§1From ore to molten iron
Every spring begins as rock. The quality ceiling of the finished wire is set surprisingly early — at the ore body, the flux quarry and the coke oven.
Iron arrives at the mill as an oxide. The classic ore is hematite, red and running roughly 30–65 % iron; magnetite is its black, magnetic cousin. As the rich direct-shipping ores thinned out, mills turned to low-grade taconite at 25–35 % iron, which is ground, concentrated and rolled into pellets about 6–13 mm across carrying 62–65 % iron — a better blast-furnace feed than most natural ore. Fine ore and flue dust are sintered into clinker so nothing is wasted.
Two more ingredients join the ore in the furnace burden. Limestone acts as the flux: it melts, floats on the bath and scavenges phosphorus, sulphur and the earthy gangue into slag. Coke — bituminous coal baked for 14–17 hours to drive off volatiles — is fuel, structural support for the charge, and the source of the carbon the iron absorbs. Carefully sorted scrap completes the recipe.
The blast furnace itself is a counter-current chemical reactor on a civil-engineering scale — a typical stack about 9 m in diameter and 30 m tall. Charge enters the top at roughly 200 °C; halfway down it passes 540 °C; in the melting zone it reaches about 1930 °C. Daily output per furnace climbed from about 1300 t of pig iron in 1950 to beyond 10 000 t by the mid-1970s, helped by enriching the blast from air's natural 21 % oxygen towards 30 %. What flows from the taphole is pig iron: iron carrying 3.5–4 % carbon plus silicon, manganese, phosphorus and sulphur — far too dirty to spring. Refining it into steel is the next step, and spring steel ends up claiming only about 2 % of everything the industry pours.
§2Steelmaking routes
Steelmaking is subtraction: burn the excess carbon and impurities out of pig iron, then add back exactly the elements wanted. The vessel used for that subtraction has changed twice in living memory.
Bessemer & open hearth
The Bessemer converter blew air up through the bath — fast, but it could not remove phosphorus and never suited quality steels; it is obsolete. The open hearth, sweeping flame across a shallow bath at about 1650 °C with a charge near 52 % pig iron and 48 % scrap, made nearly all high-carbon and alloy spring steel until the mid-1970s.
Basic-oxygen furnace
From 1976 the big mills replaced open hearths with BOF converters: a supersonic lance blows ≥98 % pure oxygen onto the bath, refining a heat in a fraction of the time — roughly three times the daily output — while accepting up to 30 % scrap. Most carbon and low-alloy spring steel now starts here.
Electric & vacuum
Electric-arc and induction furnaces melt selected scrap with precise alloy additions — practically all stainless and tool steels take this route. Vacuum melting strips oxygen, nitrogen and hydrogen from the melt for aircraft-grade fatigue performance.
For the spring designer the route matters mainly through cleanliness: electric and vacuum heats carry fewer non-metallic inclusions, and inclusions are where fatigue cracks are born. Music-wire steel, for example, is melted only by the acid open-hearth, basic-oxygen or electric processes.
§3Ingot practice — killed, rimmed, capped
Between furnace and rolling mill the steel must solidify, and how it solidifies is a specification item in its own right.
Teemed into a mould, a cooling ingot wants to do two unhelpful things. First, it shrinks, pulling a central cavity — the pipe — down its axis; a refractory hot-top sleeve, roughly 250–750 mm tall, keeps a reservoir of liquid metal feeding the centre so the pipe never forms. Second, dissolved gas makes the melt effervesce. Adding aluminium, ferrosilicon or manganese deoxidises the steel so it lies quiet in the mould — a killed steel, with uniform analysis, little segregation and freedom from ageing. Every steel above 0.20 % carbon is killed, which is to say: all spring steel is killed steel.
Rimmed steel, deoxidised incompletely so a pure low-carbon skin freezes against the mould wall, and its arrested variant capped steel, are below-0.20 %-carbon products for deep drawing — around 30 % of total output, and none of it destined for springs. Semi-killed grades sit between, in reinforcing bar and similar work. One further distinction survives from the melting shop: basic practice (limestone slag, magnesite lining) digests high-phosphorus iron and feeds most hard-drawn wire, while acid practice, run on cleaner iron with a silica lining, gives easier slag control and a particularly uniform, clean steel.
§4Billet, rod and the wire mill
Wire is not squeezed to size — it is stretched. Everything in the wire mill is arranged around that fact.
Reheated ingots are rolled on blooming and billet mills into billets about 50–100 mm square and roughly 9 m long, then rolled again into hot-finished rod and coiled. Before drawing, the coils are pickled in hot dilute sulphuric acid, rinsed, and dipped in hot borax, which dries to a lubricant-carrying film; coil ends are welded together so the machine sees one continuous rod. The rod is drawn down to a convenient size — about 6.35 mm, at which point it is formally "wire" — and passed to continuous multi-die machines.
Cold drawing elongates every grain along the wire axis and leaves the surface smooth, bright and uniform. Production speeds are startling: routine drawing runs near 366 m/min, with later water-cooled multi-die machines reaching about 610 m/min. Dies are tungsten carbide for most sizes; the finest wires are pulled through holes drilled in diamonds or sapphires, and a good diamond die can pass some 32 km of wire with no measurable wear. Round wire dominates, but square, rectangular, oval and hexagonal sections are drawn or rolled when a spring section demands it.
The industry's output is difficult to picture — enough wire annually to circle the Earth several thousand times, feeding more than 160 000 distinct products. A spring is one of the few of them asked to survive millions of full-stress reversals.
§5Patenting and the two wire routes
One heat treatment separates spring wire from fence wire. Patenting sets up a structure fine enough to survive brutal cold reduction — and it splits wire-making into two philosophies.
