The Wonder and Terror of Lithium Batteries

Welcome to my Research Project

Lithium batteries are a wonderful technology. They are lighter, charge better, and are so efficient.


You’ll get so much more life out of them, they’ll last for decades and need so much less recharging.

THEY ARE SO COMPLICATED AND EXPENSIVE. No one really knows the technology. It’s really easy to screw it up and destroy your investment. And watch out for that Chinese garbage!

Seriously, you’ll never need another set of batteries if you do this right. You’ll have so much power.

YOU’RE PUTTING THEM UNDER YOUR BED? Are you insane? The technology really isn’t ready.

Welcome to the research I’ve done for the last few weeks. There is, to say the least, a lot of information out there about battery technology. And much of it is alarming, intimidating, and concerning. But most of it is pretty encouraging, if you’ve the stomach for electrical work and a bit of technology.

Apologies in advance for the length, there is a lot of material to cover here and even at that I’ll be skipping a lot of detail. Most of it is background for our choices and why we’re making them.

The Good Lithium

Safety First

Lithium batteries come in a whole range of technologies and mixes. Some earlier lithium batteries were indeed risks for fire or explosion. Safety has been one of the many factors in developing more advanced lithium battery technology, as well as improved battery performance.

We are installing are Lithium Iron Phosphate batteries, or LiFePO4, often abbreviated to LFP. LFP batteries are one of the safest Lithium technology on the market. They are safer than many of the batteries in your house, pocket or purse now.

In the video below, some cells similar to those are installing get put to an intentional destruction test. The cell terminals are shorted, creating one of the most dangerous conditions for ANY battery, regardless of chemistry. In this case, it took thirteen minutes for the battery to “explode”, and it did so with no flames or dangerous fumes. In a real world installation, the batteries will have fuses to prevent this sort of over load within seconds of it occurring.

The next video (in Chinese), several guys try to make some LFP cells blow up or burn. They over charge them intentionally, short them out, throw then in a fire and shoot them with a gun.

Just sayin’…

These are intentional tests to try to destroy these batteries or show they are dangerous. Yes, any battery can be dangerous. You CAN make them pop and smoke by overcharging. it is possible that Lithium, which is quite flammable in its elemental state, could burn and be difficult to put out. But an overcharged Lead-Acid battery produces hydrogen gas…that may be an explosive risk, too. No battery technology is completely risk free.

The Big Performance Win

The main reason to switch to LFP technology is the performance of the batteries. If you read the first part of this series, than you are familiar with the main faults of Lead-Acid batteries: taper charging, discharge limit to 50%, decreased efficiency at high loads, requirement to periodically fully charge AGM batteries, and limited cycle life. Specifically:

  • LFP batteries do not taper charge and they can usually accept a massive charge as well. They accept the full capacity of your chargers from the moment you start charging until they are full. This is huge, since the batteries will take much less time to charge. 200Ah of capacity can be replaced with a 50A charger in four hours, a 100A charger in two hours, and in most cases even in an hour with a 200A charger. That same charge would take many, many times longer to replace with Lead-Acid cells.
  • LFP batteries can be discharged to 20% without harming the battery. So fully 80% of the battery is usable. A 200Ah Lead-Acid bank will give you about 80Ah of usable power, a 200Ah LFP bank will give you 160Ah.
  • Not only can LFP batteries take massive charge loads, they can also handle large discharge loads without the same inefficiencies and loss of power of Lead-acid. If you run a 100A load on 200Ah LFP batteries for an hour, the power consumed is about 100A. The same load on an AGM would consume much more power, and may damage the battery.
  • There is no need to fully charge LFP batteries. In fact, for storage it is better to leave them in a partially discharged state. But it is less relevant since recharging to 100% is much easier without the tapering.
  • Cycle life (complete discharge -recharge) in Lead-Acid batteries is quite low. 600-800 cycles to 50% discharge is good for many AGMs. Most LFPs are rated for 2,000 cycles – at an 80% Depth of Discharge (DoD). The particular cells we are putting in claim 3,000 cycles at a regular 70% DoD, if we choose to charge them that way.

So overall you get a battery that stores and releases more energy, is easier to charge quickly, is more flexible on charge and usage patterns, and has many, many more charge cycles of use.

With good capacity, you can also reduce the number of charge cycles needed. Our present set of AGMs required charging every 30 hours or so, +/- six hours depending on wind, sun and use. Over time, that worked out to pretty much charging every day. Over 330 charge cycles per year will kill a battery designed for 6-800 charge cycles pretty quickly. Our new battery bank will need charging every three days, cutting our charge cycles to about 125 per year. Even if the LFP batteries didn’t have a 2,000+ cycle life, that reason alone would double or triple their life span compared to AGM.

