Sourcing Gaskets and Oil Seals From China: Compression, Hardness and the Leak That Appears at 40,000 km

Seals are the cheapest component on an engine and the most common cause of a warranty claim. I have watched a supplier swap a gasket material from an FKM compound to an FKM-sounding compound to save ¥0.6 a piece, and the resulting failure showed up eight months later as an oil weep on the timing cover of several thousand units. Nobody in the factory did anything wrong on the line. The material was simply not what the drawing said.

Here is what actually determines whether a gasket or seal survives.

Compression, not just thickness

A gasket works because it is squeezed. The useful parameter is compression stop — the amount of squeeze the design allows — and for most fibre or beaded metal gaskets the working range is 20 to 30 percent of free thickness. Less than that and you have not loaded the material enough to fill the surface imperfections. More and you have over-stressed it, and it will take a permanent set and lose its ability to follow thermal cycling.

On a stamped steel or MLS (multi-layer steel) head gasket, compression is controlled by the bead geometry and the coating, and the number that matters is the installed thickness under specified bolt load. Ask for the load-deflection curve, not the free thickness. Free thickness is the number every catalogue quotes because it is easy to measure and means nothing.

Torque matters here in a way that catches people out. Gasket compression follows bolt preload, and bolt preload is set by torque spec — but on a short bolt clamping a soft gasket, most of the torque goes into friction. Use angle-controlled tightening on the last pass, or at minimum specify a thread lubricant and check the resulting clamp load on first articles. A gasket that leaks only on the units assembled on the second shift is a torque-control problem, not a gasket problem.

Rubber hardness is a moving target

Oil seals and O-rings are specified at a Shore A hardness, usually 70 to 75 for a lip seal on a rotating shaft. The number on the drawing is measured at room temperature on a conditioned sample. The number that matters is what the material does at 120 °C in hot oil for a thousand hours.

Two things to insist on. First, hardness change after heat ageing — a good FKM compound should shift by less than 10 points after 70 hours at 200 °C in air per the standard ageing test. Second, compression set, which is the percentage of thickness the material fails to recover after being squeezed hot and released. For a seal in a hot application, compression set below 20 to 25 percent is the target. A compound that hits hardness but misses compression set will seal for a year and then weep as the lip loses its grip.

Material substitution is the other trap. NBR, HNBR, FKM, ACM and silicone cover most of what you will see. They are not interchangeable. FKM resists oil and heat but is poor in hot water and steam. HNBR costs a lot more than NBR but holds up far better at 150 °C in oil. NBR is fine at 90 to 100 °C and dissolves quietly at 130 °C. If you are buying for a Chinese engine platform, ask which specification the compound is against — ASTM D2000 line callout, or an OEM material spec — and get a batch-level material certificate, not a generic one that has been photocopied since 2018.

Surface finish is half the seal

A lip seal runs on a shaft surface, and that surface has a roughness and a direction. Roughness above roughly Ra 0.8 µm wears the lip. Below Ra 0.2 µm the surface is too smooth to retain the micro-oil film and the lip runs dry and burns. The window is narrow and it is a machining parameter, so it belongs on the shaft drawing, not in the supplier's hands.

The direction matters too. Plunge grinding leaves a spiral lead that can pump oil along the shaft. A lead-free surface — usually achieved by a final belt or superfinish pass — is what good seal suppliers will ask for. If your seal supplier has never mentioned shaft finish, they have never debugged a leak.

The housing bore matters just as much. Press-fit seals need a bore roundness and finish that lets the OD hold without crushing the case. If the bore is out of round, the seal will follow the bore, not the shaft, and one side of the lip will lift. Roughness there should be around Ra 1.6 to 3.2 µm; too smooth and the seal walks out under pressure.

What I put on the drawing and the PO

Material by ASTM D2000 line callout plus any OEM spec, hardness with tolerance, compression set limit after a defined ageing cycle, low-temperature flexibility (a seal that survives 150 °C but goes brittle at minus 30 will fail in a Chinese winter or a cold-chain shipment), the mating surface finish on both shaft and bore, and the installation lubricant. Then a batch COA with actual measured values, not just pass/fail ticks.

One more thing, and it saves real money: ask for the seals to be supplied in a protective sleeve and demand the installation tooling drawing. Most lip seal failures in service were actually installation damage — a lip rolled on a shaft keyway or a burr — and that is a process problem you fix once at the assembly line rather than forever at the warranty desk.