Welding Preheat & Carbon Equivalent
Preheat and interpass temperature from carbon equivalent and COMBINED thickness — the number that makes a 12 mm fillet behave like 36 mm plate.
The steel
Type a composition instead (from the mill certificate)
A real certificate usually shows a lower carbon equivalent than the specification maximum, which can take a whole band off the preheat.
The joint
The same plate, four ways
Identical steel and identical 12 mm plate. Only the joint geometry and the state of the consumables change — and between them they move the requirement further than switching grade would.
Three things worth knowing
The thickness that matters is the combined thickness. Preheat is driven by how fast the joint cools, and heat escapes through every plate meeting at it. A fillet weld between two 12 mm plates does not behave like 12 mm material — three plate paths lead away from that joint, giving a combined thickness of 36 mm, which sits two rows further up any preheat table. EN 1011-2 defines it exactly this way: sum the thicknesses of all parts meeting at the weld, measured 75 mm out. A cruciform sums four. The commonest preheat mistake is not picking the wrong number from the table — it is reading the table at the thickness stamped on the drawing instead of the thickness the heat actually sees.
Preheat does not warm the steel up, it slows the cooling down. Hydrogen cracking needs three things at once — dissolved hydrogen, a hard crack-sensitive microstructure, and tensile stress — and removing any one prevents it. Slower cooling attacks two: it gives a softer microstructure and it gives hydrogen time to diffuse out instead of being trapped. Which is why interpass temperature matters exactly as much as the initial preheat, and why carefully preheating a joint and then letting it go cold between passes throws it all away.
- Damp consumables cost as much preheat as extra thickness. Moving from oven-held low-hydrogen electrodes to as-received rutile ones shifts the requirement two whole bands — the same move as going from light section to very heavy. A heated quiver costs less than the propane it replaces.
- Across ordinary structural steels, thickness beats grade. Everything from mild steel to HSLA sits between about 0.35 and 0.50 carbon equivalent, and over that span thickness swings the answer roughly twice as far as composition. That reverses outside it: a hardenable alloy like 4140 sits near 0.78 and needs real preheat at any thickness.
- Pcm beats the IIW carbon equivalent on modern low-carbon steel. Below about 0.18% carbon the IIW formula over-weights manganese and alloying; Pcm weights carbon far more heavily and predicts cracking better. Where the two disagree on a modern structural grade, believe Pcm.
- Measure it, don't eye it. A torch playing on one spot reads hot at the surface while the section behind it is still cold — which is how a "measured" preheat turns out not to have been one. Contact thermometer or temperature crayons, a hand-width either side of the joint.
How to use
- Pick the grade, or type a composition from the mill certificate.
- Choose the joint type — a fillet counts three plate paths, not one.
- Set the consumable hydrogen scale, which moves the answer as much as thickness does.
- Preheat to the figure shown and hold it as interpass temperature.
Frequently asked questions
What is combined thickness and why does it matter?
It is the sum of the thicknesses of every part meeting at the joint, measured 75 mm out from the weld, and it is what EN 1011-2 asks you to read the preheat table at. Preheat is driven by how fast the joint cools, and heat escapes through every plate leading away from it. A fillet weld between two 12 mm plates has three plate paths and therefore a combined thickness of 36 mm, which sits two rows further up any table than 12 mm does. A cruciform joint sums four.
What is the commonest preheat mistake?
Reading the table at the thickness stamped on the drawing rather than the thickness the heat actually sees. People pick the right table and the right steel and then look up 12 mm for a fillet joint that behaves like 36. The error is invisible because everything about the process looked correct, and it shows up later as cracking in a joint that was welded to what everyone believed was the procedure.
Does preheat just warm the steel up?
No — it slows the cooling down, and that distinction decides everything else. Hydrogen cracking needs three things at the same time: dissolved hydrogen, a hard crack-sensitive microstructure, and tensile stress. Remove any one and it does not happen. Slower cooling attacks two of them at once, because it produces a softer microstructure and it gives hydrogen time to diffuse out of the weld instead of being trapped there.
Why does interpass temperature matter as much as preheat?
Because it is the same requirement continued. If the reason for preheating is to keep the cooling rate down, then letting the joint go cold between passes puts the cooling rate straight back where it was and undoes the whole thing. A carefully preheated multi-pass weld that was allowed to cool between runs has had the benefit of preheat on the root pass only, which is rarely the pass you were worried about.
How much difference do damp electrodes make?
About the same as doubling the plate thickness, which surprises most people. Moving from oven-held low-hydrogen consumables to as-received rutile ones shifts the requirement two whole bands up the table. Low-hydrogen coatings are hygroscopic, so electrodes that have sat out over a humid weekend are no longer low-hydrogen no matter what the box says. A heated quiver costs a great deal less than the propane it saves.
Which matters more, the grade or the thickness?
Across ordinary structural steels, thickness. Everything from mild steel through HSLA sits between roughly 0.35 and 0.50 carbon equivalent, and over that span thickness moves the answer about twice as far as the grade does — the same A36 needs nothing in light section and over 100 degrees in heavy. That reverses once you leave structural steel: a hardenable alloy such as 4140 sits near 0.78 and needs serious preheat at any thickness at all.
Should I use the carbon equivalent or Pcm?
Below about 0.18 per cent carbon, Pcm is the better predictor and the IIW carbon equivalent tends to over-weight manganese and the alloying elements. Above that, the IIW figure is the one the codes are written around and the one to work to. Both are shown here for that reason. Where they disagree on a modern low-carbon structural steel, Pcm is usually the one to believe.
Is there a maximum interpass temperature too?
For several materials, yes, and it is a real ceiling rather than a guideline. Quenched and tempered steels lose the strength their heat treatment gave them if you exceed it. Austenitic stainless sensitises, forming chromium carbides at the grain boundaries that strip the corrosion resistance exactly where the weld is. Duplex has the tightest ceiling of all and a heat input floor as well, so it is constrained from both directions at once.
🔒 This tool runs entirely in your browser. Nothing you enter is uploaded, logged, or stored.