How to check EV battery health with a capacity estimate

A drive can help estimate the energy available across an EV's displayed battery-percentage range. With consistent readings and a matching reference, repeated estimates can help you track changes in available energy.

What the estimate measures

The method divides the trip computer's reported energy use by the fraction of displayed state of charge (SOC) used. It estimates the energy represented by the full displayed 100–0% window under the conditions of that drive.

Its main assumption is that each displayed percentage point represents an equal share of energy. SOC describes charge, while energy also depends on voltage. Voltage changes during discharge, and the dashboard applies the vehicle's own mapping, so displayed percentages need not track energy linearly. Analog Devices: SOC and SOH estimation

Shared measurements can produce related errors, so agreement between the readings does not independently verify battery capacity. The Battery Management System chapter explains how the vehicle estimates battery state.

Comparing the result with a matching new-battery value gives an estimate of energy retention. State of Health (SoH) can also describe power capability, which this drive does not measure. A diagnostic assessment checks faults separately from capacity and power retention. The mechanisms behind capacity loss and power fade belong in Battery Degradation.

Choose the readings and reference

Use a trip record that covers the complete drive. Check the vehicle's manual to establish what its consumption figure includes, especially cabin conditioning, battery heating or cooling, and parked use. A recent-distance average does not represent a longer trip.

Energy and SOC readings must cover the same period. For example, Tesla's US Model Y manual distinguishes energy use while driving from energy use while parked. Combining a driving-only energy figure with an SOC drop that includes a long parked interval leaves some consumption out of the calculation. Tesla US Model Y manual: energy use while driving and parked

For an energy-retention percentage, establish a new-battery reference for the same battery version and energy window:

• Identify the exact battery configuration and any software change affecting its accessible energy.
• Check the reference's start and stop limits. Gross capacity, net capacity and displayed 100–0% energy are different quantities.
• Confirm whether the reference includes energy available below displayed 0%, and whether its measurement conditions are comparable.

EVKX defines net capacity as all driver-accessible energy from displayed 100% until propulsion stops, including any accessible below-zero reserve. This drive calculation extrapolates only the displayed 100–0% window. The Battery Buffer and Usable Capacity chapter explains that boundary; Gross and Net Battery Capacity defines the headline capacity terms.

If a matching reference is unavailable, keep the estimate in kWh and compare similar drives over time. An earlier drive can provide a personal baseline, but it represents that vehicle's condition on that date, not necessarily its condition when new.

Record a drive and calculate the energy

Choose a journey with convenient charging and enough reserve for weather and route changes. A wider SOC interval reduces the relative effect of rounding, but there is no universal minimum interval that guarantees accuracy. Starting at 100% is optional; do not drive towards shutdown to complete the check.

1. Follow the vehicle's charging guidance. Choose a drive with the battery at a repeatable operating temperature and no battery or thermal-system warnings.
2. Immediately before driving, record the displayed SOC and reset the trip record for the complete journey.
3. Drive normally. Keep the energy record and SOC interval aligned through any stops; if parked energy is omitted and cannot be accounted for, use another drive.
4. Once safely parked at the end, record SOC, distance and consumption together, before charging or a prolonged rest. Also record trip duration, date, mileage, software version, ambient temperature and battery temperature when available.
5. Calculate reported energy, then divide it by the fraction of displayed SOC used.

For consumption reported in kWh/100 km:

Reported trip energy (kWh) = distance (km) × consumption (kWh/100 km) ÷ 100

For efficiency reported in miles/kWh:

Reported trip energy (kWh) = distance (miles) ÷ efficiency (miles/kWh)

A change from 90% to 20% is 70 percentage points, or a fraction of 0.70:

Estimated displayed-window energy (kWh) = reported trip energy ÷ ((start SOC − end SOC) ÷ 100)

With a matching new-battery reference:

Estimated energy retention (%) = estimated displayed-window energy ÷ matching new reference × 100

These equations use the vehicle's reported energy. Energy billed by a charger can include losses and auxiliary consumption downstream of its meter, so it cannot be substituted for the drive's battery-energy use. Charging loss explains the measurement boundaries.

Worked examples

These are hypothetical inputs demonstrating the calculation, not measurements of a particular model. Both assume a matching reference of 86.5 kWh for the displayed 100–0% window when new. Rounded results show the arithmetic; they do not establish measurement accuracy to a tenth of a percent.

