Cell Capacity Testing: Standards, Diagnostics and Safety

Capacity testing underpins virtually every other cell characterisation measurement. State of health calculations, degradation modelling, thermal management design, and pack-level energy budgets all trace back to a reliable baseline capacity figure. This page covers the standards that govern how that figure is established, the distinction between formation and lifecycle testing, the mathematics of capacity tracking, chemistry-specific performance limits, non-destructive diagnostic methods, and the safety engineering required when testing energetic cells.

For the step-by-step test procedure, see Cell Capacity Test Procedure.

Regulatory and Standards Frameworks

Capacity validation is governed by distinct frameworks depending on the application. The most important practical distinction is between performance standards and safety certification standards — they evaluate different things and the results are not interchangeable.

Consumer Electronics: IEC 61960-3

IEC 61960-3 is the primary performance standard for portable secondary lithium cells. It validates whether a cell can reliably deliver its advertised capacity and cycle life. It is distinct from IEC 62133-2, which is a safety certification standard covering overcharge, short circuit, crush, and thermal abuse.

IEC 61960-3 testing must be performed by ISO/IEC 17025 accredited laboratories. The standard includes the following sub-tests:

Rated Capacity Discharge (20°C): Following standard charging, cells are discharged at a defined constant current to their cut-off voltage at 20°C. This establishes the baseline capacity figure.

Variable Temperature Discharge: Cells are discharged at 0.2C at 0°C and at -20°C to evaluate cold-weather performance. Typical pass thresholds are at least 70% of nominal capacity at 0°C and 50-60% of nominal capacity at -20°C, reflecting how sharply usable capacity falls as electrolyte conductivity and lithium-ion diffusion slow down in the cold.

Charge Retention and Recovery: Fully charged cells are stored for 28 days at 20 ± 5°C ambient (or 7 days at 40°C for an accelerated test) to simulate shelf-life self-discharge. Passing cells must typically retain at least 70% of their charge going into the recharge step, and recover to at least 85% of rated capacity once fully recharged.

500-Cycle Endurance: Continuous CC-CV charge and discharge cycles at 20 ± 5°C, with a typical pass threshold of at least 60% residual capacity after 500 complete cycles. Run at the standard 0.2C rate this single test takes roughly 5,000 hours (around seven calendar months), which is why accelerated cycling protocols are often substituted and flagged as such on the test report.

AC Internal Resistance: Internal resistance measured using a 1 kHz AC method, verified against the manufacturer’s declared specification.

Nomenclature and Marking: Cell coding conventions (e.g. INR18650) and dimensional tolerances are verified for cross-brand interchangeability.

Automotive Traction: IEC 62660-1

IEC 62660-1 defines cell characterisation procedures for battery electric vehicle (BEV) and hybrid electric vehicle (HEV) applications. BEV capacity testing uses the full depth of discharge (DoD) to maximise range characterisation, whereas HEV protocols focus on partial charge windows that reflect hybrid usage patterns.

SAE J2288 uses the USABC Dynamic Stress Test (DST) profile, which incorporates regenerative braking, to model realistic highway and urban driving loads on the cell.

Stationary Systems: IEEE 450 and IEEE 1188

Stationary lead-acid installations in telecommunications and power utilities are governed by IEEE 450 (vented lead-acid) and IEEE 1188 (VRLA). These standards define maintenance schedules, capacity test frequencies, and test execution procedures, covered in detail in the Stationary Battery Testing section below.

Formation vs Lifecycle Capacity Testing

A capacity measurement taken during initial electrochemical formation and one taken during long-term lifecycle testing are fundamentally different operations, with different objectives and different test parameters.

During formation, the slow charge and discharge currents allow lithium ions to intercalate gradually, minimising mechanical stress and preventing early active material loss. In lifecycle testing, the pre-stabilised cell is subjected to accelerated conditions — often combined with elevated temperature — to quantify calendar and cycle life decay rates.

