Last Wednesday afternoon, July 22nd, my phone buzzed with a WhatsApp voice note from a workshop owner in Thessaloniki, Greece. His name is Dimitris, he runs a small EV-focused service shop that mostly handles BYD ATTO 3 and Zeekr 001, and the audio was 47 seconds long. The short version: a 2023 Zeekr 001 long-range came in with a battery warning message on the dash, no DTC code, no drive power. He had plugged in his OBD, scanned every module, and everything came back clean. The 12V battery was brand new. The HV contactors were clicking on and off as expected. He was about to order a full HV battery pack — a 4,800 euro mistake — when something in the wiring harness caught his eye.
What Dimitris found was a hairline crack in the HV wiring harness insulation, right at the bend where the harness transitions from the rear-drive motor conduit to the body underbody. The copper was fine, the connector was fine, but the insulation had chafed against the subframe brace over 60,000 km of thermal cycling. Every time the pack pushed current through that segment, a tiny arc would leak to chassis ground. The BMS would see inconsistent voltage at the pack output, throw the warning, kill drive power to protect the pack — and Dimitris would see no DTC because the arc was below the BMS detection threshold.
He sent me a photo. I sent him back the Zeekr 001 high-voltage wiring harness part 8893734472, a torque spec sheet, and a 30-minute video on how to drop the rear motor conduit to swap the segment. The total cost of the fix: 280 euro for the harness segment plus 90 euro for the diagnostic time. The battery pack he almost ordered: 4,800 euro. The lesson, and the reason I am writing this article: high-voltage wiring harness faults are some of the most under-diagnosed failures on Chinese EVs, and the symptoms almost always look like a battery problem until you put your hands on the harness itself.
I have been sourcing EV parts out of Shanghai since 2013. We ship to 47 countries. Every week I see three or four cases that look exactly like Dimitris — a 12V that tests fine, a BMS that scans clean, and a vehicle that mysteriously will not drive. Below is what we have learned about HV harness faults across the BYD, Zeekr, Nio, Xpeng, and VW MEB-platform cars we ship parts for. If you are a distributor, a workshop, or a fleet buyer, this is the technical depth you want before you start ordering the wrong components.
Why HV harness failures look like pack failures
The high-voltage system on a modern Chinese EV is a closed loop. The pack output goes through the HV junction box, splits into the motor inverter, the OBC, the DC-DC converter, the PTC heater, the eAC compressor, and back to the pack negative. Every segment of that loop is monitored for current and voltage drop at the BMS level. The BMS does not measure insulation resistance directly — it infers it from voltage symmetry across the two HV bus rails.
When the insulation on a harness segment chafes, cracks, or absorbs moisture, the effect is not a short circuit. It is a slow, asymmetric leakage. The pack positive rail sees a slightly lower voltage than expected because some current is bleeding to ground. The BMS logs the imbalance, throws a generic high-voltage system fault warning, and as a safety measure locks the contactors open. Drive power disappears. The 12V system is unaffected, so the dashboard stays alive. There is no DTC because no DTC threshold was crossed — the leakage is below the BMS fault threshold but above its warning threshold.
This is the failure shape that costs workshops the most money. The instinct is to suspect the pack first because the warning message is on the pack icon. The instinct is wrong. Across our 2024 to 2026 cases, the breakdown has been roughly:
40 percent — HV harness insulation damage (chafe, moisture, rodent)
25 percent — Connector interlock loop faults (HVIL open circuit)
15 percent — OBC or DC-DC isolation faults (internal to the component, not the harness)
10 percent — Contactor welding or coil failure
10 percent — Genuine pack cell imbalance or module failure
If you skip the harness and order a pack, you have an 85 percent chance of being wrong on a 4,800 euro gamble.
The four root causes we see most often
Across the , the , and the BYD ATTO 3 e-platform 3.0 cars, four root causes account for most harness failures.
- Chafe at subframe transitions
The HV harness is routed through the underbody alongside the fuel-equivalent brake lines. Where the harness transitions from the rigid battery conduit to the flexible motor conduit, there is usually a rubber grommet and a foam wrap. Over 50,000 to 80,000 km, the foam wrap compresses, the grommet hardens, and the harness starts to oscillate against the subframe brace during accelerator pedal lift events. The result is a 5 to 15 mm chafe zone on the outer insulation.
The fix is to inspect the underbody at every 40,000 km service interval. If you see wear-through to the inner insulation layer (the colored coating under the orange HV jacket), replace the harness segment. We stock the and the as separate line items because they fail on different schedules — the harness at 60 to 80k km, the ground cable typically earlier (40 to 60k) due to higher current density.
- Connector interlock loop (HVIL) faults
Every HV connector on a modern Chinese EV has a low-voltage interlock loop that runs through the connector body. When the connector is properly mated, the HVIL loop is closed. When the connector is partially unmated, dropped, or has a corroded pin, the HVIL opens. The BMS detects the open interlock, opens the contactors, and refuses to close them again until the interlock is restored.
HVIL faults are the second most common cause of no drive, no DTC cases. The diagnostic shortcut is to listen for the HVIL click when you mate the connector — there should be a clear secondary detent click after the main connector seats. If you feel only the main click, the interlock pin is not engaged. The repair is usually to replace the connector housing (not the entire harness), but you need to source the correct connector family — there are at least four different HV connector families used across the Zeekr 001, Nio ET5, and BYD e-platform 3.0 platforms.
