Why Battery Voltage Sags Under Load

Start with a bounded first-order calculation

Illustrative calculation: assume a rested 3.30 V cell and 4 mΩ effective resistance for the declared pulse condition. A 50 A discharge gives a 0.20 V drop and an estimated 3.10 V terminal voltage.

The 4 mΩ value is not a universal cell constant. Its meaning depends on pulse duration, current, SOC, temperature, history and voltage-sampling definition.

Separate fast loss from developing polarization

Fast ionic and electronic conduction combines with developing charge-transfer and transport polarization. These contributions overlap; they are not sequential switches and have no universal time boundaries.

Published pulse studies show that calculated resistance changes with the selected sampling time. A short transient and a multi-second peak therefore need different evidence.

References: A study of the influence of measurement timescale on internal resistance characterisation methodologies for lithium-ion cells; Comparison of Several Methods for Determining the Internal Resistance of Lithium Ion Cells

Check the complete pack voltage path

Busbars, welds, fuses, contactors, connectors, cables and current shunts add series loss. At 200 A, one additional milliohm produces 0.20 V drop and 40 W of resistive heat.

The system can cross its undervoltage threshold while measurable capacity remains. Diagnose voltage simultaneously at the cell stack, pack output and equipment input rather than treating recovered open-circuit voltage as proof of adequate power.

Recovery after load removal is not restored capacity

Removing a discharge load eliminates the ohmic I·R drop. Interfacial polarization and concentration gradients then relax over time, so the terminal voltage continues to recover after the current has stopped. An immediate unloaded reading is a recovering open-circuit voltage, not necessarily equilibrium OCV.

The rebound does not replace withdrawn charge. Record rest duration and preceding current history; recovered voltage alone cannot establish capability for the next peak.

References: A study of the influence of measurement timescale on internal resistance characterisation methodologies for lithium-ion cells

Check the operating window, not just room temperature

At lower SOC, reduced OCV can reduce voltage headroom even without an increase in resistance. Evaluate the actual cell OCV curve and pulse response together.

The cited automotive pouch-cell study covers three types over 10–90% SOC and −10 to +40°C. It supports condition-dependent impedance, not a universal temperature multiplier or safe operating limit.

Measure minimum loaded voltage at required SOC and temperature combinations. Stabilize cell temperature and record heating during repeated peaks.

References: Parameterization of linear equivalent circuit models over wide temperature and SOC spans for automotive lithium-ion cells using electrochemical impedance spectroscopy

Capacity and peak power need separate acceptance tests

Illustrative comparison: two aged cells retain 90% of their initial low-rate capacity. Different pulse resistances can still produce different loaded voltages under the same current.

Track usable charge and pulse response independently, with aging history and a comparable baseline. Resistance alone does not identify a unique degradation mechanism.

References: Molicel INR-21700-P45B product data sheet, revision 1.2

A cutoff depends on threshold, delay and recovery logic

Illustrative example: 3.20 V before load minus 0.40 V sag gives 2.80 V, below an assumed 2.90 V undervoltage threshold. Disconnection depends on the configured delay, filtering and logic; it is not necessarily immediate.

Check the weakest cell and equipment input as well as pack voltage. Record cell traces, fault flags and event timing to distinguish protection operation from depleted capacity.

Do not lower protection thresholds to conceal sag. Investigate load demand, cell condition and connection losses within cell and equipment limits.

References: Texas Instruments BQ76952 Technical Reference Manual, SLUUBY2B

Measure the boundary that the equipment uses

Kelvin sensing separates voltage measurement from current-carrying leads. Identify the exact sense points: a contact inside them remains part of the measured resistance.

Record cell identity, SOC method, temperature, rest, current direction and amplitude, sample windows, sense points and uncertainty. Capture stack, pack-output and equipment-input voltage together to localize loss.

References: Keysight: How to Test Coin Cell Batteries

Finish with the repeated duty profile

Include peak spacing, base load, the low-SOC endpoint and temperature rise. Constant-power equipment may draw increasing current as terminal voltage falls.

Use HPPC to map pulse capability across operating points, then validate the complete required duty including cabling, protection and thermal limits.

References: DOE/INL: Battery Test Manual for Plug-In Hybrid Electric Vehicles, revision 3 (September 2014)

Measurement & validation procedure

Continue this engineering task

Related engineering guidance