Weight Savings

How do we get to a three-day run between charge cycles? More usable capacity. How do we do that? Two ways: deeper discharge/better charging and more batteries.

LFP cells are considerably lighter and smaller than their AGM counterparts. Our current 660Ah 24V house bank weighs in at a hefty 792 lbs. For that, we get 264Ah of usable capacity. The new bank of LFP cells 720Ah 24V, but has double the usable capacity of 576 Ah. And they weigh 394 pounds – half the weight. If you look at usable Amp-Hours per pound, the LFP cells deliver power at about four times the rate (1.46 Ah/Lb) over the AGM bank (0.33 Ah/Lb). I could easily have fit more than 720Ah of capacity if I wanted to.

So you can replace your AGM bank with 1/4 of the weight in LFP cells, for the half the weight you can double your working capacity. More capacity equals fewer charge cycles, and more battery life.

The Lithium Risks

As mentioned above, the physical risks to your boat and your body using LFP technology are pretty low. Not zero, but no battery is 100% safe so we can stipulate some risks no matter which type of battery you use, and move on in the discussion.

The risks of LFP technology are more financial than physical. Improperly installed and managed, you can flush a lot of money away quickly. They are more complex to use and charge, more expensive to buy, and you need to take some steps to make them idiot proof if you want a good long life to realize your Return on Investment for the project.

The Fragile Knees

The charge state of Lead-Acid batteries can easily be checked by reading the battery voltage. A fully charged battery will read with high voltage, and as the batter drains, the voltage will slowly and predictably drop. The reverse is true for charging – the voltage rises with charging, and all the various “smart” chargers will adapt their charging behavior based on the read voltage. You end up with a pretty linear relationship of voltage to charge like this:

The near linear relation between charge state and batter voltage with Lead-Acid.

LFP batteries tend to maintain a constant voltage over their charge and discharge cycle. Voltage at 95% charged is little different from at 40% charged. From a use perspective, this is excellent – devices on boats like high voltage and work more efficiently. Your lights will be brighter, your electric pumps will be more lively.

Unlike the Lead-Acid batteries, at the points of fully charged and discharged the voltage suddenly and precipitously changes. The voltage plummets as charge level approaches zero, and it starts to shoot up right around 100% charged.

The graph above shows the difference in the voltage states, and the “knees” in the graph of the LiFePO4 cell voltage as the charge varies. The problem is that LFP batteries can get really damaged by pushing them over those knees. Discharging an LFP down past the lower knee to 0% can damage the cells and reduce life cycles. Over charging past the upward knee can very quickly destroy the cells as well, resulting in the loss of your expensive batteries.

Traditional Lead-Acid chargers can not cope with LFP charging needs, and must be reprogrammed to fake them into handling them properly. This is not always possible. But there is a way around this.

Charging and Management

Batteries are stupid. They have no inter-cell circuitry, no knowledge of their own charge state. The solution to avoiding premature LiFePO4 battery death then is to ADD a brain to the process. A Battery Management System, or BMS, becomes the heart and soul of your battery setup and prevents you from stupidly destroying your batteries through mismanagement.

In theory, you can avoid damaging the batteries by running over the “knees” as I described above. You could sit near the batteries with a Multimeter, checking the voltage of each cell repeatedly, looking for the first spike upwards when recharging, or the first precipitous dip when discharging. When you spot that spike, you can turn off the charging and all will be well. When you spot that dip, you can start the charging up to keep things from running too low.

But sitting with the batteries all day with a multimeter would get tedious.

In essence, that is what a BMS does. It constantly monitors every cell for voltage, and will cut off charging to prevent over charge. It also cuts off power out if the batteries are toollow. At its most basic level, that “brain” of the system idiot-proofs the battery bank and prevents accidental destruction of the batteries. Although a simple BMS wouldn’t start charging automatically, more likely it would sound and alarm and flash some lights telling you to start the generator before you hit the lower knee.  You really wouldn’t want your “need to charge” notification to be the complete shutdown of all electrical systems on the boat.

A more fully functional system can take it steps further than acting as a simple gate-keeper to power in/power out of the batteries. The system we are employing, the Emus BMS, can control certain chargers directly and preemptively using a protocol called CAN (Controller Area Networking – it’s used in a lot of automobiles and is the underlying infrastructure of NMEA 2000). It can also trigger relays, lights and displays, provide detailed information on the battery state, alert the operator when attention is needed, balance cells that are out of voltage (more on that later…some day), speak Bluetooth to allow battery monitoring on an android device, and even has an add-on module to accept a GSM cell SIM to enable SMS notification of battery state.

In theory, one can install and use a LFP battery bank without a BMS. But the risks to the batteries are big enough so most view it as foolhardy to do so.

