Electric Welding Temperature: Arc, Preheat & Interpass

A fabricator in Duluth welded a 1/2-inch A36 bracket out the door on a January Friday with E7018 electrodes, 120 amps, and a bead that looked textbook. By Monday morning the fillet had split open along the heat-affected zone, right where the plate was thinnest. The welder had done nothing sloppy. The problem was that nobody ever asked what temperature the steel was at. Because here is the strange part of electric welding: the arc is roughly as hot as the surface of the sun, and the plate next to it may still be nearly freezing. Once a welder understands that gap between the temperature in the arc and the temperature in the joint, most cracking, distortion and HAZ headaches start to make sense.

What Electric Welding Temperature Actually Means (And Why Your Arc Being 10,000°F Doesn’t Matter)

Search “electric welding temperature” and you get a hundred numbers with no clear answer: arc temperature, melting point, preheat, interpass, heat input. They get used interchangeably, and mixing them up is a leading cause of cracked, warped, or rejected welds.

Plainly: there is no single number called welding temperature. There are at least four separate quantities, each measured somewhere different, each set by a different person or process, and each with its own failure mode.

The Four Numbers People Call “Welding Temperature”

1. Arc (column) temperature. The plasma temperature at the middle of the arc, commonly cited in the roughly 6,000 to 10,000°F (3,300 to 5,500°C) range at the core of the arc column. Nobody sets it, you cannot control it directly, and as a quality number for your joint it is nearly meaningless.
2. Base metal temperature (preheat). The minimum temperature the base metal must reach before the arc is struck. Codes check this first, and it is the one most often skipped in a hurry.
3. Interpass temperature. The maximum temperature the metal may reach between passes. Codes cap it because the HAZ keeps getting re-heated.
4. Heat input. Not a temperature, but the calculated proxy for how much heat the joint absorbs, reported in kJ/in or kJ/mm. It tells you how fast the joint heats and cools.

The Counterintuitive Part: Hot Arc, Cold Plate

Picture a hairdryer and a blowtorch aimed at a cold steel plate. The flame is thousands of degrees, the hairdryer air is around 100°C, yet the blowtorch leaves the plate hotter. It depends on how the energy arrives: a concentrated flame hits fusion temperature instantly in a tiny spot while the rest of the plate barely moves.

That is the arc. Steel filler melts at roughly 2,700 to 2,800°F (about 1,500°C) and the arc column sits far above that, yet the plate beside the puddle may only be a few hundred degrees warmer than when you started. The heat is real and dangerous, but intensely localized.

So arc temperature is the wrong number to watch. One consequence worth internalizing: filler metal melts in the arc, then drops to base metal temperature as it flows into the joint. An overly cold joint gives instant solidification problems, cold lap and lack of fusion, because the metal arrives molten with nowhere to put its heat and freezes against a cold wall instead of fusing into the plate.

The Three Phase-Change Temperatures That Govern Everything

• Solidus and liquidus. The melting range of the metal. Below solidus it is solid, above liquidus it is fully liquid. Filler and base metal must both pass through this range for a fusion weld to form.
• Austenitizing range. The band where steel transforms into austenite and its crystal structure reorganizes, which for many carbon steels begins above 1,300°F (about 700°C). Heat the HAZ into this range and its grain structure is reset; push it higher or hold it longer and the grains coarsen, which is what high heat input and high interpass temperature do.
• Martensite start (Ms). The temperature at which austenite begins transforming to hard, brittle martensite on cooling. Cool through this range too fast on a hardenable steel and you get a hard HAZ that cracks under restraint.

Who Sets What, Where You Measure It, What Breaks

Quantity	Who sets it	Where you measure it	What breaks if ignored
Arc / column temperature	Physics, fixed by the process and arc length.	Inside the arc plasma, not measurable with shop gear.	Nothing directly; treating it as a quality indicator.
Preheat (base metal)	The WPS and the code. AWS D1.1 requires minimum preheat at a distance at least equal to the thickness of the thickest member, and not less than 3 in. (75 mm) in all directions from the point of welding.	On the base metal near the joint, before striking the arc.	Cold cracking, hard untempered HAZ, lack of fusion.
Interpass (maximum)	The WPS and code, often tied to material family and the filler’s diffusible hydrogen level.	At the weld joint, before tacking the next pass.	Grain coarsening, wide soft HAZ, sensitization in stainless, more distortion.
Heat input	The welder, via travel speed, voltage and current, bounded by the WPS.	Calculated from recorded volts, amps and travel speed.	High: wide soft HAZ, distortion, residual stress, spatter, sensitization. Low: lack of fusion, cold lap, undercut, cracking.

