Diagnostic Test 04 — Internal Resistance (Impedance) Test
| Mode number | 4 |
| Registered handler | internal_resistance_test → _run_test_04_internal_resistance in ems_diagnostic_handlers.py |
| Model actions | change_power (pulse), direct Modbus reads (cell voltage + DC current) |
| Parameters | diagnostic_test_mode.internal_resistance_test |
Theory
Every battery has an internal resistance — the opposition to current flow caused
by the electrolyte, electrodes, current collectors, terminals and internal
connections. When the battery delivers or absorbs current I, its terminal
voltage sags or rises by ΔV = I × R. Internal resistance therefore shows up as
a voltage drop proportional to current.
As a battery ages, internal resistance increases: electrodes degrade, electrolyte conductivity falls and connections loosen. A high internal resistance reduces usable power, increases heating, and is a strong early indicator of cell degradation or a loose busbar/terminal. It is measured with a short, controlled current pulse:
R = |V_rest - V_pulse| / |I_pulse - I_rest|
V_rest and I_rest are measured just before the pulse; V_pulse and
I_pulse are measured while the pulse is applied. The pulse is brief and
moderate so it perturbs the battery without charging or discharging it
significantly.
This test:
- applies a defined DC current pulse (discharge or charge),
- reads pack voltage (sum of the cell-voltage registers) and DC current,
- computes pack resistance in mΩ and derives the average cell resistance
(
pack resistance / cell count, since the cells are in series), - compares the results against configured limits and against pulse-to-pulse repeatability.
Settings needed
All values live under diagnostic_test_mode.internal_resistance_test in
emsConfig/ems/ems_control_setting.json. The nested object is canonical; flat
keys are a legacy fallback; every key also has a built-in default.
Safety / eligibility
| Key | Default | Meaning |
|---|---|---|
min_soc |
20 |
Minimum SOC (%) required before the pulse can run |
max_soc |
80 |
Maximum SOC (%) allowed before the pulse can run |
min_pack_voltage_v |
700.0 |
Minimum baseline pack voltage (V) allowed |
max_pack_voltage_v |
900.0 |
Maximum baseline pack voltage (V) allowed |
max_cell_temp_c |
45.0 |
Reserved maximum cell temperature (C), enforced by the BMS safety layer |
The test aborts with prerequisite_failed if SOC or baseline pack voltage is
outside its window.
Pulse generation
| Key | Default | Meaning |
|---|---|---|
pulse_current_a |
50 |
Target current step magnitude (A) |
pulse_direction |
discharge |
Pulse sign: discharge (positive power) or charge (negative power) |
pulse_duration_sec |
10 |
Total time the pulse is held (seconds) |
settle_time_sec |
5 |
Stabilisation delay after applying the pulse and before measuring (seconds) |
rest_time_sec |
5 |
Recovery delay between pulses (seconds) |
number_of_pulses |
3 |
Number of pulse cycles to run |
The pulse power is derived as pulse_current_a × baseline_pack_voltage / 1000
(kW), signed by pulse_direction. The PCS is always commanded back to 0 kW
after the measurement, including on every failure path.
Pass / fail thresholds
| Key | Default | Meaning |
|---|---|---|
max_pack_resistance_mohm |
50.0 |
Maximum allowed pack resistance (mΩ) |
max_cell_resistance_mohm |
10.0 |
Maximum allowed average cell resistance (mΩ) |
max_resistance_deviation_pct |
20.0 |
Maximum pulse-to-pulse spread of pack resistance (%) |
Average cell resistance = pack_resistance_mohm / cell_voltage_count. The test
passes only when every pulse for every PCS stays within all three limits.
Register map (site-specific)
These addresses and scale factors must match the actual BMS/PCS Modbus map.
| Key | Default | Meaning |
|---|---|---|
cell_voltage_start_register |
1860 |
First holding register of the cell-voltage window |
cell_voltage_count |
8 |
Number of cell-voltage registers to read |
cell_voltage_scale_mv |
1 |
Scale applied to each raw cell-voltage word (mV) |
dc_current_register |
1801 |
Holding register for DC current |
dc_current_scale_a |
0.1 |
Scale applied to the raw signed-16-bit current word (A) |
Pack voltage is computed as
sum(cell_voltage registers) × cell_voltage_scale_mv / 1000 volts. DC current is
decoded as a signed 16-bit value then multiplied by dc_current_scale_a.
