Weatherstripping, Seals and Rubber for Classic Cars

The root cause is almost always the reference part. Nobody has factory tooling, so a shop sections a physical sample and cuts a die from it — and used seals are what survive. Every defect the sample acquired over forty years goes into the tooling: decades under compression leave permanent set, and hardened rubber has shrunk in section and length. The result measures like the reference and does not seal.

Shrinkage is the second layer. Rubber contracts as it cools from cure temperature; allowances commonly fall in the region of one to three percent, EPDM often toward the lower end. Tool makers compensate by cutting the cavity oversize — a tool cut from a used sample with no allowance produces undersized parts, and the error accumulates across a whole opening.

Third is construction. An extrusion has no corners, and most weatherstrip needs them; the honest ways are a moulded corner joined to extruded legs, or a fully moulded loop. Mitring and gluing an extrusion is the shortcut, and it shows.

The commonest specification failure is asking for a durometer on a part that does not have one. Dense — solid — rubber is measured on the Shore A scale under ASTM D2240, and dense automotive weatherstrip commonly sits in the region of 60 to 80 Shore A, stated with a tolerance so a supplier can verify it.

Sponge rubber is different: a cellular seal has no meaningful Shore A hardness, because the gauge indents the cell structure rather than the polymer. ASTM D1056 applies instead, classifying flexible cellular rubber by type (1 open-cell, 2 closed-cell), by class for fluid and temperature resistance, and by grade 0 through 5, ranking it by the force needed to compress it 25 percent. That grade decides whether your door shuts, and co-extruded seals need both halves specified.

A request naming a shape and a polymer gets a price that means nothing. These fields make quotes comparable.

Rubber is one of the few categories where a sample abused for an afternoon beats any correspondence.

ISO 2230, the guideline for storing rubber products, sorts polymers into groups with an initial storage period plus an extension permitted after inspection: group A (natural rubber, SBR, polyurethane) five years plus two; group B (nitrile, chloroprene) seven plus three; group C (EPDM, silicone, fluorocarbon) ten plus five, shorter group governing blends. Conditions matter as much as the clock: below 25 °C, humidity under roughly 70 percent, out of sunlight, away from ozone-generating equipment, stored relaxed and loosely coiled rather than folded. A length that arrives kinked never fully recovers.

Three cases recur, and in each the honest answer is to say so rather than spend tooling money. When no unused reference survives anywhere, tooling cut from a used seal encodes that seal's decay, and the best available reproduction is an approximation that should be sold as one. When the original was a complex one-piece moulding — a windshield gasket with integral corners — the mould is expensive and the surviving car count will never repay it. And when the body itself has moved, a car with replaced sills or a re-hung door will not accept a correct seal; that needs bodywork, not parts.

One more limit worth naming is our own. If you are restoring a single car, this is not a sourcing exercise: a set of seals from an established marque supplier costs less than the freight on a pilot order. Overseas sourcing makes sense when you need the same profile repeatedly and hold an unused reference length nobody currently tools. If that is you, describe the part in plain English on our request page: vehicle, years, position, construction, and what reference you hold. Taiwan Supplier Hub returns a shortlist of at least three verified manufacturers that have confirmed they want the project within 14 days, US$99 at launch, refunded if we can't deliver.