How to Measure DCIR Correctly

Define the test before collecting values

Record cell identity, history, SOC method, temperature, stabilization, rest, pulse direction, amplitude, duration, voltage windows, sample rate, filtering, fixture and calculation.

An acceptance limit belongs to that frozen procedure. A changed fixture, timing or SOC method requires correlation before the same limit is reused.

Sense voltage at the intended boundary

Use separate current-force and voltage-sense conductors. Place the sense points at the terminals whose resistance is intended to be included.

Cell-terminal, module-terminal and pack-output measurements answer different questions. Label the boundary in the result.

References: Keysight: How to Test Coin Cell Batteries

Use synchronized current and voltage

Use discharge-positive current: discharge I > 0, charge I < 0. Define t0 from the measured transition, not just the command. State the pre-transition baseline and synchronized voltage/current sample windows.

The numerator is a voltage drop, Vbaseline − V(t). For a normal charge pulse both numerator and denominator are negative, giving positive resistance. A near-zero current step makes the ratio unusable.

Record the trigger rule, rise time, window width, sample rate and filtering. End-of-pulse samples must precede the falling edge. Early readings can be unresolved; longer pulses include polarization and state changes.

References: Comparison of Several Methods for Determining the Internal Resistance of Lithium Ion Cells

Reject common measurement shortcuts

Do not use two-wire voltage, unspecified tester-reported timing, unmatched temperatures or mixed charge/discharge direction. Do not average an unstable fixture into a precise-looking number.

Report uncertainty and only the digits supported by the equipment and repeatability study.

Worked discharge and charge examples

Illustrative data, not test results or recommended settings: assume 50% SOC, stabilized 25°C and prescribed rest. Define t0 at the measured step midpoint, baseline averaging from −100 to −10 ms, and loaded windows at 90–100 ms or 990–1000 ms. Verify the tester settles before the windows.

At 1 s, the discharge example gives 0.250 V / 50.0 A = 5.00 mΩ. The separate charge example uses its own prepared state and baseline.

Carry uncertainty through the ratio

Illustrative independent standard uncertainties u(ΔV) = 2 mV and u(ΔI) = 0.2 A give u(R) ≈ 0.045 mΩ at 0.250 V and 50.0 A. With k = 2 the expanded estimate is approximately 0.09 mΩ.

The resulting 5.00 ± 0.09 mΩ (k = 2) applies only to that assumed budget. Include actual correlations, timing, temperature, SOC preparation and fixture repeatability. Uncertainty of a difference is not automatically uncertainty of one reading.

References: NIST: Metrology and process control, dealing with measurement uncertainty

Set abort limits before applying current

Use qualified personnel, a rated protected fixture and cell-specific voltage, current and temperature limits. Define abort conditions and safe isolation in the approved test plan.

Stop on a breached limit, unstable connection, unexpected heating, swelling, leakage or equipment fault. A clipped, tapered or interrupted pulse is not a full-amplitude acceptance sample. Do not override protection to complete the test.

Measurement & validation procedure

Continue this engineering task

Related engineering guidance