A LiFePO4 voltage chart is the first thing most people look for after replacing a lead-acid bank, and it is also the first place the replacement stops behaving the way the old battery did. With flooded lead-acid, resting voltage is a usable fuel gauge — 12.7 V is full, 12.0 V is about half, and the slope between them is steep enough to read. Lithium iron phosphate does not work that way. Between roughly 90% and 20% state of charge, a LiFePO4 cell moves less than a tenth of a volt.
That flatness is a feature. It is why a 12.8 V pack still delivers close to full voltage at 30% capacity while a lead-acid bank at the same point is already sagging under load. It is also why a voltage reading alone cannot tell you how much energy is left, and why every serious pack ships with a battery management system that counts charge in and out rather than watching the terminals.
This guide gives the voltage tables for the three common platforms — 12.8 V, 25.6 V and 51.2 V — explains what conditions a reading has to meet before it means anything, and sets out the three situations where voltage is still the right measurement to take.
What a LiFePO4 voltage chart actually tells you
A voltage chart maps open-circuit voltage to approximate state of charge. Three words in that sentence carry all the weight.
Open-circuit means nothing is being drawn from the pack and nothing is being pushed into it. A reading taken while a 2 kW inverter is running is not an open-circuit reading, and the difference is not small.
Approximate is not a hedge. On a flat discharge curve, a measurement error of 50 mV at pack level moves the indicated state of charge by tens of percent. The chart is a sanity check, not an instrument.
State of charge is a percentage of present usable capacity, not of the capacity printed on the label. A pack that has faded to 85% retention still reads 100% at full charge — it simply holds less energy at that point than it did when new. Voltage tells you where you are in the tank. It says nothing about how big the tank has become.
The cell is the unit that matters
Every LiFePO4 pack is built from cells with the same characteristics, wired in series to reach the system voltage. The cell figures are the ones to memorise; everything else is multiplication. <div class=”sc-spec”> <div class=”sc-spec-h”>Single LiFePO4 cell — the numbers behind every chart below</div> <dl> <dt>Nominal voltage</dt><dd>3.2 V</dd> <dt>Fully charged (resting)</dt><dd>3.40 – 3.45 V</dd> <dt>Charge termination</dt><dd>3.65 V</dd> <dt>Discharge cut-off</dt><dd>2.50 V</dd> <dt>Flat region</dt><dd>3.20 – 3.30 V, roughly 20 – 90% SOC</dd> </dl> </div>
Series counts follow directly. A 12.8 V pack is 4 cells in series (4S), a 25.6 V pack is 8S, and a 51.2 V pack is 16S. This is why a “12 V lithium battery” is labelled 12.8 V and a “48 V” battery is labelled 51.2 V — the marketing name comes from the lead-acid bank it replaces, and the technical name is four or sixteen cells multiplied out.
That distinction matters when you are matching a battery to an inverter. An inverter specified for a 48 V nominal bus expects to see something between roughly 40 V and 58.4 V, which is exactly the operating window of a 16S LiFePO4 pack. The names differ; the window is the same.
LiFePO4 voltage chart by platform
The tables below give resting voltage at 25 °C with no load and no charge applied. Take the reading at least 30 minutes after either has stopped — see the section on surface charge for why.
12.8 V — four cells in series
<div class=”sc-spec”> <div class=”sc-spec-h”>12.8 V (4S) resting voltage vs state of charge</div> <dl> <dt>100%</dt><dd>13.6 V</dd> <dt>90%</dt><dd>13.4 V</dd> <dt>70%</dt><dd>13.2 V</dd> <dt>50%</dt><dd>13.1 V</dd> <dt>30%</dt><dd>13.0 V</dd> <dt>20%</dt><dd>12.9 V</dd> <dt>10%</dt><dd>12.6 V</dd> <dt>0%</dt><dd>10.0 V (BMS cut-off)</dd> </dl> </div>
Look at the span from 90% to 20%: seven tenths of a volt covers 70% of the usable energy. Now look at the last 10%: the same seven tenths of a volt is compressed into the final sliver before cut-off. That asymmetry is the whole story of LiFePO4 fuel gauging.
