Lithium battery math, better than you may think

Ben Stein

Ben Stein

Publisher of, passionate marine electronics enthusiast, 100-ton USCG master.

32 Responses

  1. Saffy The Pook says:

    Good analysis and you haven’t even assigned a value to the greater charging efficiency you get from LiFePo being able to accept higher charge current above 70-80% SOC.

    All that said, there are some downsides and tradeoffs with lithium:
    1. More complexity and ways to fail. Lots of internal connections, internal electronics, and external electronics if there’s a proper Battery Management System
    2. Lithium doesn’t like to charge below freezing and doesn’t like to float at 100% SOC
    3. “Wild West” claims from manufacturers (e.g. I would assume 2000 charge cycles)
    4. To get the most life and performance out of Lithium, you really need a good BMS, not a “drop in”
    5. Your charging system very likely needs upgrades (e.g. alt cooling and disconnect protection)
    6. Max current limitations with smaller banks having fewer batteries
    7. Possible (though unlikely) catastrophic failure modes involving fire and smoke

    Don’t get me wrong, I love my lithiums but I there’s still plenty of hype out there and people need to understand that they aren’t a panacea, just a different set of tradeoffs.

  2. Ben Stein Ben Stein says:

    STP, you’re right that I glossed over the big benefit of charging efficiency. Even without the financial benefits, the quality of life benefits of not running the generator for a couple of hours to up in a few percent is really nice.

    I think your list of tradeoffs is pretty spot on though I think several of them can be mitigated through clever design. In fact, I think Mastervolt has done that with their BMS and external disconnect relays. But, I’m about 10 cycles into a lifespan that should last hundreds of times that, so it’s far too early for me to speak to reliability.

    As for the considerations to the rest of the system, you’re absolutely correct that charging and distribution need to be addressed at the same time you’re changing your batteries.

    -Ben S.

  3. Sean L Welsh says:

    Hi Ben. Great article. Since you asked, here are a couple of errata (I acknowledge you know this already and were being concise in the name of clarity):

    1. To get to the 20-hour rate, you are discharging simply at “five amps,” not “five amps per hour.” Discharging at five amps for a period of one hour is five amp-hours. You could say “five amp-hours per hour” but that’s the same as “five amps” — the hours cancel out.

    2. Generally speaking, lowering the voltage does not increase amps, but decreases them. Specifically, entirely resistive loads such as incandescent lights, many LED lights, and “universal” or DC motors all use less current at lower voltage, and thus consume less power. In fact, the power is a square relationship: if you cut voltage in half, power consumption will drop to one quarter. This is why lights look so dim with even a small drop in voltage. There are some exceptions: AC induction motors, and devices with smart controllers or regulated power supplies, such as variable-voltage LEDs, will increase their current consumption to attempt to draw the same amount of power when the voltage is lower. So resistive devices using battery power directly will actually consume a little more power (and draw a little more current) on the higher voltage of lithiums than on lead-acid. Lights will be a touch brighter, small dc motors will run a hair faster, and anything marginal to begin with might “burn out” a tad sooner on lithium. One way to mitigate this is to use a DC-to-DC converter to supply regulated 12vdc power to DC loads.

    When I did my own analysis, I came up with similar numbers to yours for cost-per-kWh delivered over the life of the battery. But the real-world numbers get better when charging from a generator, because lithiums charge at a nearly constant rate from empty to full, whereas the “acceptance rate” of lead-acid drops rapidly as the batteries get more full. This means that you will be running a lightly loaded and thus inefficient generator periodically to deliver the “topping charge” to the lead-acid batteries. So for those of us charging from a generator, the cost of those kWh going into the batteries is lower for lithium than for lead-acid. This tips the scales even further in favor of lithium.

    The biggest advantage for me, though, is no longer having to worry about ruining a lead-acid bank through over-discharge or chronic under-charging, both problems I’ve had to combat in the past. We replaced 750 aH (at 24v nominal) of AGM with just 315 aH of LiFePO4. Very happy (so far) that we made the switch, for all the reasons you mentioned.

  4. greg young greg young says:

    very interesting analysis… and confess i had missed the “power” aspect in comparisons, but your totally correct..
    however it may be even better? than your description, given that it will actually be about “area under the curve”.
    yes the LiFePo4 start at higher voltage (good) … but im thinking they have a “flatter” curve (as compared to FLA discharge voltage curve).
    and hence the “power” delivered may be even higher … eg its about the “shape” of the discharge curve, and if i recall my theory from waaaaayyy back, i believe its about area under the curve.

