Reading a Broken Spring: Where It Broke Names the Cause

Get the two halves, put them under a loupe, and look at the broken faces. Do this before cleaning them. **A fatigue fracture** has two zones. A smooth, flat, often slightly discoloured region where the crack grew slowly, sometimes with fine curved beach marks showing where it stopped and started; and a rough, bright, crystalline region where the remaining metal finally tore. Find the smooth region and follow it back — it points at the origin, and the origin is the defect. **The smaller the rough zone, the lower the load.** A spring that broke under gentle loading crept a long crack before the rest gave way. A spring with a large rough zone broke soon after the crack started, which means the load was high. **A brittle fracture** is rough and bright all over, flat, at right angles to the wire, with no smooth zone at all. This is a single event, and on a spring it usually means hydrogen embrittlement from plating, or a badly hardened wire, or a genuinely enormous overload. If the spring was plated and broke early, suspect the plating. **A ductile fracture** is dull, torn, and the wire is visibly necked or twisted near the break. It has deformed before parting. This is honest overload, and the spring will usually also be set. **Look at the origin specifically.** On the inside of the coil is normal for fatigue. On the outside is not, and means the spring was loaded backwards, or buckled and rubbed. At a bright scar means something hit it or rubbed it. At a pit means corrosion started it.

**In the middle of the body, inside the coil.** The ordinary fatigue failure. The alternating stress is too high for the number of cycles it has done. Fix the loading, the preload or the peening — see the fatigue rung — not the steel. **One or two coils in from a ground end.** Suspect the grinding: a burnt or overheated end leaves tensile residual stress exactly where the spring bears. Also suspect an uneven seat, which loads that coil more than the rest. **At the very end coil, on a compression spring.** The seat. Either it is not flat, or it has a raised edge, or the spring is not square and is rocking on it. **At the hook bend, on an extension spring.** The commonest single failure in all of spring work, and usually a design problem rather than a manufacturing one: the bend radius is too tight. A replacement with the same hook will do the same thing. Fix it with an extended hook, a generous radius, or a threaded insert. **Where the legs leave the coil, on a torsion spring.** The stress is highest at the transition, and it is made worse by any sharpness in the bend. Also check the spring was the correct hand. **Several coils broken, or a coil ground flat.** Not a break at all — the spring buckled, or went solid, and has been rubbing on something. Look for the matching wear mark on the housing or the rod. **Every spring in the set broken.** A system problem: a resonance, a shock load, or a temperature. One failure is a spring; five identical failures is the machine.

**Pitting corrosion** is the big one. A pit is a sharp notch, it forms at the surface where the stress is highest, and on a peened spring it removes the compressive layer that was doing the work. Springs in damp places should be specified with that in mind, not simply painted — paint does not stay on a surface that flexes. **Stress corrosion cracking** needs stress and a corrosive environment together, and a loaded spring always has the stress. The cracks are branching and run across the wire, and the fracture face often has no fatigue zone. Stainless in chlorides is a classic combination, which is why 'stainless' is not automatically the right answer for a wet application. **Fretting** happens where two things rub microscopically — coil against coil in a nested pair, or the end coil against its seat. It shows as a reddish brown powder and a dull patch, and it starts fatigue cracks very effectively. **Hydrogen embrittlement**, from plating or from pickling before it. The spring breaks brittle, often within days of being fitted, at a load it should carry easily. The fix is baking after plating, and the prevention is asking whether the spring needs plating at all. **Heat.** A spring that has been anywhere near a welding torch or a brake has been tempered, and it will not be the spring it was. Look for discolouration. **Contact with anything.** A spring that touches its housing wears a flat, and a flat on the outside of a coil is a stress raiser and a sign it buckled. Find the matching mark on the other part before assuming the spring was at fault.

Six questions. They take five minutes and they are the difference between a repair and a repeat. **1. Is the replacement actually the same spring?** Measure the wire with a micrometer and the coil diameter with a caliper, and count the coils. A spring that looks right and has wire a tenth of a millimetre thinner has 78% of the rate — the fourth power again. **2. What length is it installed at?** Measure it in place, not in the catalogue. A worn seat, a missing shim or a bent bracket changes the working length, and therefore the force and the stress, as surely as a wrong spring does. **3. Can anything press it solid?** If yes, and it was not designed for that, that is your answer and a stop is the fix. **4. Is the seat flat and square?** Put the spring on a flat surface with a square against it. More than about 3 degrees of lean means it will load one side of every coil. **5. Is anything rubbing?** Look for bright marks on the coils and matching marks on the housing or rod. Rubbing means buckling or insufficient clearance, and both will come back. **6. Why now?** If the machine ran for years and has broken two springs in a month, something else changed — a speed, a load, a temperature, a different supplier. The spring is reporting on the machine. **Then write down the free length and the installed length** before you close it up. Next time, that record turns the whole of this rung into one subtraction.