at least not while we were in India.

Well, fast forward to 2023, when I decided to set this issue right, once and for all, despite the obstacles. I ordered Winston 700Ah cells from Skypower in China – delivered to Turkey where the administrative fees cost almost as much as the batteries themselves.

Also, another big piece of the puzzle was the Battery Management System (BMS), a customized JBD 250A BMS from Mueller Energy in Australia (delivered to Diane’s niece in Australia and hand carried back to Canada by her cousins Mike and Brenda Toonders!).

I picked this particular BMS for its obvious heavy duty construction, as well as passive and active cell balancing. This BMS was reviewed by Off Grid Garage, and although it has active balancing, it only “connects” the active balancer when certain voltage conditions are met. In addition to a small external monitor, through bluetooth, the BMS offers a very good app (for my iPhone) where I can see what is happening with the cells, and where I can adjust limits for safety and performance, as well as triggers for the balancer. There is even a software switch to turn off CHARGING and both an external wired switch and software switch to turn off DISCHARGE current – a feature that is very useful at the time of installation and troubleshooting.

This new Winston battery is actually described as LiFeYPO4 and not simply LiFePO4 (the cells contains Yttrium) and boasts a nominal capacity of 700Ah, yet its actual capacity is known to often exceed this nominal rating by around 12% upon initial usage. What sets these cells apart is their extended operating voltage range of 2.8V~4.0V, wider than typical LiFePO4 cells. Also, according to factory claims and real life usage, it boasts an impressive cycle life of over 5000 cycles at 80% depth of discharge (cycle life can be as much as 7000 times at 70% DOD), making it a robust choice for sustained usage. With an operating temperature range spanning from -45℃ to 85℃ (much greater than any LiFePO4 cells), it thrives in challenging environments. Its high discharge and charge rates cater to demanding high-power and fast-charging requirements.

It was here, in Monastir Tunisia that I was finally able to construct this custom battery, after returning from Canada with the BMS. The cells and other bits and pieces have been onboard since late Feb 2023. Over the past month, I first physically stacked the cells together, using the compression plates provided. Getting these compression plates connected took a few hours, it’s a very tight fit – and it should be.

Next, together with Diane, we installed 4 nylon web straps and a bottom plate of acrylic that would serve as the “case” for the battery. Then, I wired the cells in parallel and used a small 20A 3.65V cell charger (bought from AliExpress while we were in Turkey) and charged the cells up full – to 3.65V.

It took nearly 10 days to do this parallel charging with this small cell charger, but the result was worth it. While it was charging, I took the time to mount the BMS. My initial plan had been to place the BMS on the end of the battery pack, but due to “short” leads – I was forced to mount it on the side, pressing the heat sink against the aluminum battery compression plate (which I think will serve well to dissipate heat) to save space. At this point, the battery weighs about 190 pounds (21.6kg per cell. 86.4kg total), not including the aluminum compression plates, BMS and wires).

Next, I withdrew these temporary “parallel wires”, connected the manufacturer provided series interconnects and finished installation of the BMS and its balancing leads. Then, I set about actually configuring the BMS.

Here, I discovered that due to firmware limitations, the actually battery capacity entered could only be a maximum of 655.350Ah (rather than the purchased and technological limit of 700Ah). What’s more, the “cycle capacity” maximum is only 524.280Ah, again – noticeably less than the 700Ah I built! The cycle capacity built into the BMS leaves a 25% safety margin. In other words, the battery has a capacity of 700Ah, but at 0% SOC displayed by the BMS, there is still 25% (175Ah) remaining, still usable – as long as the minimum cell voltage levels are respected. I set the minimum voltage levels to Cell 2.8V and Battery 11.2V – still above the minimum limits set by the manufacturer Winston. I actually have two other current shunts (one Victron and one Balmar) that make their own SOC calculations, so I am not fussed by the issue of SOC calculations.

I then connected up my digital load tester (bought from AliExpress while in Turkey) and drew the battery down over a 4 day period. The load test was interrupted when I burned out the device at the 16 hour point (foolishly operating it at 180W, and not “under 150W” as recommended on YouTube). I figured out that a single MOSFET had burned out, and found a very good Tunisian technician who sourced and replaced it for me (in a week).

The total capacity of the bank was determined to be 82.04Ah (until the tester burned out) plus an additional 700.76Ah (after I had the tester repaired) when the BMS shut the battery down with a remaining voltage of 11.93V. This proved the REAL capacity of the bank as 782.80Ah, even further from the BMS’s “cycle capacity” of 524.280Ah, and throwing off even more it’s SOC calculations. This video explains how to use the tester.

This is a screenshot from the BMS at the end of the discharge, when it was shut down for a low voltage condition. With this BMS, most of the parameters are adjustable, so that I can run a safe battery and protect it from damage. I’ve circled three things in red: the Discharge MOSFET is now OFF, Protection (Cell UVP) is triggered, Battery voltage is 11.93V and Current is zero. The battery would still permit charging current, but no discharge, at least not until the cell voltage increases.

The next step was to fully charge the new Winston battery using a 30A LiFePO4 charger (again, something I bought from AliExpress while in Turkey), bringing it up to the top balancing all the cells.

Then, I was finally ready to remove the Indian made ULTRALIFE battery, and prepare the area to receive the new battery. The Indian battery was removed using a halyard, perfectly lined up through the overhead hatch – and then 4 people carried it up to the aft deck. Then, I cut away the old battery box to make just a little bit more room for the new battery. After cutting away 3/4 of the old battery box, I gave it a good paint job. This is a photo of the empty space, before cutting and painting.

This was what the new battery looked like from the backside, before moving it into the new space.

I also made up this functional top cover, to protect from any short circuits due to misplaced tools or hands. This photo shows the mounted discharge switch as well.

Then, Diane and I lifted the new battery (now weighing nearly 200 pounds) using a halyard and positioned it in the right spot. This photo shows the battery blocked in place and connected.

This photo shows the battery with it’s top cover in place, and the remaining items placed around it. In summary, I am very pleased with the overall installation and expect it to last for a decade or more.

I fully charged the “old” ULTRALIFE battery and load tested it at 358 Ah (when it’s BPM shut off the discharge at 10.8V).

This screenshot from the Victron monitor shows that over it’s 4 years of use, the Indian made ULTRALIFE battery consumed 584 charge cycles, suffering a deepest discharge of 297Ah and an average discharge of 168Ah – in other words, not heavy usage.

Finally, I came across this handy image a few years ago that illustrates cycle usage. A decade ago, I naively thought that each day was “a cycle” but of course this will entirely depend on the severity of the discharge – how deeply you take the battery down.

Several other cruisers at the marina have shown interest in buying it, and I am keen to see it go to a new home where it should still have many years of life left in it.

2 thoughts on “22 October 2023 – New LiFePO4 Battery Installation (Winston)”

  1. Wade
    Great post! Did you have to source the buss bars in Tunisia or did the manufacturer supply them? We too are looking to ditch the propane and go to electric cooking, which means a bigger battery bank to bridge those cloudy days.

    1. I sourced nothing in Tunisia. I brought it all with me. Skypower provided the cells, compression plates and interconnecting buss bars (for putting the 4 cells in series). The big heavy duty tin plated copper bars (on the end, the positive and negative) I made up myself using material from AliExpress. I believe these cells are the best that money can buy, and Skypower (Julia Yu sales@skypowerintl.com) is the best contact you can make. If delivery to your European location is difficult, you may also deal with GWL in Cz, many have.

Leave a Reply

Your email address will not be published. Required fields are marked *