    .. thinking out aloud .. and off to check my theory some more, interested in others thoughts if Im on the correct track..

  5. greg young greg young says:

    .. correcting myself,
    area under a voltage-time discharge curve .. at constant current will be energy delivered ..
    however as you correctly note … devices typically want “power” and hence will increase their current draw as voltage reduces..

    so a more correct comparison between the two is about “energy delivered” ….

  6. Keith Pleas says:

    I went through this analysis last year for (6) batteries on my WhisperJet and I ended up in the middle with this version of the common 100ah AGM battery:

    $174.99, free shipping, and I wasn’t charged sales tax (10% in WA) so that was a net cost. Biggest factor over the FLA is the cycles / DOD:
    1200 @ 50% DOD
    500 @ 80% DOD

  7. Joe Pica says:

    All good stuff Ben, however a good external BMS with cell resister top balancing monitors on individual cells and temperature sensing by the charger provides the greatest flexibility. Sean is correct in that users need to understand the differences e.g. don’t like to be left on float or stored at 100% SOC or their capacity is reduced. Also can not have SOC indicated by voltage, they like to be store at approx. 60 % SOC and customized charging profiles as all LFP04 batteries are not the same. I favor individual cells in a series parallel bank. That provides an alternative if single cell fails. You can then isolate it with just a reduction in capacity. After I experienced the reduction in capacity having them on float for an extended time, I isolated them from the charger leaving them discharged to 70% soc if laying say over the winter. I provided power to the ship’s 12
    volt systems(pumps, toilet, etc) by a small separate charger acting as a power supply. As you know they can be so quickly recharged by shore, generator/alternators that isn’t much of a problem. They still last and return greater watts vrs LA batteries that you are still ahead economically even if they lose some capacity. If equipped with solar, it acts to cycle them at night so reduces the chance of reduction in SOC storage at 100%soc.

  8. Paul says:

    Nice write up and comparison. I’m in the process of battery replacement after good 10+ years of service from my AGMs. In trying to justify a Lifepo upgrade I’m coming up with a large (self) install cost.
    New inverter/charger, seperate the charge and load buses, LVC/HVC, solar controllers,…….
    The total cost when you then add in the actual lithium batteries to the support equipment is hard to justify versus the cost of alternatives that are much closer to drop in.

  9. Geoff says:

    It would be interesting to see the total cost of ownership math which includes all the new controllers etc. needed. (This is the reason we did not switch to LiFePo with out latest batteries a couple of years ago (a switch probably makes more sense now).

    It would be interesting to see how the math works out for Ben’s boat when you include all the costs to change to LieFePo, the fuel usage charging etc. VS the cost if you’d stayed with Flooded. After how many years/months does it make more sense to switch. Obviously it will depend greatly on power usage etc.

    • Ben Stein Ben Stein says:

      I’m working on the second installment of the main Mastervolt upgrade series. In that piece I’m going to break down the component costs. I haven’t yet captured fuel cost differences and I’m not sure I’ll be able to do that very well. At some point it feels like I could probably bend the numbers to say whatever I want if I make enough assumptions.

      -Ben S.

  10. Keith Pleas says:

    FWIW, I also included Firefly Oasis G31 in my evaluation. I am _very_ comfortable with AGMs over almost 2 decades of use and I really wanted to talk myself into one of these newer technologies – but I couldn’t justify it. I have had LiFePo4 in my EPCarry for a couple of seasons and love it. One more consideration – what if advances in batteries make LiFePo (or even a subsequent technology) significantly cheaper in – say – 5 years?

  11. Gary Cummings says:

    We made the switch to LiFePo4 on the Katie B. Couldn’t be happier. We went with Big Battery due to their on/off switch, modular construction, built in BMS, 300 amp fuse, 10 year warranty, & Anderson connectors. We travel 3-4 months of the year by RV. On the boat & RV we have a connector for each 170 Ah battery, with equal length 2/0 marine wires, leading to our main buss bar. They just plug in like grapes, so in 30 minutes I can move my battery bank from boat to RV (47 lbs. each). One battery always stays in the boat or RV. So far 680 Ah is more than enough for us. Not affiliated with Big Battery, just was impressed with the logic & the You Tube teardown by Will Prose.

