Engineering Margins and Factors of Safety

Real hardware must survive uncertainty and off-design conditions, so we design to margins and report when a design does not meet them.

Margin is honesty about uncertainty

A margin, or factor of safety, is the gap between what a component must withstand and what it can withstand. It exists because loads, properties, and conditions are uncertain, and because real machines see off-design excursions. Choosing margins is not padding; it is quantifying how much we do not know and refusing to build to the edge of a guess.

How margins are set

• Larger where the physics or loads are less certain.
• Larger where failure is severe or hard to inspect.
• Tied to established design codes where they apply.
• Stated explicitly, so a reviewer can judge whether they are adequate.

When a design fails its margin

The standard is tested by what we do when a design does not meet its margin. The burner (Aegis / MetroVolt) plug coil is overstressed by roughly 3-3.9x at the design bore -- that is, the load exceeds capacity by several times, the opposite of a margin. We report that as a physics-and-engineering fact: infeasible as specified. We do not relax the allowable to make it pass, which would be exactly the fudging our standards forbid.

Margins are where scientific rigor meets buildability. A design that closes on paper but violates structural limits has not closed; it has only computed a number for a machine that cannot be built as drawn. Saying so is part of the standard. A margin is, in the end, a number that encodes humility: it is the amount by which we admit the world may differ from our model, sized to how much we actually trust that model rather than to how confident we would like to sound.