Cooling time t8/5 calculator

How long the joint takes to fall from 800 to 500 °C decides what structure forms in the heat-affected zone. It is the parameter a welding procedure hides behind its heat input range and preheat temperature.

Download the diagram: SVG · PNG

The formulas in EN 1011-2

The standard gives two, because heat leaves in two different ways.

Three-dimensional — thick material, heat escapes in every direction:
t8/5 = (6700 − 5·T₀) · Q · (1/(500−T₀) − 1/(800−T₀)) · F3

Two-dimensional — thin plate, heat spreads only in the plane:
t8/5 = (4300 − 4.3·T₀) · 10⁵ · Q²/d² · (1/(500−T₀)² − 1/(800−T₀)²) · F2

Which one applies is settled by the transition thickness — the thickness at which both give the same answer. The calculator works it out and shows it beside the result: below it the 2D formula rules, above it the 3D one. Under three-dimensional flow the thickness drops out of the result altogether — adding more material changes nothing.

The shape factor

A joint sheds heat the worse, the more walls meet at one place. The standard captures that with the coefficients of Table D.1.

For a T-joint the standard gives F2 as a range, 0.45–0.67. The calculator uses the upper value: it yields a longer t8/5, that is, the less cautious estimate — and that is the one worth checking.

For the T-joint the standard gives F2 as a range. The calculator takes the upper value: it yields a longer cooling time, that is, the less cautious estimate — and that is the one worth checking with a measurement, not the safe one.

What counts as a good time

Where it earns its keep

• Choosing the preheat. If the preheat temperature comes out high and there is nothing to heat with, t8/5 shows how much has to be made up with heat input.
• Qualifying a procedure. The WPQR records the actual heat input; converting it to t8/5 tells you whether the toughness and hardness stand a chance.
• Investigating cracks. A cold crack in the HAZ is almost always too short a cooling time together with hydrogen. Check both, not just the electrode brand.

t8/5 table by heat input and plate thickness

Cooling time for a butt weld without preheat (20 °C), worked out with the same formulas as the calculator — including the choice between the 2D and 3D formula by transition thickness.

t8/5 in seconds for a butt weld (F = 0.9) at a starting temperature of 20 °C. The 3D mark means the thickness is above the transition value and the three-dimensional formula applies; no mark means two-dimensional. Window 5–25 s as in the calculator.

What the table shows

On thin plate the time grows with the square of the heat input: 6 mm at 0.8 kJ/mm is 18.2 s, at 1.5 kJ/mm already 63.9 s — far outside the 5–25 s window, with a risk of coarse grain. On thick plate (3D mark) the thickness stops mattering: 20 and 30 mm at 1.2 kJ/mm give the same 5.7 s, because the heat now escapes in every direction. There the time rises only in proportion to heat input and is hard to lengthen without preheat.

Worked examples

Example 1: 10 mm plate, 1.2 kJ/mm, no preheat

These are the calculator defaults (butt weld, F = 0.9). The transition thickness comes out at 16.1 mm, so at 10 mm the 2D formula applies: t8/5 ≈ 14.7 s, an average cooling rate of 300 / 14.7 ≈ 20.4 K/s. The result sits in the middle of the working window.

Example 2: the same plate preheated to 100 and 150 °C

At T₀ = 100 °C the time rises to 21.1 s, at 150 °C to 27.5 s, which is already past the 25 s upper limit. On thin plate it is easy to overdo preheat: for 10 mm at this heat input 150 °C is too much unless hydrogen demands it.

Example 3: 20 mm S355 preheated to 120 °C

Without preheat, 20 mm at 1.2 kJ/mm gives 5.7 s (3D formula) — right at the lower limit. Preheating to 120 °C, which the preheat calculator works out for S355 with a rutile electrode, stretches it to 7.6 s, and raising the heat input to 1.5 kJ/mm takes it to 9.6 s. Both levers together move the joint well inside the window.

Example 4: the same 10 mm in different joints

At 1.2 kJ/mm and 20 °C: bead on plate (F = 1.0) — 16.4 s, butt weld (0.9) — 14.7 s, fillet in a T-joint (0.67) — 11.0 s. Three plates draw heat faster than two, so a T-joint cools a quarter faster than a butt joint in the same plate.

What heat input to choose for a target t8/5

The formulas can be turned round. With 3D heat flow, 20 °C and F = 0.9 the time is proportional to the heat input: t8/5 ≈ 4.76 · Q. To get 10 s on thick plate you need about 2.1 kJ/mm — a lot for one pass, which is why thick sections are usually slowed down with preheat instead. With 2D heat flow the time grows with the square of Q: for 10 mm t8/5 ≈ 10.2 · Q², so 10 s takes about 1.0 kJ/mm and 25 s about 1.56 kJ/mm. Heat input from your welding parameters comes from the heat input calculator.

2D or 3D formula: what the transition thickness depends on

The transition thickness grows with heat input and starting temperature. For a butt weld without preheat it is 13.1 mm at 0.8 kJ/mm, 16.1 mm at 1.2, 18.0 mm at 1.5 and 23.2 mm at 2.5 kJ/mm. The same 20 mm plate can therefore be "thick" for a root pass and "thin" for a wide fill pass — the calculator picks the formula for you and shows which one it used.

Typical mistakes in t8/5 calculations

• Arc energy instead of heat input. The formula takes Q with the thermal efficiency of the process, not bare U·I·60/(v·1000).
• Units. Q is in kJ/mm; entering kJ/cm (ten times more) gives absurdly long times.
• T₀ from the first pass for the whole weld. For later passes the starting temperature is the interpass temperature, usually higher than the preheat.
• Wrong joint type. A T-joint worked out as a butt joint overstates the time by about a third (example 4) — the assessment comes out too optimistic.
• One window for every steel. 5–25 s applies to non-alloy and low-alloy steels; quenched and tempered steels have a narrower range from the material data sheet.

Author: Dmitry Frolov, welder and metal fabricator Updated: 28 September 2026