Isolation monitoring and CableCheck

Why a DC charger runs as an unearthed IT system, how its insulation monitor tests the cable before charging and watches the whole circuit during charging, and why the test sometimes fails for no real fault.

An IT system on the DC side

The transformer in each DC/DC module separates the output of a DC charger from the mains. Neither DC+ nor DC- is tied to protective earth, which makes the output an IT system: a single insulation fault between one pole and the car body or the charger enclosure does not by itself drive a dangerous current, because there is no return path. A second fault on the other pole would. The job of the insulation monitoring device (IMD) is to find the first fault before a second one can appear.

The car's high voltage system is built the same way: floating with respect to the chassis and monitored by the car's own isolation monitor. Once the car closes its contactors, the charger output, the cable and the car's battery form one IT system.

Before and during charging

  • Before charging (CableCheck): with the car's contactors still open, the charger raises its output to a test voltage and measures the insulation of its own output, the cable and the car inlet up to the contactors. A fault here is found before the battery is connected.
  • During charging: the IMD keeps measuring. Now it sees the whole circuit, car included, so a fault in the car's high voltage system can also trip it.
  • Two active insulation monitors on one IT system disturb each other, because each injects its own measuring signal. Cars therefore commonly pause their own monitor while the charger's is in charge of the circuit.

The test voltage used for CableCheck and the rules for which monitor is active when differ between editions of IEC 61851-23 and between vehicle makers. Check them against the edition you test to.

CHAdeMO and GB/T do the same test outside ISO 15118: in CHAdeMO the charger locks the connector and tests insulation before the car closes its contactors; in GB/T 27930 the charger tests after the handshake, using the maximum voltage the car announced in BHM.

Thresholds in ohm per volt

Insulation limits are given relative to the working voltage, so the same rule covers a 400 V and an 800 V system. Multiply by the voltage to get the resistance.

LimitPer voltAt 500 VAt 1 000 V
Charger: Valid above500 ohm/V250 kilohm500 kilohm
Charger: Warning between100 to 500 ohm/V50 to 250 kilohm100 to 500 kilohm
Charger: Fault at or below100 ohm/V50 kilohm100 kilohm
Vehicle DC circuits (UN ECE R100)100 ohm/V minimum50 kilohm100 kilohm
Vehicle AC circuits (UN ECE R100)500 ohm/V minimum250 kilohm500 kilohm

The charger thresholds (500 ohm/V for a warning, 100 ohm/V for a fault) are the commonly quoted values of IEC 61851-23 and are to be checked against the edition you test to. The vehicle values are those of UN ECE R100.

What a capture shows

  1. EV: CableCheckReq, repeats while the charger measures
  2. EVSE: CableCheckRes, EVSEProcessing Ongoing, EVSEIsolationStatus Invalid, EVSE_IsolationMonitoringActive
  3. EV: CableCheckReq
  4. EVSE: CableCheckRes, EVSEProcessing Finished, EVSEIsolationStatus Valid
  5. EV: PreChargeReq, the output was discharged after the test and ramps up again

In DIN SPEC 70121 and ISO 15118-2 the result is the EVSEIsolationStatus field of DC_EVSEStatus: Invalid (no result yet), Valid, Warning (low but permitted, charging may go on), Fault (the charger must not deliver energy) and, in ISO 15118-2 only, No_IMD. The field is repeated in every DC response, so a drop from Valid to Warning during CurrentDemand is visible in the capture. The meaning of each value is in the code dictionary under isolation.

ISO 15118-20 drops the field: DC_CableCheckRes carries only the response code and EVSEProcessing, so a failed test shows as a FAILED response code or a loop that never finishes. The iso20 demo has a single DC_CableCheckRes that is already Finished.

The length of the CableCheck loop is the IMD's measuring time plus the time to charge and discharge the output. A few seconds is common; long runs near the car's CableCheck timeout point to an IMD struggling with capacitance or leakage.

The CableCheck timeout of the car is commonly quoted as 38 s for DIN SPEC 70121 and 40 s for ISO 15118-2. It is not in the limit profile of this tool yet and is to be checked against the timing tables.

False positives and real faults

CauseWhy it tripsTypical pattern
Y capacitors in the car and chargerEMC filter capacitors between the high voltage poles and chassis must be charged by the measuring signal; a large total capacitance slows the measurement and can upset some IMDsLong CableCheck, or a warning only with certain car models
Humidity and condensationMoisture in the connector, the car inlet or a damaged cable jacket lowers the surface resistanceFaults in the morning, in rain, or after the cable lay in water; gone after drying
Damaged cableA crushed or cut cable has a real leakage pathFault on every car at that charger
Fault in the carLeakage in the battery, heater or drive inverterWarning or Fault only after the car closes its contactors, on every charger
Symmetric faultEqual leakage from both poles; some measuring methods see it poorlyReported late or only by the car's monitor

The quickest way to tell the parties apart is the one in the code dictionary: a fault during CableCheck lies in the charger or cable, a fault that appears only once the car is connected lies in the car, and a fault that follows one car across chargers belongs to that car.

Sources

  • IEC 61851-23, insulation monitoring and cable check of DC charging stations
  • IEC 61557-8, insulation monitoring devices for IT systems
  • UN ECE Regulation No. 100, isolation resistance of the vehicle high voltage system
  • ISO 15118-2:2014 and DIN SPEC 70121:2014-12, isolationLevelType, CableCheck