PROCESS ENGINEERING · August 15, 2026 · 8 MIN READ

Hot Forging Temperature Control: Start and Finish Windows, Soak Time, and Why 20 °C Matters

Carbon steel C45 is forged from 1,250 °C down to 800 °C; 42CrMo from 1,180 °C to 850 °C; austenitic stainless 304 from 1,200 °C to 900 °C. Finishing below the recrystallisation range raises flow stress by 40–60 % and is the most common single cause of die overload, laps and internal cracking — which is why the finish temperature, not the start temperature, is the number to control.

WHERE THIS APPLIES AT FORGEPRESS CHINA

ForgePress China runs induction and furnace heating with logged billet and finish temperatures; process cards for carbon, alloy and stainless grades are issued before the first article. Capacity: HKM3 hot-die forging line 1,000–5,000 t, part weight 0.1–500 kg, maximum envelope 1,200 × 800 mm, as-forged to ISO 8062 class C with critical dimensions machined to ±0.2 mm, process control to ISO 9001, first article 4–8 weeks from die design release.

KEY NUMBERS

Finish temperature matters more than start temperature: below the recrystallisation range, flow stress rises 40–60 %.
Overheating starts around 1,270 °C for plain carbon steel; burning (incipient melting) around 1,300–1,350 °C.
Induction heating a 60 mm billet to 1,200 °C takes 60–90 s; a furnace soak for the same section takes 45–60 min.
Scale loss is 1–3 % of billet mass at normal temperature; every extra 50 °C above target adds roughly 0.5 percentage points.

The three temperature numbers that matter

Forging temperature is not one number. It is a window defined by three points: the maximum safe start temperature (above which grain growth and overheating begin), the minimum finish temperature (below which flow stress climbs and ductility falls), and the preheat temperature of the die (which controls how fast the billet surface cools on contact).

Most quality problems blamed on "temperature" are actually a finish-temperature problem. A billet that starts at 1,250 °C and is still being worked at 720 °C is being forged outside its window, and no amount of start-temperature accuracy will fix it. When a part shows laps, incomplete filling at the last station, or shear cracks at the flash, measure the finish temperature before changing anything else.

Reference windows by material

Material	Start temperature	Finish temperature	Notes
C45 / AISI 1045	1,200–1,250 °C	≥ 800 °C	Widest window of the common grades
40Cr / AISI 5140	1,180–1,220 °C	≥ 850 °C	Narrower; sensitive to overheating
42CrMo / AISI 4140	1,150–1,180 °C	≥ 850 °C	Keep below 1,200 °C to limit grain growth
20CrMnTi / AISI 8620	1,180–1,220 °C	≥ 880 °C	Case-hardening grade; finish high to avoid surface decarburisation
304 / 06Cr19Ni10	1,150–1,200 °C	≥ 900 °C	High flow stress; higher finish temperature is mandatory
316L / 022Cr17Ni12Mo2	1,150–1,200 °C	≥ 900 °C	Even higher flow stress than 304
6061 aluminium	400–480 °C	≥ 380 °C	Narrow window; 20 °C overshoot causes incipient melting
Ti-6Al-4V	950–980 °C	≥ 900 °C	Below beta transus; window is only about 60 °C wide

Typical values for open-die and closed-die hot forging. Always confirm against the mill certificate and the applicable material standard before releasing a process card.

Heating: soak time, gradient and the section rule

A furnace load is not at temperature when the thermocouple is. For alloy steel, allow roughly 0.6–1.0 minutes of soak per millimetre of section thickness after the surface reaches set point, so a 100 mm billet needs 60–100 minutes of soak plus the ramp. Induction heating collapses that to seconds but introduces a through-thickness gradient: a 60 mm bar can show 80–120 °C difference between surface and core at the coil exit.

That gradient is not academic. A billet that is 1,230 °C at the skin and 1,100 °C at the core will forge with a hard core and an overhot skin, which produces the classic "surface crack, sound centre" defect pattern. Equalising for 20–40 s before the first blow costs nothing and removes most of it.

Where measurement error hides

Infrared pyrometers read surface temperature, and surface temperature is depressed by scale and by air cooling during transfer. Emissivity set to 0.9 on a scaled billet will read 30–60 °C low. A contact thermocouple reads accurately but only at one point, and it cannot follow a moving billet. The practical answer is to calibrate the pyrometer against a thermocouple on the same material once per shift, and to measure at the die, not at the furnace door.

For processes where the finish temperature is a customer requirement, log it. A chart of finish temperature per piece over a shift will show the real process capability, and it is the only evidence that survives a dispute.

Overheating versus burning

Overheating is metallurgical: austenite grain growth above roughly 1,270 °C in plain carbon steel, which cannot be reversed by later heat treatment and leaves a coarse, low-ductility microstructure. Burning is worse: incipient melting at grain boundaries around 1,300–1,350 °C, visible as a network of voids and as a fracture that crumbles rather than tears.

A burnt billet cannot be salvaged. An overheated one usually cannot either. The only reliable control is a hard upper limit in the furnace controller, with the thermocouple in the load rather than in the chamber, and an alarm that stops the line rather than a light that someone ignores.

Frequently asked questions

What is a normal forging temperature for 42CrMo?

Typically 1,150–1,180 °C at start and not below 850 °C at finish. Staying under 1,200 °C limits austenite grain growth, and finishing above 850 °C keeps flow stress inside the range the die was designed for.

Why does finishing too cold cause cracks?

Below the recrystallisation range the worked metal does not recrystallise during deformation, so strain accumulates at grain boundaries and flow stress rises 40–60 %. The die then has to deliver more force than it was designed for, and the part develops surface laps or internal shear cracks even though the die is undamaged.

How much scale loss should I expect?

One to three percent of billet mass under normal furnace practice. Every 50 °C above target adds roughly half a percentage point, and long soak times at temperature add more — which is why scale loss is a useful indirect check on whether the furnace is running hot.