Chapter 1 of 10 · 10 min read
How a forge weld forms
See how two pieces of hot iron become one without melting: clean contact, temperature, pressure and time, the welding range of iron by colour, and why oxide, not heat, is the forge welder's enemy.
A forge weld looks like magic to anyone seeing it for the first time. Two bars come out of the fire glowing almost white, the smith lays one on the other, a few quick blows of the hammer, and they are one bar. Neither was melted, no filler was added, and when the weld is good, nobody can find it afterwards. The explanation is not mysterious, and understanding it changes how a smith works, because it shows that what makes or breaks a forge weld is not mainly the heat, but the cleanliness of the surfaces being joined.
This chapter explains how metal surfaces bond in the solid state, what temperature, pressure and time contribute, how iron’s colour shows its temperature and its welding range, and why oxide is the problem every forge welding practice is designed to solve.
Atoms that touch
A piece of metal holds together because its atoms are bonded to their neighbours throughout. At a surface, the atoms have neighbours on one side only. If two perfectly clean, perfectly flat metal surfaces could be brought together so that their atoms were as close as the atoms inside the metal, they would bond, and there would be no surface left: the two pieces would be one. That is the principle of every solid-state weld, from the blacksmith’s scarf to the friction welds and diffusion bonds of modern industry.
Two things stand in the way. Real surfaces are rough, so when two pieces are pressed together they touch only at the tips of their high points, a tiny fraction of the apparent area. And real surfaces are covered with films: oxide, which forms on iron at once in air and grows rapidly when it is hot, together with dirt, grease and moisture. Atoms separated by a layer of oxide do not bond. Every successful solid-state weld overcomes both obstacles, by making the metal soft enough to be squeezed into full contact and by getting the films out of the way.
What heat does
Heat helps in three ways. It softens the metal, so that a hammer blow or a squeeze flattens the high points and brings the surfaces into contact over their whole area. It speeds the movement of atoms across the interface, so that a bond, once contact is made, grows quickly into a continuous structure, with grains growing across the original boundary. And, for iron, it makes the oxide behave differently: near welding heat the scale on iron can be dissolved by a flux into a fluid slag that the hammer squeezes out of the joint, as Chapter 3 explains.
The lesson that heat is not everything comes from an experiment Hart recorded in 1910. Two pieces of iron, planed and highly polished, were pressed together for several days at about 650 °C, far below the welding heat of the smith, and they partly welded. With surfaces clean and flat enough, even modest heat and steady pressure were enough to start a bond. The smith’s white heat is needed because the smith’s surfaces are neither clean nor flat, and because the smith has seconds, not days.
| Forge welding | Fusion welding | |
|---|---|---|
| Metal melted | no, joined in the solid state | yes, a pool of molten metal |
| Filler | none | usually a filler wire or rod |
| Surface films | must be removed by flux, pressure and squeezing out | float to the pool's surface as slag or are shielded off |
| Pressure | essential, by hammer or press | not needed |
| Heated zone | a long length of each piece | a narrow zone beside the weld |
| Joint preparation | shaped ends: scarfs, laps, clefts | bevels or square edges |
| Typical weakness | trapped scale and unbonded patches | porosity, slag, cracks, lack of fusion |
What pressure does
Pressure brings the surfaces into contact and does the squeezing. The hammer’s first blows, struck at the middle of the joint, press the surfaces together there and drive the fluid slag outwards towards the edges, so that clean metal meets clean metal. Hart described exactly that practice: start the weld at a middle point and work outward, driving the slag away. Later blows consolidate the joint and work the metal, closing any small gaps and refining the structure. A weld that is struck hard at the edges first can trap slag in the middle, where it stays as an unbonded patch.
The forge welder also deforms the metal deliberately. A forge weld with no deformation would be like two pieces of paper laid together; one with plenty of deformation stretches the interface, breaking up any thin oxide that remains and exposing fresh metal on both sides. That is why the scarfed ends of Chapter 4 are made thicker than the finished bar, by upsetting, so that the hammering that welds them can also draw them down to size.
The energy of a blow
The hammer’s work can be put in numbers. A 1 kg hand hammer swung so that its head strikes at about 8 m/s carries about ½ × 1 × 8² = 32 joules; a 6 kg sledge swung by a striker at about 7 m/s carries about 150 joules; a power hammer delivers far more, blow after blow. This is an illustration with typical speeds. At welding heat, steel’s resistance to deformation is a small fraction of its strength when cold, so even a hand hammer’s blow flattens hot metal visibly, and most of the blow’s energy goes into squeezing the joint together and spreading it. As the bar cools, its resistance rises steeply, and the same blow does less and less, which is the hammer’s own warning that the heat is going. Large welds need heavier blows, from a striker or a power hammer, because a hand hammer cannot deform a thick section enough to close it before it cools.
Contact over the whole face
The aim of all the hammering is contact over the whole of the joint’s face, not just at its centre. A weld that has bonded over half its area carries roughly half the load of the bar, and the unbonded half acts as a crack waiting to grow. The smith checks for full contact by watching the joint’s edges as the blows land: flux and slag squeezing out all round, then the edges closing to a fine line that disappears as the joint is worked, show that the faces have come together everywhere.