One nuance from the AWS Welding Digest on preheat and interpass: for carbon steels, exceeding the specified minimum preheat by about 100°F (40°C) is generally acceptable. Quenched-and-tempered steels are the exception; they need both a maximum and a minimum preheat established, and both followed closely.

What This Article Covers

We start with the physics of the arc, then what AWS D1.1 and AWS D14.8M actually require, then heat input math, temperature measurement with real instruments, diagnosing cracking and distortion from their pattern, per-material temperature tables, heat and fume safety, and a bench checklist.

New to the process? Get the basics of stick welding down first. Our guide to the advantages of stick welding explains why SMAW is forgiving with heat, and the all-position welding rod guide covers rod selection.

Electric welding temperature is not one number, it is four, and the one your joint actually cares about is how hot the metal is and how fast it cools. Control that, and the rest takes care of itself.

The Physics of the Arc: How Hot Does Electric Welding Really Get?

Watch a welder work and one question always comes up: how hot does that get? The answer sounds absurd and is completely true. The arc column in an electric welding process can run around 10,000°F, hot enough to vaporize the electrode tip outright, while the steel plate on the other side of the joint might still be sitting at 70°F. Both statements are true at once, and that gap is where most weld defects live.

An electric welding arc is not a flame and not a glowing piece of metal. It is a sustained plasma discharge: ionized gas carrying current between electrode and workpiece. Electron collisions strip gas molecules of their electrons, and the plasma is hot enough to melt and vaporize metal on contact. The electrode tip sits in a zone where it is vaporizing rather than just melting.

Steel’s solidus to liquidus range sits near 2,700 to 2,800°F (about 1,500°C), so the arc runs roughly three to four times hotter. Filler metal melts inside the arc, crosses as a droplet, and mixes with melted base metal in the pool. The base metal only has to reach its own melting point at the fusion line, which is why preheat and interpass temperature exist as separate control variables from anything inside the arc.

Arc Length Changes Everything

Arc length and arc voltage are directly linked. Hold the electrode closer and the arc is short, voltage drops, and heat concentrates into a small, deeply penetrating footprint with a narrow bead. Lengthen the arc and voltage climbs, heat spreads wider, and you get a flatter bead with less penetration, more spatter, and more atmosphere on the molten pool.

Current behavior depends on the power source. A constant-current machine, typical in SMAW and FCAW shops, holds amperage at the dial and lets voltage float with arc length; push the electrode out and current barely budges. Constant-voltage machines, common in GMAW setups, work the opposite way. A wandering arc length means a wandering heat input, which is why procedures pin down arc voltage and travel speed.

This is why “arc temperature” is a poor proxy for joint temperature. What moves the base metal is how much energy you put in and how fast you travel.

Where the Energy Actually Goes

Not all of the power you dial in ends up in the plate. Industry accounts for this with a thermal efficiency factor, an efficiency multiplier applied when calculating heat input, using values from EN 1011-1:2009 as cited in AWS Welding Digest (June 2025):

Process	Thermal efficiency factor
SMAW, FCAW, GMAW	0.8
GTAW and plasma arc welding	0.6
Submerged arc welding (SAW)	1.0

The rest does not disappear. It leaves through the arc column, through spatter, through heated shielding gas, and as radiant and convective heat off the plate. In GTAW, roughly 40 percent of the power you switch on never reaches the workpiece. In SAW the assumption is that essentially all of it does, which is one reason submerged arc welding runs hot and fast.

Emissivity: Why Surface Readings Lie

Infrared thermometers measure radiant energy and convert it to temperature using an assumed emissivity. Bare, shiny, freshly ground steel reflects most incoming radiation instead of emitting it, so the instrument reads far too cold. Oxidized or mill-scale steel is the opposite problem and still needs correction. A thermocouple clamped to the plate has no emissivity problem, which is why shops doing serious preheat control land on the clunky, reliable option. That gap is central to preheat measurement generally.