Charging With the Oldies

If it’s not clear yet, there are some risks using “old school” charging techniques with your flash new LFP batteries. Charging “profiles” from “Smart Chargers” will kill your LFPs in short order. Many devices simply do not know about the “knees” and will try to charge by the old rules.

On Evenstar we have the following means of charging the 24V house bank:

  • Victron Multiplus 24/3000 Inverter/Charger (70A)
  • Victron Phoenix 24/25 (24A)
  • 75A Large Frame alternator on the engine
  • Small frame alternator on the generator
  • 400W wind generator
  • 260W solar panels with a Blue Sky MPPT charge controller.

Not one of those charge sources knows the first thing about LFP batteries or their “knees”. Most of them have the potential to damage the batteries. In some cases, like the alternators, the batteries are quite likely to destroy the charge source by over drawing power from them. LFPs being charged will take every Amp they can, and any alternator that run at 100% output for any length of time will overheat and burn out.

So I’ve got to work out the charging. The Victron chargers will be replaced by new CAN-Enabled chargers. The alternator on the generator will be disabled and removed. The engine alternator needs to be de-tuned so it doesn’t cook itself, and be protected against a back surge if the BMS shuts off the charging. The wind generator will free-spin and burn itself out if the BMS shuts of the charge source, so I have to come up with something clever with that. The solar we may just leave attached to the 12V system all the time.

But it all needs re-thinking and every device, no matter how old and stupid, needs to be brought into line and taught to behave or get replaced.

The Expense

The cost is really the big nut when it comes to the downsides of LFP technology. There are three basic ways to approach LFP, and none of them will be anywhere close to the cost of just dropping in a new set of AGMs. Which isn’t cheap either – a single 4D sized 200Ah AGM costs over $1,000AUD (about $750 USD). We need six of them to replace the house bank which could run over $5,000.

The Proprietary Install

The most expensive way to convert to LFP is with a complete soup to nuts solution from a single source, installed by professionals. It will work, but it will cost a LOT. This big dollar solution would be several times the cost of what we are doing. There are some good companies doing this stuff, but you need the budget.

The Drop-in/Prebuilt Assemblage

There are really two flavors of this, but they both involve pre-built batteries (a battery is just a group of cells, wired together). Many companies have made pre-built drop-in batteries which claim to have a built-in BMS on board. They also claim compatibility with existing charging systems, but there is no guarantee. The batteries are quite expensive, and results are varied and wild as it’s tough to validate claims about them. But without a comprehensive BMS and complete integration with  the charging systems, it is difficult to see how this will really be stable and cost-effective.

The alternative “flavor” of this is to buy pre-made LFP batteries from a more known and reliable vendor like Victron or Mastervolt and work with their BMS, chaging and other supporting systems. These batteries are crazy expensive, but should work well if you use all their high quality equipment.

Do-It-Yourself (DIY)

Or the “Roll your own” battery bank for the somewhat less reverent. This is what we are doing on Evenstar. It requires that you order your own individual cells and wire them together into a battery, install and configure your own BMS, and integrate your own charging system. This is not a project for the faint of heart. If you can’t explain Ohms laws without looking it up or don’t grasp basic electricity it may not be for you. If you can’t crimp a cable or operate a multi-meter, it’s probably not for you. Even if you have a firm grasp on electrical systems, it still may not be for you.

Years ago I received an A+ in my “Electrical Systems” course at New England Tech, I installed all the instruments on the boat now, I’ve upgraded the charging and inverting systems on the boat, and I’ve crimped a LOT of cables. I’m mostly (but not completely) sure I can handle it. I’ve had to learn a lot, and immerse myself in some new concepts and ideas. It’s fascinating, and a bit intimidating. I do intend to detail our DIY install in later posts, so I won’t spell it out here.

It’s complex, and it’s a lot of work. At the end you get a completely custom setup tailored for your boat, for a lot less money.  And like all DIY projects on your boat you know where all the bodies are buried, because you put them there.

The Return on Investment

Any of those approaches are workable, and your mileage will vary in your results as will your costs. But what you need to look at is the ROI to see if it’s worth the money for you. This technology is not for every boat or every boating application. It would be overkill for most weekend and seasonal sailors, for example.

In my last post I mentioned that we’ve spent an average of $2,200 a year on new batteries for the last five years. That figure does not include the diesel I’ve burned running the generator every day and the extra wear and tear for the 2,300 hours we’ve put on the new genset since we installed it in 2015.

So – if sticking with AGMs costs $2,200/year + diesel and shortened life…how long until the batteries pay for themselves if I change to LFP?

I figure it’s about five to seven years to break even. My outside/worst case budget on this project is $20,000 – though I hope to bring it in closer to $17K-$18K.