Time
At welding heat, the bond forms in a fraction of a second once clean surfaces touch, because atoms move very quickly at those temperatures. The problem with time in forge welding is the opposite one: the smith has very little of it. The moment the bars leave the fire they begin to cool and to oxidise, and within a few seconds they are below the temperature at which the weld will take. Chapter 5 calculates how fast a bar cools, and shows why the smith’s movements from fire to anvil are drilled until they take no thought.
The colours of hot iron
Before thermometers could be used in the forge, smiths judged temperature by colour, and they still do. Hot iron glows dull red, then cherry red, orange, yellow and, at the highest temperatures, white. The colours depend on the light in the shop: the same bar looks redder in sunlight than in a dim forge, which is why many smiths keep the forge area shaded.
Hart described the welding range of nearly pure wrought iron as unusually wide, from imperfect welds at a cherry red to good welds at a dazzling white, just before the iron begins to burn and melt. Steel’s range is narrower, and it narrows further as the carbon content rises, as Chapter 6 shows. In practice, smiths weld wrought iron and mild steel at a light yellow to white heat, roughly 1,250 to 1,350 °C, and judge the moment by the colour, the look of the flux, which turns runny and bubbling, and the first few sparks from the bar’s surface.
Burning
Above the welding range, iron burns. Its surface oxidises so fast that it throws off bright sparks, and the oxygen penetrates along the grain boundaries, so that the metal becomes brittle and crumbles under the hammer. Burnt iron cannot be welded and must be cut off. Hart noted that dissolved rust makes a burnt, brittle joint. A smith who watches a bar begin to spark in the fire has seconds to take it out; one who leaves it in has a piece of scrap. The narrowness of the margin between welding heat and burning, especially in higher-carbon steel, is one reason forge welding takes long practice.
Scale and why it matters most
Iron forms scale, a layer of iron oxides, whenever it is hot in air, and the hotter it is, the faster the scale grows. At welding heat a bar can grow a visible layer of scale in the time it takes to lift it from the fire. If that scale stays between the surfaces, they cannot bond; the weld either fails outright or has unbonded patches that break under load. Everything in forge welding practice is arranged to deal with scale: the reducing fire of Chapter 2 that slows its growth, the flux of Chapter 3 that dissolves it, the rounded joint faces of Chapter 4 that let it squeeze out, and the speed of Chapter 5 that leaves it no time to form.
Forge welding beyond the anvil
The same principles work at an industrial scale under controlled conditions. In resistance butt and flash welding, two bars are clamped end to end, heated by an electric current, and pushed together, with the flash process expelling molten metal and oxide from the joint before the upset. In friction welding, one part spins against another until the interface is hot and plastic, and then the two are forged together. Roll bonding and diffusion bonding join sheets or parts by pressure at high temperature in clean or protective conditions. Hart’s account of the welding of pipe, at first lap-welded by hand over a coke fire and then butt-welded in long lengths by machine, traces the path from the smith’s hammer to the mill. Resistance Welding by the Numbers covers the electrical processes.
The running case: the first lesson
This is an illustration. The shop’s owner begins every forge welding class with two experiments before any student strikes a weld. The first heats two pieces of mild steel to welding heat without flux, lays them together and hammers them; they stick weakly and break apart at the vice, showing a dull, scaled interface. The second repeats it with flux and a clean, quick transfer; the weld holds, and when it is bent in the vice it bends with the bar. Students who have seen the difference learn to watch their flux and their timing, not just their colour.
Why forge welds vary
Because a forge weld depends on cleanliness, contact and timing at the moment of hammering, and those vary from weld to weld, forge welds are more variable than most modern welds. Two welds made by the same smith from the same stock can differ markedly in strength, and the difference cannot be seen from the outside once the weld is finished. Chapter 7 shows how large the variation was in tests of experienced smiths’ work, and why forge welding is no longer used for safety-critical industrial parts except where the process is controlled and tested.
A failure: the cold shut
This is an illustration. A student in an early class, pleased with a scarf weld on a fire poker, drew it out and finished it neatly. When the poker was bent cold in the vice as a test, it opened along a clean, flat plane at the weld, showing a grey, smooth surface with a dark rim.
The two scarfs had touched only at their centres and had been hammered flat without bonding over most of the area, because the bars had cooled below welding heat during a slow transfer. The flat grey surface was the original scaled faces, never joined. The owner calls it a cold shut, the forge welder’s commonest defect, and uses the broken poker in every class since.
Exercise: colour, range and margin
A smith welds mild steel at about 1,300 °C and finds that the bar begins to spark and burn at about 1,450 °C, and that the weld will not take below about 1,200 °C. What is the working range in degrees, and what fraction of the welding temperature is the margin above it before burning? If a higher-carbon steel’s burning point is about 1,330 °C and its lowest welding temperature about 1,150 °C, where should it be welded, and with what margin?
The mild steel’s working range is about 1,450 − 1,200 = 250 °C. Welding at 1,300 °C leaves about 150 °C below burning, roughly 12% of the welding temperature, and 100 °C above the bottom of the range.
The higher-carbon steel’s range is about 1,330 − 1,150 = 180 °C. Welding near the middle, about 1,240 °C, leaves about 90 °C to burning and 90 °C to the lower limit, a much tighter window. The smith must take it out at a lower colour, a yellow rather than a white, and work more quickly, which is why higher-carbon steels are among the hardest to forge weld.