Capacity Tracking: State of Charge Mathematics

Coulomb Counting

State of charge is defined as the ratio of remaining releasable capacity to the maximum actual capacity under equivalent health and environmental conditions:

SoC(t) = SoC(0) – (1 / Q_max) x ∫ η_c x I(t) dt

Where I(t) is cell current in amperes (positive for discharge, negative for charge), Q_max is the maximum capacity in Ah, and η_c is the coulombic efficiency.

In embedded battery management systems, this continuous integral must be discretised. The simplest approach is left-hand Euler rectangular integration:

SoC[k] = SoC[k-1] – (Δt / Q_max) x I[k-1]

Where Δt is the sampling interval in hours and I[k-1] is the current sample at step k-1. This method accumulates discretisation error under dynamic loads with rapid current transitions.

Trapezoidal integration reduces this error by averaging consecutive current samples:

SoC[k] = SoC[k-1] – (Δt / Q_max) x (I[k-1] + I[k]) / 2

This acts as a low-pass digital filter on the current signal, reducing drift during regenerative braking spikes or high-rate pulse events without significant additional computational load.

Error Sources in Coulomb Counting

Pure current integration is susceptible to unbounded drift from four main sources:

Current sensor bias: Small measurement offsets integrate linearly over time, producing progressive SoC error. A 1 mA offset in a 100 Ah cell accumulates 0.024 Ah per day of continuous operation.

Integration approximation: Rectangular summation misses transient current variations, particularly under pulsed or dynamic loads.

Capacity uncertainty: Using a fixed nominal capacity value ignores cell aging, causing systematic SoC overestimation as the cell degrades.

Clock drift: Jitter in the integration timer introduces error in the effective Δt, particularly relevant for low-power microcontroller implementations with relaxed oscillator specifications.

Modern battery front-end (BFE) ICs address current bias using dedicated sigma-delta integrator ADCs. SoC drift correction requires closed-loop estimation — typically an Extended Kalman Filter — that uses real-time terminal voltage feedback against an OCV-SoC lookup table to bound the accumulated error.

Capacity Retention and Degradation Metrics

For lifecycle testing, three metrics are tracked at each capacity checkpoint:

Capacity retention: CR[n] = C[n] / C[0]

Capacity fade rate: ΔC[n] = (C[n-1] – C[n]) / C[n-1]

Coulombic efficiency: CE[n] = Q_discharge[n] / Q_charge[n]

For calendar aging, empirical capacity decay follows:

C(t) = C_0 x e^(-k x t)

Where k is a temperature-dependent decay rate governed by the Arrhenius equation:

k = A x e^(-Ea / RT)

Where A is a pre-exponential factor, Ea is the activation energy (typically 0.5 to 0.6 eV for LFP cells), R is the universal gas constant, and T is absolute temperature in Kelvin. This relationship explains why calendar life approximately doubles for every 10°C reduction in storage temperature.

Lead-Acid Temperature Correction (IEEE Standards)

IEEE capacity measurements for lead-acid systems must be corrected to a 25°C baseline. Time-adjusted capacity:

C_adj = C_actual x (T_actual / T_rated) x K_T

Where T_actual is the measured discharge duration, T_rated is the manufacturer’s rated time, and K_T is the temperature correction factor. Rate-adjusted current:

I_adj = I_rated x K_T

Chemistry Performance Envelopes

Capacity measurement results are only meaningful in the context of chemistry-specific operating boundaries. Testing outside these envelopes does not just produce degraded results — it triggers distinct failure modes that permanently alter the cell.

Qualification Temperature Ranges

• EV traction systems: pack-level operating and storage envelopes are commonly specified from around -40°C to +60°C in OEM requirements, though this is a system-level design target rather than a fixed test condition. The IEC 62660-1 cell-level performance test itself runs at a narrower 23 ± 2°C standard condition, with -10°C or -20°C as the optional low-temperature points.
• Consumer electronics: typically -20°C to +60°C
• Defence and aerospace: down to -55°C

Low and High Temperature Failure Modes

Below 0°C, slow diffusion kinetics cause lithium ions to deposit as metallic lithium on the anode surface rather than intercalating into the graphite. This lithium plating causes permanent capacity loss and elevated internal resistance. It is a particular risk during charging at low temperature, not just discharge.

