← Back to Archive

Why roundness is a structural property

Where manufacturing tolerance enters a strength calculation.

Roundness is a structural property because a cylinder's ability to carry external pressure by compression alone depends on every point of its wall lying on a true, unbroken circle, and any departure from that circle, however small, gives the load a place to concentrate. How round a part actually is therefore decides how much pressure it survives, alongside how thick and how strong it is.

How a slightly oval wall starts to bend

A perfectly round cylinder loaded from outside carries the load the way an arch does, in pure compression running around its circumference, each point of the wall pressing evenly against its neighbours. Once the cross-section drifts from a true circle, even by an amount too small to see, that compression stops being even. A slightly oval section starts to bend at the points where its curvature is least, and the bending grows worse under load, since a flatter region deflects further under the same pressure than the rounder wall beside it, pulling more of the load toward itself. Left unchecked, that feedback ends in the sudden collapse described earlier in this set.

The imperfection driving it is rarely chosen. It is usually an ordinary trace of how the part was made: slight spring-back after machining, uneven clamping during welding, or thermal movement as the part cooled.

A tall stack of coins shows the same sensitivity. It balances easily while every coin sits squarely on the one below, the weight running straight down through the middle, but let one coin drift a millimetre off centre and the stack becomes far more likely to topple, though no coin got weaker or lighter. A small departure from circularity gives the compressive load running around a housing wall the same kind of off-centre step, and like the leaning stack, a drifting cross-section compounds under load instead of settling back.

Roundness as its own line on the drawing

Because a cylinder's buckling pressure falls away sharply once its out-of-roundness becomes a sizeable share of its wall thickness, a housing with a generous margin on paper can still fail well below its predicted pressure if its roundness tolerance was left loose. Pressure housing drawings therefore specify roundness as its own controlled dimension, and the manufacturing process includes an actual roundness check, on the machine that made the part or on dedicated measuring equipment afterward. The check itself is simple in principle: rotate the part through a full turn and record how far the wall's radius wanders from its average, since a calliper measuring diameter alone can miss some out-of-round shapes entirely.

That is an unusual demand. Most mechanical parts tolerate ordinary dimensional drift without their function changing, and a designer who treats roundness like any other secondary dimension, worth checking eventually, quietly reintroduces the vulnerability this whole set has been building toward. A drawing that lists wall thickness to three decimal places and leaves roundness implied has specified only half of what decides whether the part survives, and a housing can satisfy every explicit number on it while still being unfit for its pressure.

Thick walls, spare margin, and barrel hoops

Roundness matters most for thin-walled cylinders well inside the buckling-dominated regime, and progressively less as the wall thickens relative to the diameter, since a given deviation from true is then a smaller fraction of the wall's own thickness. It also matters less for a housing with a generous margin against buckling, since a shape working at the edge of its capability is far more exposed to a small error than one with headroom, which makes the margin chosen at the start of a design a cushion against every imperfection the finished part will carry.

Barrels, the subject of the next article, attack the same problem from the opposite direction, accepting that a set of individual staves will never sit perfectly round on their own and adding an external constraint that pulls the whole assembly back toward a true circle whether the staves cooperate or not.

More on Holding a shape