A shaft is in two pieces. One camp says "it was overloaded — the coupling must have jammed." The other says "it's been cracking for months — maintenance should have caught it." The commercial consequences of the two stories are completely different, and the fracture surface almost always knows which one is true.
Overload: the one-event fracture
An overload failure happens once, when load exceeds capacity. In a ductile material it announces itself: visible stretching, necking, twisting — the part deformed before it broke, and the fracture is dull and fibrous, often at roughly 45° where shear finished the job. In a brittle material (or an embrittled one), there is little warning deformation; the fracture is flat and crystalline-bright, sometimes with chevron marks pointing back to where it started.
Either way, the key signature is uniformity: the whole surface formed in one event, so it tells one story with no history written on it.
Fatigue: the failure with a diary
Fatigue is progressive — a crack grows a little with every load cycle, sometimes over millions of cycles, and the surface keeps a diary:
- Beach marks — curved bands spreading from the initiation point, written whenever crack growth paused or conditions changed. They point back to the origin like ripples run in reverse.
- Ratchet marks — small steps at the origin where several cracks started close together and joined, a sign of multiple initiation and often of a rough surface or sharp detail.
- A smooth region and a rough region. The smooth part is the fatigue crack that grew slowly; the rough part is the final overload of the remaining ligament. The proportion between them says how much load the part carried at the end — a large final-fracture area means high service stress; a tiny one means the crack grew long at low stress.
- Little or no overall deformation. Fatigue parts usually break without stretching, which is why they surprise people.
Why the distinction changes everything
An overload verdict points at an event: what applied the abnormal load, or why was the component weaker than intended? A fatigue verdict points at a condition: a cyclic stress that has been there for a long time, concentrated at a detail — and replacing the part without changing the detail schedules the next failure. The same logic decides warranty and supplier questions: a manufacturing defect at the fatigue origin is a very different conversation from an operating excursion.
Real failures also mix the two: a defect initiates fatigue, fatigue thins the section, overload finishes it. The investigation's job is to find the initiating cause, not the final act.
When to contact MTIS
If the "fatigue or overload?" argument has commercial or safety consequences, do not settle it by eye or by committee — the surface features that decide it reliably need proper fractography, and they are easily destroyed by handling. Preserve the part (how →) and request an investigation. See also: how MTIS investigates failures →