Kilometres and kWh/100 km

A drive covers 308 km at a displayed average of 25 kWh/100 km, from 100% to 5%:

Reported energy = 308 × 25 ÷ 100 = 77.0 kWh

SOC interval = 100 − 5 = 95 percentage points

Estimated displayed-window energy = 77.0 ÷ 0.95 ≈ 81.1 kWh

Estimated energy retention = (77.0 ÷ 0.95) ÷ 86.5 × 100 ≈ 93.7%

The estimate is about 6.3% below the reference. That difference combines any permanent change with the effects of conditions and estimation error.

Miles and miles/kWh

A separate hypothetical drive covers 140 miles at 2.5 miles/kWh, from 90% to 20%:

Reported energy = 140 ÷ 2.5 = 56.0 kWh

SOC interval = 90 − 20 = 70 percentage points

Estimated displayed-window energy = 56.0 ÷ 0.70 = 80.0 kWh

Estimated energy retention = 80.0 ÷ 86.5 × 100 ≈ 92.5%

This estimate is about 7.5% below the reference. The drive starts below full charge and ends with a reserve. The two examples use different inputs; they are not a unit conversion of the same journey.

Interpret variation between drives

Rounding and the SOC interval

Suppose each SOC endpoint is rounded to the nearest whole percentage point. Each reading can differ from the underlying estimate by up to half a point, so their difference can be wrong by up to one point from rounding alone.

Holding reported energy fixed, using an interval of 9 or 11 points instead of 10 changes the calculated energy by about +11% or −9%, respectively. Using 79 or 81 points instead of 80 changes it by about ±1.3%. These ranges illustrate rounding alone; they exclude BMS estimation error, consumption rounding and omitted energy.

Consumption and available energy

Wind, hills, tyres and speed change the energy needed for a journey. Higher consumption alone does not imply a smaller battery: if the trip record captures the energy and the SOC change tracks it correctly, the calculation accounts for that increased use.

Missing energy is a different problem. If a load reduces SOC but is absent from the recorded trip energy, the numerator is too small and the calculated window energy is understated. Check the vehicle's energy accounting before adding any correction; do not assume a standard loss percentage.

Temperature and discharge rate can also change how much energy the battery delivers before reaching its limits. The DOE battery test manual specifies temperature and voltage conditions and separates capacity, available-energy and power tests for this reason. US Department of Energy: Battery Test Manual for Electric Vehicles

SOC corrections and operating-window changes

The BMS estimates charge from measurements and models. Changes in its estimate can move the displayed percentage without adding energy or restoring cell capacity. Analog Devices: SOC and SOH estimation Record an unexpected SOC jump or a correction after parking separately; do not combine that later SOC with the earlier trip's energy and treat the result as a comparable test.

Software may also change the accessible window. Balancing can recover access to energy restricted by unequal charge levels, while leaving the cells' lost storage capacity unchanged. Cell balancing Keep software changes and relevant service work in the log so that a changed estimate is not automatically labelled degradation.

Compare several drives with similar SOC windows, temperatures and energy accounting. Keep the individual results and their conditions rather than averaging incompatible tests. A sustained reduction under comparable conditions gives a stronger reason to investigate than the lowest isolated reading.

When to obtain a diagnostic test

If an unexpectedly low result accompanies battery warnings, repeated abrupt SOC changes or abnormal power limits, follow the vehicle's service guidance. Repeating a long drive cannot rule out a fault. Montes and colleagues' battery-assessment study distinguishes ongoing health estimates from capacity and internal-resistance testing, and considers faults alongside retained energy. Montes and colleagues: assessing batteries after EV use (2022)

Where available, a manufacturer test provides a defined procedure. Tesla's US Model Y manual separates a BMS-based health evaluation from an optional Battery Health Test. On equipped vehicles, the test can take up to 24 hours on an AC connection and may discharge the battery to displayed 0% while the vehicle remains parked and connected. It reports energy retention compared with new and may recalibrate the displayed range estimate. Tesla recommends it only if energy retention is a concern. Follow the manual's full prerequisites and instructions. Tesla Model Y Owner's Manual: High Voltage Battery Health

For a purchase, repair or warranty decision, ask for a report identifying the vehicle and battery, test date, measured quantity, reference value, method, conditions and relevant faults. For a capacity claim, include the applicable pass/fail threshold. Before paying for a test intended to support a claim, confirm which procedure the manufacturer accepts. Battery warranty explains the separate requirements for capacity guarantees and defect coverage.