25.6 V — eight cells in series
<div class=”sc-spec”> <div class=”sc-spec-h”>25.6 V (8S) resting voltage vs state of charge</div> <dl> <dt>100%</dt><dd>27.2 V</dd> <dt>90%</dt><dd>26.8 V</dd> <dt>70%</dt><dd>26.4 V</dd> <dt>50%</dt><dd>26.2 V</dd> <dt>30%</dt><dd>26.0 V</dd> <dt>20%</dt><dd>25.8 V</dd> <dt>10%</dt><dd>25.2 V</dd> <dt>0%</dt><dd>20.0 V (BMS cut-off)</dd> </dl> </div>
51.2 V — sixteen cells in series
<div class=”sc-spec”> <div class=”sc-spec-h”>51.2 V (16S) resting voltage vs state of charge</div> <dl> <dt>100%</dt><dd>54.4 V</dd> <dt>90%</dt><dd>53.6 V</dd> <dt>70%</dt><dd>52.8 V</dd> <dt>50%</dt><dd>52.4 V</dd> <dt>30%</dt><dd>52.0 V</dd> <dt>20%</dt><dd>51.6 V</dd> <dt>10%</dt><dd>50.4 V</dd> <dt>0%</dt><dd>40.0 V (BMS cut-off)</dd> </dl> </div>
On a 16S pack the entire flat region spans about two volts. A voltmeter with 1% accuracy at 52 V has an error band of half a volt — a quarter of the whole usable range. This is not a case for buying a better meter. It is a case for not using voltage as a fuel gauge at all.
Why the flat curve breaks voltage-based gauges
Any device that estimates state of charge from voltage alone — a panel meter, a solar charge controller’s display, an inverter’s battery bar — was almost certainly calibrated against a lead-acid curve. Pointed at a LiFePO4 pack, it produces one of two failure modes, and installers see both.
The first is a gauge that sits at 100% for most of the discharge and then falls off a cliff. Because the pack holds above 13.2 V (or 52.8 V) for the majority of its usable range, the meter has nothing to report until the knee, at which point the remaining energy disappears in minutes.
The second is a gauge that reads low from the moment a load is applied. Under a heavy draw, terminal voltage drops by the load current multiplied by the pack’s internal resistance. On a 100 Ah pack pulling 100 A, that sag can be several tenths of a volt at 51.2 V — enough to move the indicated state of charge from 70% to 20% with no energy actually consumed.
Neither is a fault in the battery. Both are a measurement method applied to a chemistry it does not fit.
The BMS avoids the problem by counting coulombs: it integrates current in and out over time and corrects the count at the top of each full charge, where voltage does rise steeply enough to be meaningful. This is why a pack with Bluetooth or a screen reports a state of charge you can trust, and a voltmeter clipped across the terminals does not.
Resting, loaded, and surface charge
Three readings of the same pack within five minutes can differ by a volt. Knowing which one you are looking at is most of the skill.
Resting voltage is what the tables above describe. No load, no charger, and enough time for the cell chemistry to settle. Thirty minutes is a reasonable minimum; an hour is better.
Loaded voltage is lower than resting by the sag term. It is the right measurement when you want to know whether the pack can support a load, and the wrong one when you want to know how full it is.
Surface charge is the trap. Immediately after charging, a pack reads high — often 13.8 V or 55 V or more — because the outer layer of active material is still at charge potential. It settles within the hour. A reading taken straight off the charger will suggest a state of charge the pack does not have, and it is the single most common reason someone concludes their voltage chart is wrong. <div class=”sc-verify”> <p><strong>If a reading looks wrong, check the conditions before you check the battery.</strong> Was anything drawing current? Had the charger been off for at least half an hour? Was the pack near 25 °C?</p> <p>Cold changes the picture too. Below about 10 °C, internal resistance rises, sag under load increases, and a pack that is genuinely half full will read lower than the table says.</p> </div>
Charge and cut-off voltages are not preferences
The tables above describe where the pack sits. Two other figures describe where equipment is allowed to push it, and those are set by the cell, not by the installer. <div class=”sc-spec”> <div class=”sc-spec-h”>Charger and inverter settings by platform</div> <dl> <dt>12.8 V — charge</dt><dd>14.6 V</dd> <dt>12.8 V — cut-off</dt><dd>10.0 V</dd> <dt>25.6 V — charge</dt><dd>29.2 V</dd> <dt>25.6 V — cut-off</dt><dd>20.0 V</dd> <dt>51.2 V — charge</dt><dd>58.4 V</dd> <dt>51.2 V — cut-off</dt><dd>40.0 V</dd> </dl> </div>
A charger left on a lead-acid profile will try to hold an absorption voltage for hours and then drop to a float voltage. LiFePO4 needs neither. Once the pack reaches its charge voltage and current tapers, charging is finished; holding it there does nothing useful and accelerates ageing. Any charge controller or inverter being used with a lithium pack should be set to a lithium or user-defined profile, with equalisation and temperature compensation turned off.