- Moisture ingress at the AC charging socket
When CP and PP signal pins corrode, the OBC cannot complete the handshake with the EVSE. The car throws a charging fault, but if the customer only charges at home on a 7kW wallbox and the corrosion is intermittent, the fault may only appear once a month. By the time the customer brings the car in, the connector body is showing green corrosion on three pins. The fix is to replace the socket harness and reseal the body grommet with fresh butyl rubber — not silicone, silicone degrades the existing EPDM boot.
- The 12V battery that looks fine
The last root cause is not technically an HV harness fault, but it presents identically. A weak 12V battery will let the BMS initialization sequence fail halfway. The HV contactors will not close, the BMS will throw a generic high-voltage warning, and the car will not start. The 12V tests at 12.4V, looks fine on a basic multimeter, but cannot sustain the 80A inrush that the DC-DC converter demands during pre-charge.
The diagnostic shortcut here is a load test, not a voltage test. We have seen customers replace a perfectly good HV battery pack because the workshop skipped the 12V load test and went straight to HV warning equals HV problem. The is a 60Ah AGM unit, but the same failure pattern shows up on VW ID.4, BYD Dolphin, and most Xpeng models.
What a proper harness replacement looks like
For workshops and distributors who want to do this right, the procedure on a Zeekr 001 is roughly 3 to 4 hours for the rear segment, 6 to 8 hours for a full underbody run. The non-negotiable steps are below.
Lock out the HV system. Ignition off, key fob 5+ meters from the car, wait 10 minutes for the contactors to de-energize. Measure at the HV service disconnect with a CAT III 1000V meter. Both rails should read 0V to chassis ground before you touch anything.
Document the routing before disassembly. Photograph the harness path from at least four angles. The factory routing is not always obvious — there are three foam dampers and two grommets per segment. If you forget the routing pattern, the replacement harness will sit 2 to 3 mm off the body and chafe twice as fast.
Replace the grommets and foam wrap with the harness. Used foam wrap has compressed. Reusing it gives you 30 to 50 percent of the original vibration damping. New wrap costs 8 euro and adds two years to the harness life.
Torque the HV connector bolts to spec. On the Zeekr 001 the HV connector bolts are M8 at 12 Nm. The interlock pins must click into the secondary detent — confirm with a 0.5 mm feeler gauge between the connector body halves.
Run the BMS initialization sequence via the diagnostic tool, not by ignition cycling. The BMS needs to see the HVIL closed, the contactors commanded closed, the pre-charge current within 2 percent of expected, and the isolation resistance above 100 kΩ per volt before it will authorize drive. Without the tool you can cycle ignition 20 times and the car will still say system check failed.
The and the sit on a parallel HV branch and have their own interlock pins. On the Zeekr 009 the equivalent parts are the and the . The procedure is similar but the connector family is different, and the wrong family will seat mechanically but fail the HVIL check.
Distributor-side QC: what to inspect on incoming HV harness
If you are a distributor receiving HV harness shipments for resale, the failure modes are different from workshop-side. The two issues we see most often on incoming goods:
Counterfeit connector bodies. The HV connector shells look identical to the OEM but the HVIL pin is a dummy — it seats, but it does not complete the interlock loop. The harness will work in a bench test (because the high-current path is intact) but the car will refuse to enable drive. Always pull one connector per shipment and confirm the HVIL continuity with a multimeter between the interlock pin and the connector body.
Stiff or cracked insulation jacket. The orange HV jacket has a shelf life. If the harness has been stored in direct sunlight for more than 6 months, the jacket can stiffen and crack during cold-weather installation. The crack is invisible from outside until the harness is flexed. Bend-test every harness segment through a 30-degree arc before shipping.
A 200 euro bench test kit (12V supply, multimeter, insulation tester) catches both issues. A 5,000 euro diagnostic session on a customer car does not.
The bottom line
If a car comes into your shop with an HV warning, no DTC, and a 12V that tests fine — before you quote a battery pack, pull the underbody trim and inspect the HV harness at every grommet and subframe transition. The fix is usually a 200 to 400 euro part and 3 to 4 hours of labor, not a 4,800 euro pack and a week of downtime. The four root causes above — chafe, HVIL faults, moisture at the AC socket, and a weak 12V — account for the vast majority of these cases we see on Chinese EVs.
If you are sourcing parts, stock the high-failure items as separate SKUs: the rear HV harness segment, the main ground cable, the AC socket harness, and the 12V battery. They fail on different schedules, and bundling them into one HV system kit makes your inventory look more comprehensive but leaves customers waiting for a part you do not actually stock.
And if a customer in Thessaloniki, Casablanca, or Dubai sends you a photo of a warning message on the dash and asks whether to replace the battery — ask them to send a photo of the underbody harness first. It is the cheapest diagnostic in the EV repair business, and Dimitris’s 280 euro fix is the kind of outcome that brings a customer back for ten years.
— Lao Zhou, founder of JLL Auto Parts · 13 years sourcing EV parts from China · jllautoparts.com