  12. Val Vechnyak says:

    While BattleBorn batteries is a great benchmark it should be noted that off the shelf drop-in solutions is not the only game in town. I’ve put together my own 200Ah Lifepo4 battery with external BMS for just over $500. The discharge numbers are mind blowing. In my tests I easily run 21A load for 8-9 hours!

    One thing I wish I would think more about ahead of time is how I would charge it. Things get tricky and costly. I started looking at high output alternators but even with the cost aside, I really dont want to be pushing 140-180A (thats a lot of energy!) through the wires on my sailboat. I would have to get 3/0AWG wire and even then I would worry about resistance/heat at connections. Brands like Balmar would set me back another $1500-1800 for all.

    Instead I am perfectly happy with only 30-40A(max) out of my alternator at full RPM but the issue now is my 60A Hitachi/Yanmar alternator might burn itself trying to do it for long periods of time. At this moment I dont have a solution for this one…yet. Still researching….

  13. Val Vechnyak says:

    Wanted to add one thing that’s not immediately obvious to some people. Lithium battery cannot be used to start an engine. it cannot deliver the hundreds of CCA (cold cranking amps) the engine needs.

    • Ben Ellison Ben Ellison says:

      Val, they may be true of some lithium batteries but certainly not all. Good example here:

      But lithium doesn’t make much sense for fixed starting batteries because you’d be paying a lot for features starting batteries don’t generally use or need. Also, engine alternators need to be protected from possible damage if a lithium battery management system suddenly shuts the batteries down while they’re being charged.

    • Ben Stein Ben Stein says:

      From Joseph Pica:

      “…Lithium battery cannot be used to start an engine….”

      I had no problem at all with starting my two 4jh4 Yanmar Diesels from my 740 amp hr LFp04 bank. They will provide ample amps to start a diesel and certainly did so for me. Think of the popular compact lithium jump packs they are currently selling.

  14. Val Vechnyak says:

    Ben, very true. I only meant lifepo4 can’t start an engine. For example, Lithium titanate LTO would be great for it at their 10C discharge rate.

  15. Saffy The Pook says:

    LiFePo most certainly can start an engine, as long as the bank is configured appropriately. My house bank consists of three 143Ah/12V modules in parallel and each is rated for 1C continuous output, so the bank can deliver over 400A continuously. That’s more than enough to start my Yanmar 4JH3-HTE even under difficult circumstances.

    I don’t have a separate starting bank and protect my alternator with a circuit that kills the field on a warning from the BMS, backed up by a failsafe dump diode.

  16. Val Vechnyak says:

    I believe the overall Ah of your battery bank is clouding the confusion. Lifepo4 batteries should not be charged or discharged at higher rate than 1C. LTO chemistry, on the other hand, can do 10C.

    And, yes, as stated by others, given 400ah bank 1c would be 400a but this is hardly the “norm” because a) placement relatively close to the engine all that 400ah bank or a really big wire gauge. B) BMS that’s rated for 400a

    • Ben Stein Ben Stein says:


      I don’t think that’s the case. Mastervolt rates the MLi 12/5500 that I have installed in my boat at 1,800 amps for 10 seconds and 500 amps of continuous load. 1,800 amps is 4.5C and 500 is 1.25.

      Plus, the MLi 12/2750 with half the capacity is still rated for 500 amps of continuous discharge and 1,800 amps peak ( I’m wondering if that’s a misprint, but even the 90 amp hour MLi-E 12/1200 is rated for 200 amps of continuous discharge and 380 amps for 30 seconds (

      -Ben S.

    • Saffy The Pook says:

      I don’t know how you define “norm” but a 300-400Ah bank isn’t that unusual on cruising boats in the US.

      Placement of the bank relative to the engine is immaterial if you use an appropriate wire gauge. In my case, the run from the batteries to the engine is over 2m but it’s not a problem with 2/0 wire, which isn’t exceedingly big.

      As for the BMS, it doesn’t handle the large currents directly, it only monitors cells and provides balancing charge. My BMS is rated only for the string voltage and the number of batteries: 1000VDC or 65 batteries in any XSXP configuration, whichever is less. The BMS does control a contactor that sees all the current, but it’s rated for 500A continuous.

