Isolation Fault That Only Appears When It Rains
A 2020 electric hatchback comes in with an intermittent HV isolation warning lamp. The owner reports it appears after rain or a car wash and usually clears by the next day. The scan tool shows a stored isolation DTC with freeze-frame data captured at the event: isolation resistance 78 Ω/V at a pack voltage of 388 V. Live isolation value in the dry bay is 2,300 Ω/V. The car has an aftermarket dash camera hard-wired to a fuse-box tap and aftermarket LED headlamps. FMVSS 305 requires at least 100 Ω/V for DC high-voltage circuits. The owner wants the code cleared so the car passes inspection tomorrow.
1. Interpret the freeze-frame data. Is this a real defect or a nuisance code?
Read the worked answer
Real defect. The numbers settle it.
- 78 Ω/V during the event is below the 100 Ω/V minimum. At 388 V pack voltage that is roughly 30 kΩ of isolation resistance across the whole HV system - a genuine leakage path to chassis, well into the range where a person contacting chassis and a live conductor could be injured.
- 2,300 Ω/V when dry is fine, and that is precisely why this is dangerous rather than reassuring: the vehicle is compliant in the bay and non-compliant in the rain. A dry-bay pass proves nothing.
- The correlation with water is a diagnostic gift, not a coincidence. Water ingress into an HV connector, the charge port, the PTC heater, the A/C compressor, the motor terminal area, or a chafed harness section is the classic mechanism.
Clearing the code does not change any of this. It removes the evidence, leaves the leakage path in place, and sends a car out that fails the legal isolation requirement whenever it gets wet.
2. How would you localize the fault, and what does 'segment testing' actually mean here?
Read the worked answer
Split the HV system and measure each piece, in a defined order, with the system disabled and re-verified at every step.
- Follow the OEM isolation-fault location procedure first. Many manufacturers publish a sequence and a specified tester voltage for each component. Deviating risks damaging electronics and risks a wrong answer.
- Isolate loads one at a time - charge port, cabin PTC heater, A/C compressor, DC-DC converter, motor/inverter assembly, then the pack and its internal harness. Measure insulation resistance from the live conductor to chassis on each isolated segment and compare to the OEM limit.
- Do not megger the pack itself unless the OEM procedure says to. A tester applied to a battery pack can stress the BMS and the modules.
- Chase the water. Look for standing water or dampness in connector boots, a missing or pinched seal, a cracked charge-port housing, a plugged body drain, and harness chafe at wheel-well and rocker routing. Spray-test the suspect area with the OEM-approved method and watch the live isolation value on the scan tool - a value that falls as you spray is your answer.
- Aftermarket accessories: the dash cam is a 12 V tap and unlikely to cause this, but check that the installer did not pierce an HV harness or displace a body grommet. Same for the lamps and any non-factory wiring in the front clip.
3. The owner needs the car tomorrow. What do you tell them, and what would a defensible release look like?
Read the worked answer
Tell them the car fails a federal isolation requirement when wet and that clearing the code does not repair it. It is not released until the leak is found and the repair is verified.
A defensible release looks like this:
- A located and corrected root cause - a resealed connector, a replaced charge-port housing, a repaired harness - not a cleared code.
- Post-repair verification that reproduces the original condition: water test or wet-road drive with the scan tool recording, isolation value staying above the 100 Ω/V minimum throughout, and no returning DTC after a full drive cycle.
- Documented pre- and post-repair isolation values, the freeze-frame data, and the fact that the original condition was reproduced and eliminated.
A cleared code with no repair hands the customer a car that may strand them in a rainstorm and puts the shop's name on a known safety defect. If the customer declines diagnosis, that refusal belongs in writing, and the vehicle still does not leave with the fault documented as repaired.