The Thermal Cycle Inside the Joint

At the fusion line, metal melts almost instantaneously, then heat conducts into the plate along a steep thermal gradient. Just outside the weld, metal that melted or partially melted and then cooled forms the heat-affected zone; farther out, metal saw only a temperature rise and slower cool, so it transforms only partly. That gradient of properties is why the HAZ is the weak link in most failures.

Three terms describe the zone. Arc energy is raw energy per unit length of weld. Weld pool temperature is the temperature of the molten metal, typically well above the liquidus and often in the 2,900 to 3,000°F range for steel. Solidification cooling rate is how fast the pool freezes, and it sets the grain structure of the deposit.

Arc temperature is fixed by physics and chemistry, so you cannot change it. The heat delivered to the plate is under your control: amperage, arc length, travel speed, preheat, pass sequence. A procedure that caps heat input is not worried about the arc. It is worried about how fast you heat and cool the metal next to the joint, and that is a number you control with travel speed.

For process-specific heat behavior, see the advantages of stick welding breakdown. For nonferrous work the physics changes again, as covered in the aluminum MIG welding guide.

Preheat and Interpass Temperature: What the Codes Actually Require

Two numbers decide whether a carbon steel weld holds up: how hot the base metal must already be when

Preheat and Interpass Temperature: What the Codes Actually Require

Two numbers decide whether a carbon steel weld holds up: how hot the base metal must already be when you strike the arc, and how hot you are allowed to let it get between passes. Get them backwards, or treat them as optional, and you get delayed cracking that shows up days later.

Preweld (preheat) temperature is the minimum temperature the base metal must reach before welding starts. Interpass temperature is the maximum temperature the metal is permitted to reach between passes. A cold plate pulls heat out of the pool faster than you can feed it; a hot plate softens and distorts. Two halves of one thermal control problem.

Why cap the temperature you may reach between passes, when hotter metal sounds safer? Because every pass reheats the heat-affected zone of the pass before it. Repeat that cycle and HAZ grains keep growing, the zone keeps softening, and residual stress and distortion accumulate. On stainless steel the same reheating can trigger sensitization; on quenched-and-tempered steel it can temper away the strength the material was heat treated to hold. A ceiling on interpass temperature is a ceiling on damage.

Where the numbers live: AWS D1.1 Table 5.8

For structural steel, the industry’s go-to reference is AWS D1.1 Table 5.8, “Prequalified Minimum Preheat and Interpass Temperature,” reached through Clause 5.7. Table 5.8 sorts steels into Categories A, B and C, grouping them by susceptibility to hardening (carbon and alloy content) and by the diffusible hydrogen level of the consumable. Each category carries its own minimum preheat and maximum interpass values, and because the table is part of the prequalification basis, a joint built to a prequalified WPS inherits those limits automatically.

Here is the point most shops never hear stated plainly: almost nobody at the bench recomputes preheat. You get the number from the WPS on your job traveler, or from the contract spec. That is not shortcutting, it is the system working. The engineering happened when the procedure was qualified. If you want to understand what your WPS is protecting you from, the electrode classification and its hydrogen rating are worth a look, and why SMAW still wins in structural work covers the consumable side of the trade.

Annex B: the alternative routes

AWS D1.1 Annex B, “Guidelines on Alternative Methods for Determining Preheat,” is a normative annex, so it carries real code weight. It offers two paths to establishing preheat without relying on tabulated values:

• Heat-affected-zone hardness control method, which is limited to fillet welds. It proves acceptability by measuring HAZ hardness instead of matching a table.
• Hydrogen control method, which works from composition, thickness, diffusible hydrogen level and the assumed degree of restraint.

Both demand more documentation and care than reading a table, which is why most fabricators reach for them only when the job calls for it.

