The Pain Point: The Gasket That Weeps in Winter

End tank leaks are the most common radiator warranty claim and the least satisfying to investigate. The tank itself is intact, the core is intact, and yet there is a film of coolant along the crimp line after the first cold snap. The assembly passed a pressure test at 20 °C in the plant. It fails at −15 °C in a customer's driveway.

The reason is that the joint was qualified at one temperature and one pressure and is used at another. A crimped joint relies on a compressed gasket that must maintain sealing force across a temperature range of more than a hundred degrees, against two materials with different expansion coefficients, while the tank itself expands and contracts at a different rate from the header it is clamped to. When the gasket loses compression, the joint does not fail catastrophically; it weeps.

The end tank is where all the difficult requirements of a radiator meet: a pressure boundary, a sealing interface, several ports, mounting points and a brazed or welded connection to the core.

The Principle: Sealing Is a Force Balance, Not a Material Property

A gasket seals by maintaining contact stress. The rubber or elastomer must stay compressed between two surfaces through the full operating envelope. Sealing force comes from the crimp or the bolt preload, and it is consumed by thermal expansion, creep of the elastomer and relaxation of the joint.

Thermal expansion works against you. An aluminium header and a glass-filled nylon tank expand at very different rates. Over a 100 °C swing, the differential movement at the joint can reach several tenths of a millimetre. The gasket must absorb that movement without losing contact stress, which is why gasket geometry - not just hardness - determines success.

Ports and vents create local stress. Every inlet neck, outlet neck, drain cock boss and bleed vent interrupts the wall. In a cast or machined end cover, the material around a port must carry the pressure load that the removed material no longer carries. A generous boss with a fillet is not cosmetic; it is the difference between a pressure vessel and a crack initiation site.

Buffer volumes change the pressure history. Some designs include a buffer or expansion chamber in the end cover to damp pressure pulsations from the water pump. This is a genuine functional feature, but it also creates an internal volume where air can collect and where the wall must resist pressure cycling without support from the core.

Seals need somewhere to sit. A sealing ring works only if the groove that holds it has the right depth, the right surface finish and the right lead-in chamfer. The same applies to a dust ring or a wiper on a sliding interface: if the groove is machined with a sharp edge or a rough floor, the seal will cut itself or leak past its own seat.

Guiding features carry load that the seal should not. Where a component includes a guide sleeve or a spigot, its purpose is to take transverse load and keep alignment so that the seal is never asked to act as a bearing. The classic materials for such sleeves are sintered oil-retaining alloys and bronze castings, chosen because they tolerate sliding contact and retain lubricant. In an aluminium housing, that insert must be a separate part: aluminium against aluminium in sliding contact will gall.

Equipment Classification: End Tank and Cover Types

Crimped glass-filled nylon tanks on a brazed aluminium header. The volume leader in passenger cars. Light, cheap, and entirely dependent on gasket compression and crimp quality.

Welded aluminium end tanks. Used where pressure or temperature exceeds what a gasket can hold. Welded tanks remove the gasket but introduce heat-affected zones and distortion.

Cast aluminium end covers and caps. Common on heavy-duty radiators, oil coolers and charge air coolers. Die casting replaced malleable iron specifically to cut weight and eliminate rust, at the cost of porosity risk.

Machined billet covers and manifolds. Low volume, high pressure, and often the route for prototype and motorsport hardware.

Small brass covers on miniature units. Where the size makes die casting uneconomic and where the thermal load is modest, brass remains the practical choice even in an aluminium system - with all the galvanic consequences that implies.

Practical Guide: Getting the Joint Right

Specify gasket compression as a range, not a target. State the installed compression, the minimum retained compression after ageing, and the test method. A gasket that is over-compressed takes a set and never recovers.

Design the crimp for uniform load. Crimp tabs spaced unevenly produce a joint that seals at the tabs and weeps between them. Model the load distribution and check it on a sectioned sample.

Control the header flange flatness. A wavy flange cannot be sealed by any gasket. Specify flatness over the full sealing line, not just locally.

Put radii on every port transition. A 2 mm radius at the base of an inlet neck can reduce the local stress concentration by a factor of two compared with a sharp corner.

Provide a proper seal groove specification. Depth tolerance, floor finish, side wall draft and lead-in chamfer all belong on the drawing. A seal groove drawn as a rectangle without dimensions is a leak waiting to happen.

Isolate dissimilar materials. Where a brass fitting or a steel stud enters an aluminium tank, use a sealant or a coating that prevents galvanic contact, and remember that the thread itself is a crevice.

Test at temperature, not just pressure. A pressure test at ambient proves the joint geometry. A thermal cycling test followed by a pressure test proves the joint design.

Leave the tank able to breathe during filling. A vent or bleed port is not optional on a system that traps air. Air pockets cause local hot spots and accelerate corrosion from the inside.

Field Experience: The Cast Cover That Cracked at the Boss

A heavy-duty radiator used a die-cast aluminium end cover with an integral inlet neck and a drain boss on the same wall. Units began cracking at the base of the drain boss after about 3,000 operating hours, always on the same side, always starting at the same corner.

Two factors combined. The casting had a wall thickness change from 3.5 mm at the boss to 2.2 mm in the surrounding panel, with no taper, so solidification left a shrinkage void just under the surface at that transition. And the drain boss had been positioned directly opposite the inlet, so pressure pulsations from the pump struck it at the point of highest stress.

The fix had three parts: the wall transition was tapered and radii were added, the boss was moved 40 mm downstream to a region of lower pulsation, and the casting was changed from a gravity-fed die to a vacuum-assisted process which reduced porosity. Field failures stopped within one production quarter.

The interesting detail was that the section that cracked had passed a burst test at three times operating pressure. Static strength was never the issue. Fatigue under pulsation at a porosity defect was.

Summary

Aluminium parts for radiators are judged at their end tanks. The core determines how well the unit cools; the tank determines whether it survives. Sealing is a force balance across a temperature range, ports are stress concentrations, and seals only work when the groove that holds them is specified as carefully as the surface it seals against. Every one of these is a drawing decision, made long before the first part is cast.

Outlook

The direction is toward higher system pressures and more aggressive thermal cycling, both driven by tighter packaging and by electrification. That will push more designs from crimped polymer tanks toward welded or cast aluminium, and it will make porosity control a qualifying capability rather than a quality bonus. Expect also to see seals specified with documented compression-set data and end covers validated by thermal cycling rather than by burst pressure alone - because the failures that matter here are the ones that appear at 40,000 kilometres, not at the end of the line.