Worked example: pre-charge stalls at 312 V
The precharge-fail demo read step by step: an ISO 15118-20 DC session that passes every step up to DC_PreCharge, then 40 pre-charge requests while the charger output stays at 312 V of 400 V, until the car ends the session.
The capture
- Load the demo Pre-charge stalls (precharge-fail): 11.2 s of an ISO 15118-20 DC session over plain TCP, as test benches record it with TLS switched off.
- Verdict: Never charged, Pre-charge failed, likely party EVSE, confidence medium.
- Stage strip: SLAC, SDP, TCP / TLS, Handshake, Setup, and Service and schedule passed; Energy transfer not reached; Stop passed, because the car ended the session in order.
- The Capture tool grades the file good: no missing bytes, no retransmissions, and VT-02 finds all 100 messages decoded.
Up to the pre-charge
- EV: supportedAppProtocolReq, 1.700 s: ISO 15118-20 DC first, ISO 15118-2 second
- EVSE: supportedAppProtocolRes, ISO 15118-20 DC selected
- EVSE: AuthorizationSetupRes, EIM only, no certificate installation
- EVSE: AuthorizationRes, OK, Finished at the first answer
- EVSE: ServiceDiscoveryRes, DC and DC_BPT offered
- EV: ServiceSelectionReq, DC, parameter set 1
- EV: DC_ChargeParameterDiscoveryReq, 200 to 900 V, up to 400 A and 150 kW, TargetSOC 80
- EVSE: DC_ChargeParameterDiscoveryRes, 150 to 1000 V, up to 500 A and 350 kW
- EV: ScheduleExchangeReq, Scheduled, DepartureTime 7200 s, EVTargetEnergyRequest 40 kWh
- EVSE: DC_CableCheckRes, 2.760 s: OK, Finished at the first answer
HS-03 warns twice: the car offered, and the charger selected, ISO 15118-20 over plain TCP. That is the bench setup, not the fault the verdict is about.
The pre-charge loop
- EV: DC_PreChargeReq, 2.850 s: EVProcessing Ongoing, EVPresentVoltage 380 V, EVTargetVoltage 400 V
- EVSE: DC_PreChargeRes, 40 ms later: OK, EVSEPresentVoltage 312 V
- both: 38 more pairs, one request every 200 ms, every value unchanged
- EV: DC_PreChargeReq, 10.650 s: the 40th request
- EVSE: DC_PreChargeRes, 10.690 s: still 312 V
- EV: SessionStopReq, 10.950 s: ChargingSession Terminate
- EVSE: SessionStopRes, 11.010 s: OK
- both: TCP connection closed, 11.150 s and 11.180 s
- Every response is OK. No response code will tell you anything here; the fault is in a value.
- The charger answers quickly, 40 ms per pre-charge response and 80 ms at the slowest in the session (TIM-01). This is not a slow charger but an output that does not move.
- EVSEPresentVoltage is 312 V in all 40 responses. It does not creep up.
- The target is inside what the charger announced: 400 V lies between its EVSEMinimumVoltage of 150 V and EVSEMaximumVoltage of 1000 V, and the Negotiation tool finds no conflict between the two sets of limits.
- The car never sends PowerDeliveryReq. It keeps its contactors open while the voltages differ and ends the session about 8 s into the loop.
What DC-01 says
DC-01 reads the last DC_PreChargeRes: after 7.8 s and 40 PreCharge requests the charger reported 312 V against the 400 V the car asked for, a difference of 88 V where the profile allows 20 V, and no PowerDelivery Start followed. The loop also ran longer than the 7 s pre-charge timeout of the profile. The output never got close, so the check names the charger.
dc.prechargeToleranceV (20 V) and dc.prechargeTimeoutS (7 s) are flagged for verification in the limit profile. The car's own pre-charge timer is not in the capture; here the car stopped 260 ms after the last response.
Confirming the cause
- Measure the DC output during pre-charge, with the charger's own log or a meter at the connector. There are two outcomes: the output really stops at 312 V, or it reaches 400 V and the charger reports a value that is not live.
- If it really stops, look for what caps it below the target: a power module range or a voltage limit set lower than the range announced in DC_ChargeParameterDiscoveryRes.
- If the value is stale, the charger reports a number it does not update. A value that does not change at all over 40 responses is worth checking against a meter.
- Try another car with a similar battery voltage. If it stalls at the same 312 V, the car is not involved.
What a fix looks like
- Make the output reach any target inside the range the charger announces, or announce only the range it can reach. EVSEMinimumVoltage and EVSEMaximumVoltage are a promise.
- If the reported value was stale, fix the measurement path so that EVSEPresentVoltage is a live reading; the car relies on it before it closes its contactors.
- After the fix, expect EVSEPresentVoltage to climb to within the tolerance of 400 V over a few responses, the car to report EVProcessing Finished, and PowerDeliveryReq with ChargeProgress Start to follow.
- For a charger in the field, switch TLS 1.3 on; HS-03 then passes. DC session stalls in CableCheck or PreCharge covers the other ways CableCheck and PreCharge stall.
Sources
- ISO 15118-20:2022, DC_CableCheck, DC_PreCharge and PowerDelivery
- IEC 61851-23, pre-charge of a DC charging station
- SPEC.md of this tool, FR-ENC-05, the failing demo sessions