  17. Jorgen BC says:

    My favorite kind of installation is having a LiFepo4 bank serving the house needs. But you retain an AGM as a starter battery and you let your alternator (and perhaps also solar, water gen, ..)feed into this battery. Then you put a battery-to-battery charger between the two battery universes. When the AGM gets close to full, the B-2-B charger ships the excess power over into to lithium.
    In this way you retain the advantages of lead to keep your alternator happy, always have a fully charged starter battery and enjoy the happiness of lithium-power for the rest of the boat.

    • Val Vechnyak says:

      I’m doing exactly the same but I have heard of setups where DC-DC charges the starting battery instead. The thinking is that Lifepo4 can take whatever the alternator gives and no sense of limiting the output.

      • Jorgen BC says:

        Val, this sound as an worth investigating. I am not too worried because the B-2-B charger I use can handle 60A (Stirling Power). But I must admit this is less than the alternator at 115A. Most alternators, howeever, do not output their max for very long time. (If they haven’t been “fixed”, but this is another topic).

  18. Wolfgang Jansen says:

    Hi Ben, thank you for the math.
    When I read this (and some more information, lot to be found on the internet), this made me consider replacing the lead-acid battery with lithium battery.
    However, there are some assumptions questionable.

    Depth of discharge:
    Never use a FLA below 50%. Why not? See your deep cycle battery service life graph. It does not matter if you discharge 50% or 100%, in both cases you will get the same amount of energy out of the battery. (50% x 400 or 100% x 200).
    Real deep cycle batteries (not the cheap so called like the Duracell), are rated 1500 cycles at 80% DOD, 3000 cycles at 50%, etc. See for example Hoppecke.
    In contrary to Lithium batteries, lead-acid will survive 100% discharges. Absolutely not recommended, but they are recoverable. As long as they are immediately recharged to 100%. A Lithium battery 100% discharged (< 2V/cell) is chemically dead.
    So lead-acid can be used for 80%, recommended is 50%. Lithium should stay between 20-80%, you will have about 60% real capacity.

    Lifetime of real deep cycle batteries are rated 15-20 years, see for example Hoppecke. Lifetime of LFP types are never documented, only the number of cycles (which Lithium is good at). Keep in mind, testing the number of cycles by producers is a continues process and not like lot’s of people think a daily cycle or even a 2-3 days cycle. 13.000 cycles they can do within one half year (for the smaller cells). See for example: (this test actually is a very good result, the manufacturer Winston rates their own cells with C/2 or C/3, not with 1,5C like this test was done whit.)
    There is non evidence a Lithium battery will have a better lifetime than a lead-acid battery. Probably in both cases they last 10 years for sure.

    Efficiency (and Peukert):
    LFP batteries do not have 100% efficiency nor Peukert coefficient of 1. There is an internal resistant in every battery, causing heat and Peukert < 1.
    Victron states an efficiency of 92% for their LFP batteries.,8-Volt-lithium-iron-phosphate-batteries-EN.pdf
    The real deep charge batteries capacities are rated C/10, sometimes C/5, where the cheap so called ones are rated C/20.
    When you have a 900 or 1000Ah lead-acid battery bank, with C/10 (standard) you can use continuously 100A and get your rated capacity. Lithium can perform better in this perspective, but probably the cabling/fusing on your boat will be the real limiting factor.

    Both LFP as well as lead-acid should be loaded properly. For example: For LFP better not stored at 100%, where lead-acid should be stored at 100% charge. For lead-acid loading the UI characteristic is clear. For Lithium depends on the manufacturer, but a maximum constant voltage is recommended (and then stop charging, no float charging is allowed).

    Weight & maintenance:
    Lithium batteries have a huge advantage when weight is a real factor.
    Same counts for maintenance, Lithium batteries themselves hardly need any maintenance. You do not need them fully loaded, no float or raised absorbsion necessary. Deep cycle lead acid need, depending on the use, every 14 days a raised absorbsion voltage to keep them in optimum shape.

    My situation regarding lead-acid vs lithium:
    Maintenance is no issue, I have an automatic water fill system in place. Battery compartment is nicely ventilated, no issues in that perspective either.
    Battery weight is no issue, the ship is already over 40 tons (tanks empty), battery weight does not really matter.
    DC connection is an issue, when I wanted to switch to lithium. I have to split the DC bus into charge and load busses.
    Lifetime is no issue, the current lead-acid batteries last for minimum 12 years (only one time replacement in 25 years up to now, 24V 90Ah bank).