Patenting austenitises high-carbon rod well above the upper critical — commonly about 1000 °C — then quenches it into molten lead near 477 °C. Cooling that fast, at that temperature, martensite cannot form; the steel transforms instead to an extremely fine, equiaxed pearlite that combines high strength with the ductility needed to take pass after pass through the dies, each pass cutting section by roughly 20 %. An English refinement improves torsional quality: austenitise at about 900 °C, quench in lead at about 530 °C, then hold the finished wire at 315 °C for three minutes after the final pass.
From patented rod, two routes diverge. The soft process draws to size and then restores softness — either a spheroidise anneal in protective atmosphere, giving annealed wire that is coiled first and hardened and tempered afterwards to about HRC 42–46, or a continuous harden-quench-and-temper line through lead baths that ships the familiar oil-tempered spring wires. The hard process keeps every bit of the cold work: patented rod is simply drawn until the accumulated strain hardening delivers the target tensile strength. Music wire, hard-drawn MB, the 302/304 stainless wires, Monel, Inconel, phosphor bronze and spring brass all take this route.
Hard-process wire hides a design trap: its elastic modulus varies with the amount of drawing. Diameter reductions up to about half raise E by nearly a tenth; drawing further surrenders most of that gain. A worked case from the mill floor — music wire drawn 2.8 → 1.5 mm (a 46.8 % diameter reduction) shows about 1917 MPa tensile with E ≈ 187.5 GPa; five more passes to 0.99 mm (64.9 % reduction) lift tensile to about 2413 MPa while E falls to roughly 183.4 GPa. No single modulus fits all sizes, which is why load-test results on real springs drift from textbook formulas — and why the property tables in the sheets that follow quote E and G by diameter band.
§6What each element contributes
Iron is never listed in an analysis — it is the balance. Everything else is there in carefully rationed percentages, each buying a specific property.
For scale: an ordinary machine steel such as SAE 1020 carries 0.15–0.25 % carbon; music wire carries 0.80–0.95 % — four to five times as much. That single ratio explains most of the difference in care, cost and heat-treatment discipline between structural steel and spring steel.
| Element | Typical range | What it buys | Watch-outs |
|---|---|---|---|
| Carbon | 0.55–1.05 % (springs) | Tensile strength, elastic limit, hardness; eutectoid near 0.85 % | Below ≈0.40 % hardening is unreliable; above ≈1.05 % brittleness rises with no hardness gain |
| Manganese | 0.60–1.20 % (music wire 0.20–0.60 %) | Sound ingots; rolls and draws cleanly; deep hardening | Deep hardening forces oil quenching — water cracks spring sections |
| Silicon | 0.10–0.30 %; 1.80–2.20 % in silico-manganese grades | At high levels lifts tensile strength without costing ductility; aids hot work | Never used alone — paired with 0.60–0.90 % Mn |
| Phosphorus & sulphur | ≤0.055 % each; usually far less | Nothing — residual impurities | Cut strength, ductility and shock resistance; free-machining additions have no place in springs |
| Chromium | 0.80–1.10 % (alloy grades); 12–20 % in stainless | Strength, toughness, deep hardening, heat and corrosion resistance | Raises the hardening temperature |
| Nickel | to ≈2 % in low alloy; 7–8 % with 18 % Cr | Toughness and wear; with high Cr produces austenitic (non-hardenable) stainless | Little effect on hardenability on its own |
| Vanadium | ≥0.15 % | Elastic limit, impact resistance, fine grain; with Cr, fatigue life | Guards grain size during accidental overheating |
| Mo · Co · W | speciality grades | Hardenability and hot strength (Mo); red-hardness for tools (Co, W) | Cost — reserved for tool and high-speed steels |
Nearly every one of these alloying metals was, and largely remains, an imported commodity for major steelmaking nations — manganese, chromium and cobalt almost entirely so. Alloy surcharges and substitution grades (Chapter 4's Ni–Cr–Mo bar steels are a wartime example) trace straight back to that supply map.
§7Why the metallurgy matters to the designer
None of this chemistry is trivia. Four practical rules fall straight out of the production chain.
Specify the melt, not just the number
Fatigue-critical springs deserve the cleaner electric or vacuum heats; inclusions seed cracks. For ordinary service the BOF product is entirely adequate — and priced accordingly.
Respect the carbon window
Order a tight carbon range. A wide specification band lets two deliveries of the "same" steel harden and relax quite differently, and load tolerances suffer for it.
Never assume one modulus
Hard-drawn products carry a size-dependent E and G. Use the banded values on each material sheet, not a single textbook constant, or measured loads will not match calculated ones.
Oil, never water
Manganese makes spring steel deep-hardening by design. The same property makes a water quench a cracking hazard — every hardening schedule in this series assumes oil.
With the production chain established, the sheets that follow take the materials one family at a time — beginning with the plain high-carbon wires that carry the bulk of the world's springs.
§8Quick reference
Ore → wire
Pellets 62–65 % Fe → blast furnace (≈1930 °C melt, pig iron 3.5–4 % C) → BOF (O₂ ≥98 %) → killed ingot → billet 50–100 mm → rod → patent ≈1000 → 477 °C → draw ≈20 % per die.
Always true of spring steel
Killed steel, every time. Carbon 0.55–1.05 %. Mn-driven deep hardening → oil quench only. P and S ≤0.055 % each, the lower the better.
Modulus drifts with drawing
Heavy reductions move E by several per cent — 187.5 GPa at 46.8 % reduction vs 183.4 GPa at 64.9 % in the worked case. Quote E and G by wire-size band.
