Springs — surface quality, set, and a bake that must not undo the temper
A spring stores energy in its surface: decarburization ruins it, a barrel nick starts a fatigue crack, and the de-embrittlement bake that saves a bolt can relax a spring. Wire standards, load at test length and the non-electrolytic option.
At a glance
| Thread / dimensional standard | EN 10270 / JIS G 3522 · no thread |
|---|---|
| Tolerance class | No thread — nothing in ISO 965 or ISO 1502 applies. |
| Barrel window | not used — coils tangle and nick |
| Rack / basket | wire ø0.5 – ø6 mm, free length ≤ ≈ 100 mm |
| Strength specification | no property class — wire grade + load at test length |
| ISO 4042 relief bake | High risk — consider a non-electrolytic coating |
| Reviewed | 2026-08-04 |
A spring stores its energy in the surface of the wire. That one fact explains why decarburization ruins it, why a nick picked up in a barrel is a fatigue crack waiting to start, and why the de-embrittlement bake that saves a bolt can flatten a spring. Nothing else on this list is this sensitive to what happens in the last few hundredths of a millimetre.
Why this shape is processed the way it is
Coils tangle. A barrel loaded with compression springs comes out as a knot of nicked wire, and every nick sits on the outer surface of the coil — exactly where torsional stress peaks in service. Springs are therefore run spaced, in baskets or on racks, and the handling decision costs more and matters more than the chemistry does. Extension springs with hooks add a second problem: the hook bend is already the highest-stress point on the part before anything is done to it.
Grades we run in this shape
Where these parts go
Threads, tolerances and dimensional standards
Springs are specified from the wire upward rather than from a thread. EN 10270-1 covers patented cold-drawn unalloyed wire in the SL, SM, SH and DH grades, EN 10270-2 oil-hardened and tempered wire; JIS G 3521 and JIS G 3522 cover hard-drawn and piano wire, JIS G 3560 oil-tempered wire and JIS G 4801 the SUP hot-rolled spring steels. Design and tolerance come from EN 13906-1 and EN 15800, and disc springs from DIN 2093. There is no gauge to pass — the acceptance characteristic is a load at a stated length.
Size band and handling route
Wire diameters of roughly half a millimetre to six millimetres, with free lengths up to about a hundred millimetres, are the practical window for small-parts finishing. The limit is not furnace size or tank size, it is whether the spring can be held and moved without tangling or touching its neighbours. That is a handling constraint, and it is why spring work is quoted per piece rather than per kilogram.
Strength specification
A spring has no ISO 898 property class, because it is not certified by a tensile test on the finished part. Oil-hardened and tempered spring wire is normally used in a hardness region well above the embrittlement threshold, and the design works instead from the wire tensile grade, the stress at solid length, and the load at a stated test length to EN 15800. After coiling the part is stress relieved — a lower temperature for patented cold-drawn wire than for oil-tempered wire — and then preset, so that it takes its permanent set in the factory rather than in the customer assembly.
No property class exists for this part type. Anything on a drawing that reads like one is a specification error worth fixing before the order runs.
Hardness: oil-hardened and tempered wire typically 44–52 HRC; stress relief ≈ 200–300 °C (patented cold drawn) or 380–420 °C (oil tempered)
Coating thickness and what limits it
Springs run at the highest surface stress of any part type on this list, at a hardness well above the embrittlement threshold, which makes electroplating the least comfortable option available. Where it is specified, the deposit is kept thin, and it should be understood that the inside of a coil will always carry less than the outside — current does not reach into a tight helix. Where the corrosion requirement is genuinely heavy, a zinc flake system applied and cured with no electrolytic step, of the kind ISO 10683 describes, is the route the fastener industry uses precisely because it introduces no hydrogen at all.
- ASTM B633 SC 1 — Fe/Zn 5
- zinc flake to ISO 10683 where the corrosion target is heavy
Hydrogen embrittlement relief
Two rules collide on a spring, and this is the part type where that collision is worst. ISO 4042 wants relief as soon as possible and normally within 4 hours at 190–230 °C, because the part is far above 390 HV. But the spring was already stress relieved after coiling — for some wire types at a temperature not far above the bake window — and re-heating close to that temperature relaxes it: the free length shortens, the load at test length drops, and the part is scrap for a reason that has nothing to do with corrosion. The bake temperature has to be chosen against the wire type and its original relief temperature, and the load has to be re-checked afterwards.
Hydrogen de-embrittlement bake checkerHow this part fails
- Delayed fracture after plating
A spring cracks hours or days after finishing, often in the coil rather than the hook. High surface stress plus hardness above the relief threshold plus electrolytic hydrogen is the worst combination in fastener work, and springs sit at the top of it.
- Fatigue crack from a decarburized surface
A soft, carbon-poor skin on spring steel cuts fatigue life dramatically because the highest stress is at the surface. The part passes a bulk hardness check and fails a cycle test, and only a metallographic section shows why.
- Load loss and a shortened free length
The spring measures the right dimensions and delivers the wrong force. Usually a heat exposure above the original stress-relief temperature — often a de-embrittlement bake chosen for a different part on the same order.
- Pitting at a coil contact point
Where coils touched during processing, the deposit is thin or absent, and corrosion starts there. On a working spring, a corrosion pit on the outer coil surface is a fatigue initiation site, so a cosmetic defect becomes a mechanical failure.
What to check on the lot
- 1Decarburization depth on a sectionISO 3887 on a mounted, etched section. For a spring this is the most informative single test there is, and it is invisible to every other check — a bulk hardness average will not show a surface layer that is a fraction of a millimetre deep.
- 2Load at test length, before and afterEN 15800 gives the tolerance framework. Measuring load before and after any thermal step is the only way to prove that a bake or a cure did not relax the spring, and it is cheap compared to a field return.
- 3Hardness on a wire sectionTaken on a mounted cross-section rather than on the curved outer surface, per ISO 6507-1 or ISO 6508-1. A reading taken on the round outside of small wire is biased and usually reads low.
- 4Coating thickness on the outside of the coilReport where the reading was taken, and expect the inside of the helix to be thinner. A salt spray result to ISO 9227 means very little unless the exposed orientation and the acceptance surface are both stated.
Test methods
Standards cited on this page
Questions engineers ask
Can springs be zinc electroplated at all?+
Yes, and it needs to be a deliberate decision rather than a default. The part is above the ISO 4042 relief threshold, so de-embrittlement is mandatory and the clock matters; the bake temperature then has to be checked against the wire's original stress-relief temperature so the spring does not relax. For heavily loaded or high-cycle springs, a non-electrolytic coating system of the kind ISO 10683 describes avoids the conflict entirely.
Why did our springs lose load after plating?+
Almost always the bake, not the plating. If the relief temperature used is at or above the temperature the spring was stress relieved at after coiling, the part relaxes: free length shortens and load at test length falls. Send the wire type, the original stress-relief temperature and the load specification with the order and the bake can be set against them.
What kills a spring first, corrosion or decarburization?+
On a cyclically loaded spring, decarburization usually wins, because it removes carbon from precisely the surface layer that carries the peak stress and it is already there before the part ever sees weather. Corrosion is the slower killer and the more visible one. A drawing that specifies salt spray hours but says nothing about surface decarburization is controlling the second problem and ignoring the first.
We reply to quote requests within 24 hours.
Guidance based on published standards and general fastener metallurgy. Every figure is a typical range that depends on the part, the material and the governing specification — the customer drawing and the applicable standard always prevail. Coating designations describe what a drawing can call for, not a result guaranteed on your parts without testing.