Above 45°C, accelerated SEI growth consumes active lithium inventory and increases internal resistance. Sustained operation above 60°C risks separator degradation and thermal runaway onset.

Chemistry Comparison

Depth of Discharge and Cycle Life

DoD has a non-linear effect on cycle life. A typical NMC cylindrical cell:

These figures are indicative for modern automotive-grade NMC cylindrical cells at 25°C and 1C/1C charge-discharge. Results vary significantly with cell design, temperature, C-rate, and the SoC window within which the DoD is applied — a 20% DoD window centred at 90% SoC ages the cell faster than the same window centred at 50% SoC.

Shallow cycles reduce the volumetric expansion and contraction of the electrode host structure. Deep cycling drives particle cracking, active material loss, and accelerated capacity fade. State-of-health thresholds in automotive applications are typically set at 80% retained capacity precisely because the degradation rate accelerates below that point.

Incremental Capacity Analysis and Differential Voltage Analysis

Incremental Capacity Analysis (ICA) and Differential Voltage Analysis (DVA) are non-destructive diagnostic methods that extract internal degradation information from charge-discharge voltage curves.

ICA computes dQ/dV from the voltage-capacity data. Flat voltage plateaus — characteristic of LFP — produce sharp, distinct peaks that shift or reduce in amplitude as the cell ages. DVA computes the inverse (dV/dQ) and is more useful for chemistries with sloped discharge profiles.

Four degradation modes can be isolated from peak behaviour:

Loss of Lithium Inventory (LLI): Irreversible consumption of cycleable lithium through SEI growth, electrolyte decomposition, or lithium plating. Reduces the total area under the IC curve and shifts peaks along the voltage axis.

Loss of Active Material at the Anode (LAM_NE): Graphite exfoliation, particle cracking, or pore blocking. Shrinks the specific peaks corresponding to anode-phase transitions.

Loss of Active Material at the Cathode (LAM_PE): Structural collapse or transition metal dissolution in the positive electrode. Reduces amplitude of cathode-phase change peaks.

Resistance Increase (RI): Elevated internal resistance increases overpotentials, causing IC peaks to shift horizontally, broaden, and decrease in height.

Practical Limitations

Mathematical differentiation amplifies measurement noise. A 200-point moving average filter is typically required to produce clean peaks from real cell data. To achieve distinguishable peaks from a vehicle battery, cells must be charged at a very low C-rate (C/25 or lower) under stable temperature — in practice this means a vehicle parked in a temperature-controlled workshop before the diagnostic charge. These requirements make ICA/DVA impractical for in-service monitoring without purpose-built algorithms and dedicated hardware.

Stationary Battery Discharge Testing

Stationary lead-acid installations in telecommunications and power utilities operate under IEEE 450 (vented lead-acid) and IEEE 1188 (VRLA). Both standards define structured maintenance schedules and capacity test frequencies.

Maintenance Schedule

• Every 4 months: verify station DC supply voltages, electrolyte levels, and ground fault status
• Every 18 months: measure individual cell float voltages, internal ohmic values, inter-cell connection resistances, and check physical integrity of battery racks and connections

Capacity Test Frequency

• Baseline capacity test within the first 2 years of service
• Subsequent tests at 25% intervals of the battery’s expected service life
• Annual testing once capacity drops below 90% of rated value, or when the battery reaches 85% of its expected service life

Pre-Test Charging Requirements

IEEE 450 and IEEE 1188 point the test engineer to the battery manufacturer’s specified equalising voltage rather than fixing one value in the standard itself; in practice this commonly falls in the 2.33-2.40 V per cell range for vented lead-acid, applied for 72 hours or longer, followed by a continuous float charge of at least 72 hours immediately prior to discharge.

