An empty drink can is strong until you dent it

How shape rather than material carries an external pressure load.

An empty drink can is strong until you dent it because its strength comes almost entirely from its shape holding a smooth, unbroken curve, and very little from the thickness or toughness of the aluminium making it up. The instant that curve is interrupted by even a small dent, the can loses the geometric trick that let it carry weight in the first place and collapses under a load it supported easily a moment before.

The first pressure housing I sized was checked against the strength of the metal, which was the wrong calculation, since a cylinder loaded from outside cares far more about whether its round shape can be pushed a fraction of a millimetre out of true than about how much force the material could withstand in a straight pull.

A wall as thin as paper holding up a person

An unopened can stood upright on a table can support an adult standing on it carefully, even though its aluminium wall is roughly a tenth of a millimetre thick, about the thickness of a sheet of paper, on a can some six hundred times wider than that. A smooth cylindrical wall carries a downward load by spreading it evenly all the way round the circumference, so that no single point of the metal ever bears more than its own tiny share. That arrangement is extraordinarily efficient, which is why so little material holds so much weight, but it depends on every point of the wall staying exactly where the round shape says it should be.

The moment a dent pushes one small region out of that shape, the load stops spreading evenly and concentrates on the distorted region instead, which folds further under the very load it used to help carry, and the whole can collapses in an instant. The aluminium itself is unchanged in that instant, with no crack opened and no material torn. The wall has simply stopped doing the one thing its shape had been doing for it, and the load has nowhere left to go.

One thumbprint, and a housing under water

Press a thumb gently into the side of an empty can, then stand the same weight on it, and the dented can collapses at a fraction of the weight the undented one shrugged off, even though not one gram of aluminium was removed. A submerged cylindrical housing faces the same test at every point on its surface at once, with water pressing inward from every direction. A housing that is round to a tight tolerance resists that pressure the way the undented can resists a standing weight, sharing the load evenly around its circumference, while a housing with even a small out-of-round region concentrates load there exactly as the dent does.

A thin-walled cylinder can fail this way, buckling suddenly into a collapsed, non-round shape, at a pressure far below what would be needed to crush the same metal in a straight compression test, because buckling is a loss of stability. The shape loses its ability to hold its own geometry long before the material is asked to carry anything close to its rated capacity. That gap is why a pressure housing cannot be sized the way a plain tension or compression member is, by checking only that the material's cylinder stress stays below an allowable limit.

Roundness written on the drawing

Because the failure is governed by shape, roundness becomes a specification worth stating and checking on a pressure housing in a way it never needs to be for a bracket or a shaft. A housing machined from a strong alloy but held to a loose roundness tolerance can fail at a lower pressure than one made from a weaker alloy held to a tighter one. That reordering, shape ahead of raw material strength, is the governing idea behind everything else in this set. A designer who has only ever sized parts against yield or tensile strength brings a well-practised habit to the problem, and that habit is exactly what has to be unlearned, since checking the material's strength on a housing that is about to buckle answers a question nobody asked.

Stocky walls that crush instead

The shape argument dominates for thin-walled cylinders, the geometry most housings use, and weakens as the wall grows thick relative to the diameter. A sufficiently stocky cylinder runs out of material strength before it becomes unstable enough to buckle, the same way a short, fat pillar crushes under a compressive load instead of bowing sideways. Most housings are kept well away from that regime, since a wall thick enough to fail by crushing is usually far heavier than the job requires, which is why buckling remains the calculation that governs almost every real housing built to work underwater. The rest of this set follows that thread, first separating stability failures from strength failures in general terms, then asking what property of a material resists buckling once shape has been accepted as the thing that matters most.