Formation and grading data are measurements before they become cell judgments. Channel accuracy, probe contact, range, temperature, and time synchronization can create false capacity, resistance, and curve anomalies.
How to calibrate channels
Cover the voltage and current ranges actually used, not one point
Formation cyclers commonly use multiple current ranges. Calibration only at a high-current point can leave large errors during low-rate formation, small constant-voltage tail current, and rest voltage. Cover the product voltage window, each critical current range, and both current directions with traceable meters or shunts, retaining standard value, channel reading, error, environment, and before/after adjustment.
Beyond voltage and current, confirm sampling time, step transitions, and channel clocks. Capacity integrates current over time, so timestamp or missing-data faults directly change the result. Bind calibration to equipment, module, channel, and software/firmware state instead of marking only the whole cabinet as passed.
| Calibration item | Coverage | Record |
|---|---|---|
| Voltage | Near lower, middle, and upper product voltages | Standard, reading, error, polarity |
| Current | Critical ranges and charge/discharge direction | Range, shunt, error, settling time |
| Time/sampling | Long/short steps, transitions, interruptions | Timestamp, interval, gaps, recovery |
Verify before every production start
Standards, contact checks, and channel swap tests quickly isolate channel bias
Before production, inspect probe contamination, spring force, wiring, polarity, and temperature control. A stable standard or reference cell can quickly compare channel voltage, drop, and current response. It detects inter-channel differences but does not replace formal calibration.
When one curve is abnormal and safety permits, run a channel swap with a qualified cell of the same type. A fault that follows the channel points first to equipment, contact, or temperature; one that follows the cell points to welding, filling, resistance, or material. Keep protocol and temperature fixed and retain the relationship between original and retest positions.
- —Use references for quick comparison only
- —Never overwrite the original with a swap test
- —Check temperature and contact for positional faults
How to read a curve anomaly
Mark the step, duration, and accompanying heat before assigning a cause
Do not compare final capacity alone. Align cells by protocol step and review starting OCV, constant-current voltage slope, constant-voltage duration and cutoff current, relaxation, first-cycle Coulombic efficiency, temperature rise, thickness/pressure, and gas. Sudden jumps often involve contact, wiring, or sampling. Persistent high voltage can involve contact drop, poor wetting, resistance, or temperature. Low efficiency with heat or swelling requires safety quarantine and checks of moisture, electrolyte, and side reactions.
First see whether anomalies cluster by channel, tray position, assembly station, electrode section, or material lot. Equipment faults often cluster by channel or position; process faults by weld, fill, or roll section; material shifts more often move the lot distribution. This order reduces the risk of rejecting a cell lot for an equipment fault.
| Curve pattern | Equipment first | Then cell/process |
|---|---|---|
| Instant voltage jump or gap | Probe, wiring, sampling, network | Protection event or true soft short |
| Persistently high voltage in one step | Contact drop, channel current, temperature | Wetting, weld, compaction, material |
| Low capacity with similar shape | Current range, time integration, cutoff | Balance, active loading, first-cycle loss |
| Low efficiency with heat/swelling | Sensor and positional temperature bias | Moisture, electrolyte, side reaction, micro-short |
Preserving raw data
Keep raw curves, protocol ID, ranges, and every retest relationship
For every cell, retain unsmoothed raw voltage, current, capacity, time, and temperature histories with equipment, module, channel, tray position, protocol version, range switches, alarms, and interruptions. Derived slopes, constant-voltage time, dQ/dV, or anomaly scores do not replace the raw curve.
A retest creates a new record linked to the original; it never overwrites the first result. Calibration, probe replacement, software upgrade, and repair also form a timeline, showing whether a fault disappears after maintenance or follows the cell lot. Automatic release should block missing curves, mismatched barcodes, unknown protocol IDs, and unexplained interruptions.
Specified accuracy does not guarantee long-term channel accuracy on the floor. Set calibration intervals, quick checks, and diagnostic rules from equipment stability, duty, and product risk.
Bibliography
References
- 01Data-driven approaches for the fast formation of lithium-ion pouch cells using operando gassing analysis
- 02Development of a Self-Learning Process for Determining Quality in Battery Cell Production Based on Optical Thickness Measurement
- 03Simulation of the Production of Lithium-Ion Cells and Battery Packs
- 04NEWARE CT/CE-4000 Series and CT-9000 Ultra-High Precision Testing Series
- 05How to Evaluate Battery Test Equipment
Updated: 2026-08-26