Discharge Procedure and Cell Bypass Protocol

Standard capacity tests run at constant current for 1 to 8 hours, with a target cell voltage cut-off typically around 1.75 V per cell, though the exact figure is manufacturer-specified and rate-dependent. Individual cell voltages may drop prematurely during the test. Under IEEE guidelines the test should not be interrupted unless a cell approaches reverse polarity, conventionally treated as 1.0 V per cell or below for both VRLA (IEEE 1188) and vented lead-acid (IEEE 450) strings. At that point:

1. Pause the discharge
2. Bypass the failing cell using a jumper cable of adequate current capacity
3. Resume discharge within 6 minutes of pausing

Only one bypass is permitted per test. A second cell failure requires the test to be terminated and the battery bank assessed for replacement.

Modified Performance Test

The Modified Performance Test verifies that the battery bank can meet the actual site duty cycle, not just a constant-current discharge.

Type 1 (High Initial Load Step): Apply a high initial current step for time T1 to simulate critical emergency starting loads, then transition to a rate-adjusted constant current for the remainder of the test.

Type 2 (Complex Dynamic Sizing): Apply high-current spikes that may exceed the manufacturer’s sustained current rating for the duration required by the duty cycle. The capacity removed during spike events is subtracted from 80% of rated capacity, and the remaining capacity divided by remaining test time determines the constant current rate for the rest of the test.

Instrumentation: Four-Wire Kelvin Sensing

Measuring the low internal impedance of high-capacity cells requires four-wire Kelvin connections to prevent lead resistance and contact resistance corrupting the terminal voltage measurement.

In a standard two-wire setup, the test leads carry both the charge/discharge current and the voltage sense signal. The voltage drop across the lead resistance appears in the measurement as apparent cell voltage, inflating internal resistance readings.

The Kelvin method separates these functions:

Force leads: A low-impedance current path connects the tester’s source/sink to the outer boundary of the battery terminals.

Sense leads: A separate pair of voltage-sensing leads connects directly at the cell terminals, routed to a high-impedance internal voltmeter.

Because the voltmeter input impedance is very high, the current flowing through the sense leads is negligible. The voltage drop across the sense cables is therefore effectively zero, and the voltmeter reads only the true terminal voltage of the cell.

Four-wire Kelvin sensing is mandatory for DC internal resistance (DC-IR) pulse measurements and for Electrochemical Impedance Spectroscopy (EIS). For standard capacity discharge measurements on large-format cells, the error from two-wire sensing may be acceptable at low C-rates, but at 1C and above on low-impedance pouch or prismatic cells the error becomes significant.

Test Chamber Safety Engineering

Electrical capacity testing of lithium-ion cells carries risk of electrolyte leakage, toxic gas emission, and thermal runaway. Testing operations classify failure severity using EUCAR hazard levels:

Chamber Design Requirements

Sheath heaters with surface temperature limits: Standard heating elements can exceed 200°C surface temperature, well above the auto-ignition point of typical lithium-ion venting gases. Safety-configured chambers use heaters with surface temperatures capped below the auto-ignition threshold of volatile organic gases released during venting.

Spark mitigation: Intrinsically safe electrical barriers must be applied to all internal sensor paths. Fans must use non-sparking blades. Internal lighting must be removed.

Gas detection and automated ventilation: Continuous monitoring for H2, CO, and HF. On detection, the system automatically opens inlet and exhaust dampers to dilute flammable mixtures below their lower explosive limit.

Inert gas flooding: Nitrogen or CO2 purging systems displace oxygen and suppress flame propagation if gas detection thresholds are exceeded or a rapid thermal escalation rate is detected. Liquid nitrogen surface cooling can also be deployed to arrest thermal runaway propagation between adjacent cells.

Pressure relief ports: Gravity-controlled ports (typically 6 to 8 inches in diameter) vent positive pressure events without requiring manual intervention or replacement after operation.

Safety door interlocks: Electromagnetic locks prevent personnel from opening the chamber during active tests or immediately after a safety event, maintaining isolation until venting and atmosphere normalisation are complete.

Related Pages

• Cell Capacity Test Procedure
• Cell Testing
• Incremental Capacity Analysis
• Kelvin Resistance Measurement
• Lifetime
• IEC 61960-3
• IEC 62660-1
• SAE J2288

See also: Recovery Capacity, Battery Capacity