Lithium, cell by cell.

First, the chemistry — and the word everyone gets wrong

Lithium iron is not lithium ion

The lithium that catches fire and the lithium in a well-built boat bank are not the same chemistry.

The one that makes the news is lithium-ion in the narrow sense — cobalt-based cells, NMC and LCO. Phones, laptops, e-bikes, electric cars. Fantastic energy density. The problem is thermal runaway: damage one, overcharge one, or let one get hot, and the cell can generate its own oxygen and burn in a way that water won't stop.

What's in this boat is LiFePO4 — lithium iron phosphate. No cobalt. The iron-phosphate cathode is far more thermally stable, it tolerates abuse that would put an NMC cell into runaway, and it gives up perhaps 30% of the energy density in exchange. On a boat, where you have room for another cell but not for a fire, that is an easy trade.

LiFePO4 also behaves differently in use. A cell sits at roughly 3.2 V nominal and holds that voltage almost flat across most of its capacity, then falls off a cliff at each end. That flat curve is why you can't judge state of charge from voltage the way you could with lead acid, and it's why you need a BMS that counts.

Why go lithium at all?

Everyone has an opinion. Here's ours.

Ask ten sailors and you'll get ten answers — weight, cycle life, usable capacity, no maintenance. All true. But for us the real reason is the one people talk about least: how fast you can push charge in and pull it out.

The number that matters is C-rate

C-rate is current expressed as a fraction of capacity. For a 400 amp-hour bank, 1C is 400 amps, 0.5C is 200 amps, 0.1C is 40 amps. It's the honest way to compare batteries of different sizes.

Put real numbers on it. A 400 Ah lead bank gives you perhaps 200 Ah you can actually use, and it might accept 60 amps early in the charge and a trickle by the end. A 400 Ah LiFePO4 bank gives you 320 Ah or more, and will swallow 200 amps right up until it's nearly full.

That isn't a slightly better battery. It's a different category of thing.

The bit people with lithium still get wrong

Plenty of boats fit lithium and then carry on charging it the way they charged lead — same small alternator, same modest solar array, same habits.

That throws away most of what you paid for. The reason the taper matters so much is afternoon solar. A lead bank reaches absorption by lunchtime and spends the rest of the day refusing most of what the panels are making. Lithium takes everything the array can produce, all day, right to the top.

So with lithium you want to oversize every charging source you have. More solar than seems reasonable. Wind if it suits your cruising. Shore power that can actually deliver. And the engine giving you real amps rather than a polite trickle.

And yes — the biggest alternator you can

If the bank will take 200 amps, why hand it 40? Every hour of engine time should be doing as much work as it can, because engine hours are the most expensive amps on the boat.

So: get the biggest alternator your engine, belts and regulation can carry, not simply the biggest your batteries would accept. The batteries are almost never the weak link — everything in front of them is.

What's actually on this boat

Three banks, doing three jobs

Why there's still a lead-acid battery aboard

That AGM buffer is the part people miss. A lithium BMS can disconnect the bank instantly — that's its job. If an alternator is pushing 50 amps into the bank at the moment it opens, the alternator is suddenly driving an open circuit, and the voltage spike that follows can destroy its diodes.

So the alternators charge the AGM buffer, which can never disconnect, and DC-DC chargers feed the lithium from there. The lead-acid battery is there purely to give the alternator somewhere safe to push.

The charging chain

Solar and shore power feed the system through Victron equipment, and the engine feeds it through Victron Orion DC-DC chargers — 24 V, 50 A each — taking alternator output and delivering a properly regulated lithium charge profile instead of whatever the alternator felt like producing.

Some of those chargers come on automatically with the engine. Others are on manual switches, so we can decide when to pull hard on the alternator and when to leave it alone.

Cell balancing, and why it decides everything

A bank is only as good as its worst cell

Cells in series all carry the same current, but they do not all hold the same charge.

Put eight cells in series for a 24 volt bank. Manufacturing tolerance means one of them has slightly less capacity than the others. On every charge it fills first; on every discharge it empties first. The BMS has to protect that cell, so it stops charging when the weak one hits its limit and stops discharging when the weak one hits the floor.

The whole bank is now the size of the worst cell, and the gap widens over time if nothing corrects it.

What the balancer does

A passive balancer bleeds charge off the highest cell through a resistor as heat. Simple, slow, and standard. An active balancer moves charge from the high cell to the low ones instead of wasting it — much faster, and the reason the JK boards are popular.

Balancing only happens at the top of the charge, where the voltage curve is finally steep enough for the BMS to tell cells apart. That's why a bank that never gets charged all the way will quietly drift out of balance.

Getting the JK BMS onto the boat's own screen

Custom page, running on the Raspberry Pi

The JK BMS has an app. The app wants your phone, Bluetooth range, and your attention. That's no good for something you want to glance at from the nav station, or trend over a month, or alarm on at three in the morning.

So the Pi reads the BMS and serves its own page: every cell voltage at once, the spread between highest and lowest, pack voltage, current, state of charge and temperature. It's a web page, so it comes up on the helm display, a phone, a laptop — anything on the boat's network — without installing anything.

Because it all lands in Signal K alongside everything else, the battery data sits in the same place as depth, wind and position. Which means alarms, logging and trends come essentially for free.

How Signal K ties it together →

The two mistakes that kill lithium

Both are about where you leave it sitting

1. Storing it full

Lead acid wants to be left at 100%. Lithium does not. A LiFePO4 cell held at full charge ages faster — calendar ageing, and it happens whether you use the boat or not. Leave a bank on float at full charge all winter and you will lose capacity you never spent.

2. Leaving it flat

The other end is worse but faster to notice. Run a cell below about 2.5 V and you risk permanent damage; leave a bank deeply discharged for months and the BMS's own draw can take it lower still. A bank that self-discharges into the floor over a winter may not come back.

So change your settings for storage

This is the bit almost nobody does. If the boat is hauled out or you're not aboard for a season, reduce the charge voltages so the bank rests part-charged instead of being held full.

The short version: for a season on the hard, aim to leave the bank around half to sixty per cent and let it sit there. Not full, not empty, not cycling.

Why build the bank rather than buy one

Honest trade-offs

A drop-in lithium battery is a sealed box with a BMS you cannot see into. It works, and when it stops working you buy another box.

Raw cells plus your own BMS cost significantly less per usable amp hour, and — the part that actually matters — you can see every cell. When one starts to drift you know months ahead, and you replace a cell rather than a bank.

The cost is that you are now the one responsible for getting it right. Top-balance the cells before you commission the bank, set the BMS limits properly, give the alternators somewhere safe to push, and don't leave it sitting full. Get those four right and the thing will outlast the boat's wiring.

← All projects The gear we use