Example configuration:
"diagnostic_test_mode": {
"internal_resistance_test": {
"min_soc": 20,
"max_soc": 80,
"pulse_current_a": 50,
"pulse_duration_sec": 10,
"pulse_direction": "discharge",
"settle_time_sec": 5,
"rest_time_sec": 5,
"number_of_pulses": 3,
"max_pack_resistance_mohm": 50.0,
"max_cell_resistance_mohm": 10.0,
"max_resistance_deviation_pct": 20.0,
"min_pack_voltage_v": 700.0,
"max_pack_voltage_v": 900.0,
"max_cell_temp_c": 45.0,
"cell_voltage_start_register": 1860,
"cell_voltage_count": 8,
"cell_voltage_scale_mv": 1,
"dc_current_register": 1801,
"dc_current_scale_a": 0.1
}
}
Model action required in emsConfig/pcs/power_sequences/<MODEL>.json:
change_power— used to apply the pulse power and to return to0 kW.
Prerequisites
Before the test runs:
- At least one active PCS is configured and Modbus-reachable.
- Every available PCS returns a valid SOC, and every SOC lies within
[min_soc, max_soc]. - Every available PCS returns valid cell-voltage and DC-current register reads,
and its baseline pack voltage lies within
[min_pack_voltage_v, max_pack_voltage_v]. - The PCS model must define the
change_poweraction. - Every register required by the test —
read_soc,change_power, the configured cell-voltage range and the DC-current register — exists in the PCS device register map (emsConfig/pcs/<pcs>/pcs_register_full.json, falling back toemsConfig/pcs/pcs_register.json). A missing register aborts the test and is reported inregister_checks. pymodbusis installed.
Testing procedure
- Phase 1 — Prerequisite check. Confirm at least one PCS is available, then
verify every register required by the test (SOC, power setpoint, cell voltage
and DC current) is defined in the device register map. A missing device or
missing register aborts the test as
prerequisite_failed. - Phase 2 — SOC safety window. Read SOC from every PCS; abort if any value
is outside
[min_soc, max_soc]. - Phase 3 — Baseline measurement. Read pack voltage and DC current from
every PCS; abort if any pack voltage is outside
[min_pack_voltage_v, max_pack_voltage_v]. - Phase 4 — Pulse loop. For each pulse:
- Command the pulse power (derived from
pulse_current_aand the baseline voltage, signed bypulse_direction). - Wait
settle_time_secfor the current to stabilise. - Read pack voltage and DC current.
- Hold the remaining
pulse_duration_sec - settle_time_sec(if positive). - Command
0 kWand waitrest_time_sec. - Compute
R = |V_rest - V_pulse| / |I_pulse - I_rest|per PCS. - Always command
0 kWat the very end, even if a step failed. - Phase 5 — Evaluation. Compare pack resistance, cell resistance and pulse-to-pulse deviation against the configured limits.
Report output
The report (measurements) contains:
pulse_direction,pulse_current_a,pulse_power_kw,number_of_pulses.max_pack_resistance_mohm,max_cell_resistance_mohm,max_resistance_deviation_pct.max_pack_resistance_measured_mohm,max_cell_resistance_measured_mohm,resistance_deviation_pct.results— per-PCS, per-pulse record withv_rest_v,i_rest_a,v_pulse_v,i_pulse_a,delta_v_v,delta_i_a,pack_resistance_mohm,cell_resistance_mohm.
Graphs:
ir-pack-resistance— bar chart of average pack resistance (mΩ) per PCS.ir-voltage-current— line chart of pack voltage and DC current vs elapsed time (baseline + pulse samples).
Pass / fail summary
| Result | Condition |
|---|---|
passed |
Every pulse for every PCS within all three limits and sequence completed |
failed |
A resistance limit or the deviation limit was exceeded |
error |
A read/command failed or the pulse sequence did not complete |
prerequisite_failed |
No PCS, SOC outside window, baseline voltage outside window, or a required register missing from the register map |
Register-map caution. The default register addresses (
1860,1801) and scale factors are placeholders. Confirm them against the real BMS/PCS Modbus map before the first run, otherwise the measured resistance will be wrong.