Cut-off is the BMS’s job, not the inverter’s. The BMS will disconnect the pack to protect the cells, and it is designed to. But an inverter that keeps drawing until the BMS opens the contactor produces an abrupt loss of output rather than a controlled shutdown. Setting the inverter’s low-voltage disconnect slightly above the BMS cut-off gives you a warning instead of a trip.
Reading the LiFePO4 voltage chart in the charging direction
Every table above describes a pack coming down. Going up looks different, and the difference explains several behaviours that get reported as faults.
Charging happens in two phases. Through the constant-current phase the charger delivers whatever current it is set to and voltage climbs slowly along the same flat plateau — from around 52 V to perhaps 54 V on a 16S pack, covering most of the capacity being returned. Then the knee arrives. In the constant-voltage phase the charger holds at 58.4 V while current tapers towards zero, and this final phase returns only the last few percent.
Two consequences follow.
Charging appears to stall near the top. A customer watching a display see amps falling from 100 A to 10 A over the last half hour and concludes something is wrong. Nothing is. That taper is the pack finishing, and it is also the only part of the cycle where the BMS can recalibrate its charge counter against a known reference point.
Interrupting the taper leaves the gauge drifting. A system that never quite completes a full charge — because a cloudy week keeps the array from reaching absorption, for example — gives the BMS no calibration point. The counted state of charge slowly diverges from reality until the next complete charge corrects it. If a pack’s reported percentage seems to be drifting, a single deliberate full charge usually resolves it.
The same asymmetry explains why a pack that has been sitting at 100% for weeks may read slightly below 13.6 V or 54.4 V. Self-discharge on LiFePO4 is low — a few percent a month — but it is not zero, and the flat curve means a small energy loss shows up as a very small voltage change. That is normal, and it is not a fault worth chasing.
What voltage is still good for
Having spent this article arguing against voltage as a fuel gauge, three cases where it is exactly the right measurement:
Checking cell balance. If the BMS exposes per-cell voltages, a spread of more than about 50 mV between the highest and lowest cell at full charge indicates the pack is drifting out of balance. This is the most useful diagnostic a voltage reading gives you, and it is invisible at pack level.
Confirming a pack is alive. A pack reading near zero at the terminals has either an open BMS or a genuine fault. A pack reading in the normal range but refusing to deliver current is a different problem entirely. One measurement separates them.
Verifying storage state. A pack going into storage should sit near 50% — around 13.1 V, 26.2 V or 52.4 V. Storing at full charge shortens calendar life; storing near empty risks the cells falling below cut-off over months of self-discharge.
How to read a voltage chart on a supplier datasheet
Most datasheets give charge voltage and cut-off voltage but not the full SOC curve, which is reasonable — the curve is a property of the chemistry rather than of the pack. What a datasheet should tell you is the series configuration, so you can derive the rest.
If a specification lists a 51.2 V nominal pack but does not say 16S, ask. And if the stated charge voltage is not 3.65 V multiplied by the series count, ask about that too: 58.4 V for a 16S pack is 3.65 V per cell, and a figure meaningfully above it is either a typo or a different chemistry.
Cycle life figures are quoted against a test that assumed particular charge and discharge voltages, which is one reason a cycle number cannot be compared between suppliers without its conditions — a point covered in more detail in our guide to reading a cycle life claim. The safety standard those tests reference, IEC 62619, defines the requirements for industrial secondary lithium cells and batteries and is the standard most EPC and telecom buyers ask for by name.
FAQ
What voltage is a 12 V LiFePO4 battery when fully charged? About 13.6 V at rest, measured at least 30 minutes after charging stops. Straight off the charger it will read closer to 14.6 V because of surface charge.
Why does my 48 V battery read 52 V? Because a “48 V” LiFePO4 battery is nominally 51.2 V — sixteen cells at 3.2 V. A reading of 52 V puts it somewhere around 30–50% state of charge, which is normal.
Can I use a voltmeter to check how full my LiFePO4 battery is? Only roughly, and only at rest. Across the middle 70% of the range the voltage changes by less than a volt on a 51.2 V pack, so a voltmeter cannot resolve state of charge usefully. Use the BMS reading.
Is it bad to leave a LiFePO4 battery fully charged? For short periods, no. For long-term storage, around 50% state of charge is better for calendar life than sitting at 100%.
What is the minimum safe voltage for a LiFePO4 cell? 2.5 V per cell is the normal discharge cut-off, which is where the pack-level figures of 10 V, 20 V and 40 V come from. Below that, cells can be damaged permanently.
Do I need to change my charger when switching from lead-acid to LiFePO4? Usually you need to change its settings rather than the unit. Set a lithium profile, use the charge voltage for your platform, and disable equalisation, float and temperature compensation.