High-Temp O-Ring Solution for Under-Hood Connector Seals: An HNBR Case Study

The problem: a connector that only failed when the engine was hot

Under-hood connectors sit in one of the least forgiving sealing environments in a vehicle. The seal works a few centimetres from a heat source cycling between ambient and well over 150 °C, it is washed in oil vapour and fuel permeation, and it is clamped between two moulded housings whose grooves were designed before the seal was specified.

This case looks at an under-hood wiring harness connector program originally specified with a standard NBR O-ring at 70 Shore A. New parts passed bench leak tests. Field trials came back with two symptoms: a slow weep at the connector interface after sustained high-temperature running, and a permanent set that made re-mating hard on the assembly line.

Separating three failure modes

• Thermal ageing — the compound hardening and losing recovery near the heat source.
• Compression set — permanent deformation in a shallow groove, so sealing force collapses after each heat cycle.
• Fluid attack — swelling or extraction in oil vapour, which changes hardness and groove fill.

Only the first looks like a temperature problem. The field failures were driven by the second and third, which is why a higher temperature rating alone did not stop the leak.

Material decision

HNBR was the only candidate holding all three requirements at once: heat ageing close to FKM, oil and fuel vapour resistance better than NBR, and enough cold flexibility for winter assembly. Hardness stayed at 70 Shore A, so groove fill did not change.

Groove and size work

1. Groove fill — HNBR swells less than NBR in the same media, so installed volume stayed inside the safe fill window and no housing rework was needed.
2. Cross-section tolerance — the groove is shallow, so a few hundredths of a millimetre of cord variation changes squeeze. Cords were held to a tight band rather than a general tolerance.
3. Size standard — the program stayed on AS568 dash numbers and metric cross-sections, so seals came from existing tooling instead of new moulds.

Validation that mattered more than a data sheet

• Heat ageing with hardness change and tensile retention recorded.
• Compression set at actual service temperature, not at a comfortable lower one.
• Oil and fuel vapour immersion with volume swell and hardness tracked, confirming the groove fill assumption.
• Low-temperature recovery, to protect assembly force in cold plants.
• 100% inspection of production batches with dimensional reports per lot — human and machine, not sampling.

Why this did not need a new mould

• More than 15,000 moulds in the tool room, covering O-rings and custom parts, including the complete AS568 range from -001 to -932.
• No tooling fee on any size already in mould.
• Not in mould? Two routes: pay the one-time mould cost and start small, or reach USD 2,000 on that size and the mould cost is waived on request.
• Minimum order value USD 100–120 per size by material, not per piece — so five sizes can be qualified in one order instead of committing volume to one dash number.
• Quotation per size, material, hardness and colour within eight working hours.

For a validation program that needs five sizes at low volume, that structure removes the usual barrier: no mould investment has to be justified before the seal proves itself in a heat cycle.

What changed

HNBR at the same hardness and the same groove dimensions addressed both field symptoms without touching the housing design. Re-mating force came back inside the target band, and the weep at the connector interface stopped appearing after sustained high-temperature running.

The transferable lesson: specify against the failure mode, not against the temperature rating. A seal rated for the heat but taking a permanent set in a shallow groove will still leak; a seal that resists oil but hardens in the cold will still stop your line.

Sealing an under-hood connector?

Send the groove drawing, media list, service temperature range and target hardness. We will tell you which candidate fits, whether the size is already in mould, and what the per-size minimum looks like — usually within eight working hours. If we cannot make your size, we will say so in the same eight hours instead of leaving you waiting.

Ask for the size list too. Our published list covers more than 5,400 sizes, but it is never the whole story — we open new moulds every month.