    My conclusion:
    Do your research carefully regarding switching over to lithium of not. It really depends more of your situation. Both battery chemistry have their own pro and cons. In both cases, if not properly maintained, your will not get 10+ years of lifetime.

    • Wolfgang Jansen says:

      mmm, my formatting is gone.
      Secondly, it is a 24V 900Ah battery bank.

    • Ben Stein Ben Stein says:

      I think you’re making a lot of assumptions in order to weight this towards lead-acid. First and foremost, the cost of what you call real deep cycle batteries will be higher than my cheap Duracell example.

      Additionally, looking at companies like Rolls Surrette’s spec chart doesn’t support the notion that you get the same amount of energy out of the battery no matter how deeply you discharge it. They show 2,000 cycles at 50% DoD and 1,000 cycles at 80% DoD. So, you’re giving up 50% of your cycles for only 30% more energy per cycle.

      I’m curious if you can provide more foundation to the statement that LiFePO4 needs to stay between 20 and 80% SOC. That’s not in the literature of any of the manufacturers I’ve looked at and not reflective of how the vast majority of these batteries are used in the field.

      I certainly agree that based on the installation there are cases where different chemistries make sense and no one chemistry is a solution to all problems.

      -Ben S.

      • Wolfgang Jansen says:

        Regarding cycles, you are right:
        Hoppecke gives 2800 at 50% and 1500 at 80%. 15% difference in energy.

        Actually, same counts for LFP. Mastervolt states 7000 cycles at 50% and 3500 at 80%.

  19. Wolfgang Jansen says:

    Hi Ben,
    The 80% SOC you will not find in any literature or documentation. Same as life expectancy, no information there either.
    Car manufacturers like Tesla, VW, Ford expect you to charge to 80% SOC for optimal lifetime of the battery, and only charge up to 100% before doing a long road trip.
    In the information of Victron you can read they expect 3000cycles at 70% DoD. However they do not specify the starting point (100 or 90% SOC).
    With Mastervolt you can also set the “full” level below 100% SOC. Why would they do this, if you can charge a LFP battery always to full? They give 2 years warranty (in Europe), for sure with the 100% SOC setting your life expectance will exceed this 2 years.
    Mastervolt as well as Victron specify 50% DoD for optimal life cycle.

    Tesla and some other manufacturers give 8 year battery warranty, but with the SOC normally at 80%. That is the main source of the 80% SOC for a long lifetime.
    (Their 0% SOC is the same as Mastervolt, it is not really 0%, but 10-15% to prevent damage to the battery.)

    Lots of practical information regarding LFP for marine you can find here:
    (They state 90% SOC for daily use of the LFP in marine environments.)

    Back to my situation:
    My battery bank was sized to last 2 days without charging or cruising. There is more electronisch these days, 1 day is still no problem. It is a motoryacht, I upgraded the alternators that time too, to match the battery bank. (With external Mastervolt regulators, proper sized wiring to the battery, etc.)
    The battery bank has cost € 2000.00 (11 years ago, the second one). Nowadays around € 2500.00 for a 24V 900Ah/C20 battery bank.
    Hoppecke (my battery bank is 11 years old now, I was inquiring it), will cost me around € 8000.00 for a 900Ah/C10 battery bank.
    Comparision, one Mastervolt MLI Ultra 24/5500 costs € 6000.00. That is only 24V/200Ah. I need minimal 2, better 3, to replace my FLA bank.
    For a proper lithium installation, I have to change/upgrade also all charging equipment and split the DC bus.
    I just cannot justify Lithium in my case.

    • Saffy The Pook says:

      Lithium chemistry isn’t new and isn’t that mysterious. It’s pretty well understood that the last 10% at the top (90% -> 100% SOC) and the last 10% at the bottom (10% -> 0%) is where the damage is done to cycle life. That’s where the 80% DoD recommendation for maximum cycle life comes from.

      These values are pretty evident on a graph of voltage and SOC versus charging time for a fully depleted lithium battery. The voltage curve is highly nonlinear at the extremes but linear in the middle. The transitions between the linear and non-linear regions occur right around 10% and 90%. SOC

  20. Charles says:

    I could be mistaken, but I think I read somewhere that Deka (aka Duracell) are good for 500 cycles to 50% DOD.

Join the conversation

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