This upgrade should save me about 700-800 generator hours per year. Call that 1,000 Liters of diesel at $1.00 or so a Liter (prices vary wildly by location). It’s also saving me 3-4 oil changes on the generator (about $30-$35 each for oil & filter), and is extending the life of my generator by a couple of years every year.

If you assume 650 hours of generator time used/year, a 10,000 hour life on a generator that costs about $20,000 to replace, for annual cost savings by doing a LFP installation you can assume:

Generator Life saved, $2/Hr $1300 $20,000 Replacement cost over 10,000 hour life, 650 hours saved/year
Fuel Savings, $1.00/L $975 ~975L
Oil Change Savings $115 3-4 oil changes
AGM Annual Expense $2,200 Not buying new AGMs in 2-3 years!
Total $4,590 Annual costs recouped by LFP install

With a $20,000 expenditure that’s a five-year return; I’m thinking seven because I’m probably overlooking a few things and being too optimistic. With the expected cycle life of 2,000 cycles, these batteries should last sixteen years, not the two to three we’ve gotten from the AGM batteries. If they last sixteen years, my annualized expense on the install is $1,250. Or a little more than half of what I spend on AGMs now.

Note that these numbers work GREAT for full-time liveaboards like us. If I were a seasonal coastal cruiser putting 50 charge cycles on my AGMs a summer and fully recharging them when I returned to my slip on Sunday night, they’d make zero sense. Keep the $20K, your batteries will probably last a decade.

But on the whole, if you live like we do, have our various requirements and limitations on power generation, and aren’t intimidated by the DIY approach it can work out well.

I hope…

Further reading:

LiFePO4 Lithium Battery Installation and Research

Related Posts Plugin for WordPress, Blogger...
This entry was posted in 24V, Batteries, Lithium. Bookmark the permalink. Follow any comments here with the RSS feed for this post. Post a comment or leave a trackback: Trackback URL.


  1. Viki Moore says:

    Lithium technology certainly seems to have its advantages. Would you mind if I put a link to your blog post in my battery study notes – here:

  2. B.J. says:

    Of course not! Link away…

  3. We have been living with our homebuilt LFP system as full-time cruisers for a year now, and it is the bomb! When we installed it, we were most looking forward to having so much usable power, but the rate of charge has been the real gainer. We spent less than $4000 for the whole shebang. We don’t use an active battery management system, we don’t seem to need one with our simple 4 cells in series system. One year of fulltime use and the battery bank is still balanced.

    1. B.J. says:

      That’s encouraging. I’m installing a bank of thirty-two 180Ah cells. It will be a 4P8S 24V bank. I’m planning to use a BMS because I don’t trust myself not to screw things up with the charging.

      How do you avoid over charging or over draining the bank? The BMS is supposed to be the safety against doing those and pushing the batteries over the edge. Without a BMS I’d be likely to forget to check constantly.

      1. Our system is only four cells in series. They are 400 amp hour Winston cells. By the book fully charged it is 14.8 volts. For safety, my solar charge controller is set to full bulk charge to 13.8 volts. I can’t turn off the float charge but I can turn it down to 0.2 amps and with a fully charged battery (+- 13.8v) it doesn’t put much more in. It can’t fight the voltage. I have a Sterling Power alternator to battery charger, set to bulk charge to 14 volts. So when the 180 amp alternator is switched on it “beats” to solar controller. The Sterling Power unit has a thermo sensor and backs off the draw if the alternator gets hot. In my observations our system does unbalance when charging. We have one cell that tends to charge 0.2 volts higher then the others. With the float setting on the alternator, the highest we see is 14.4 volts still below max charge. As a back up I was going to install a diy BMS from HouseBMS. I got as far as installing the individual cell shunts which turn on and shunt 0.7 amp if a cell goes above 3.7 volts. So far this has never happened. Early on I was a human BMS, neurotically checking state of charge voltages with the multimeter sometimes as often as every ten minutes. My observation is that as soon as you stop charging and start discharging the one high cell discharges faster then the rest and the system balances itself. This has been my experience with cells in series; I suspect it would also be true for eight cells as well. Once you start paralleling cells I am not sure how things would work out. As far as discharge I check the victron state of charge meter every night and morning; if we are low I run the motor. We rarely get below 70% SOC. We’ve only run the motor exclusively to charge the batteries once in a year.

    2. B.J. says:

      That’s encouraging. I’m installing a bank of thirty-two 180Ah cells. It will be a 4P8S 24V bank. I’m planning to use a BMS because I don’t trust myself not to screw things up with the charging.

      How do you avoid over charging or over draining the bank? The BMS is supposed to be the safety against doing those and pushing the batteries over the edge. Without a BMS I’d be likely to forget to check constantly.

Leave a Reply

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