AWS D14.8M and the four carbon equivalent methods

For systematic preheat determination, AWS D14.8M (ISO/TR 17844) provides four methods, each built on a carbon equivalent that predicts hardenability:

Method	Origin	Formula
CE	European	C + Mn/6 + (Cr+Mo+V)/5 + (Ni+Cu)/15
CET	European	CE + 3Nb
CEN	Japanese	C + (Mn+Cr+Mo+V)/10 + (Ni+Cu)/20 + N/600 + Nb/300
Pcm	AWS D1.1 Annex B	C + Si/24 + Mn/10 + Ni/40 + Cr/20 + Mo/12 + V/10 + Nb/4 + 3B/4

The CEN method applies to material thickness from 3/8 to 8 in., to steels and filler metals with yield strength up to 110 ksi, and to effective heat inputs up to 125 kJ/in, and it uses a 0.9 factor for submerged arc welding rather than the 1.0 factor applied to CE and CET. D14.8M also carries a CE-based method in Chapter 2 that sets preheat from a critical HAZ hardness value established from joints 1/4 to 4 in. thick.

The trade-off nobody wants to talk about

Preheat is not free. It eats production time, consumes fuel gas or electricity, and adds a setup step to every joint. The counterweight is that alternative methods, combined with modern low-carbon steels and H4/H8 low-hydrogen filler metals, can deliver real time, cost and environmental savings. The catch is that misapplied, they can also put a susceptible microstructure exactly where you did not want one. The economics work; the metallurgy still has to.

Here is how a shop team decides. A 1-in A36 bracket carrying service load into a 1/2-in attachment gets a preheat call, because thickness, restraint and cold-weather plate combine to give a cooling rate that can produce hard HAZ microstructures. The 3/8-in fillet on a non-critical bracket usually does not, particularly with a low-hydrogen electrode and controlled heat input. Same day, same shop, same welder, two different calls, because the variables are thickness, restraint, hydrogen level and material cleanliness, not habit.

Heat Input: The Number That Really Controls Weld Temperature

Arc temperature gets all the attention, but heat input decides how hot your joint gets and how the steel reacts. Instead of “110 amps on 3/16-inch E7018,” you describe the pass as “1.1 kJ/mm.” One is a setting, the other is what the metal experiences.

The Formula, in Both Unit Systems

Heat input is power multiplied by time, divided by the distance the arc traveled:

H = (60 × V × A) / (1000 × travel speed)

V is arc voltage in volts, A is current in amps, and travel speed in mm/min gives kJ/mm. Use inches per minute for the same math in kJ/in.

A real pass worked step by step: 24 volts and 110 amps on a fillet weld at 140 mm/min travel speed.

• Step one: power = 24 V × 110 A = 2,640 watts, or 2.64 kW.
• Step two: 2.64 kW ÷ 140 mm/min = 0.0189 kJ/mm of arc time per mm of travel.
• Step three: multiply by 60 seconds per minute = 1.13 kJ/mm.
• Step four: convert to imperial = 1.13 × 25.4 = 28.7 kJ/in.

An ordinary mid-range fillet heat input, produced by an ordinary setting. Nothing at the machine looked special, and that is the point: the setting and the result are not the same thing, and heat input connects them.

Arc Energy vs. Effective Heat Input

Two “correct” numbers can describe the same weld, which confuses welders until they understand why.

ASME Section IX qualifies procedures using arc energy, the raw output of that formula with no correction. ISO 15614-1 and EN 1011 instead use effective heat input, which multiplies the raw result by a process efficiency factor covering how much arc energy lands in the workpiece. The rest goes into spatter, the arc column, shielding gas and radiant heat.

Common factors are 0.8 for SMAW, FCAW and GMAW, 0.6 for GTAW and PAW, and 1.0 for SAW, where the flux blankets the arc. The AWS Welding Digest on preheat and interpass shows how those factors feed the carbon equivalent methods used to set preheat requirements.

Apply 0.8 to the example and effective heat input becomes about 0.91 kJ/mm, or 23 kJ/in. Same weld, same arc, two legitimate numbers depending on the code language.

What Each Variable Actually Does

Voltage and current set the instantaneous power. Travel speed is the lever the welder actually controls, and because heat input is inversely proportional to travel speed, slowing down raises heat input in direct proportion. Double the travel speed, halve the heat input. The most common cause of an overheated weld is not a wrong amperage setting but a welder “lugging” along the joint because fit-up is poor or the torch got heavy by hour two of a long production run.

Other levers exist. Longer arc length raises voltage and spreads heat less concentratedly. A larger electrode diameter concentrates more heat at the root, and switching from SMAW to FCAW changes deposition rate and thermal profile. The case for stick welding is worth reading for a thermal comparison, since SMAW’s high heat per pass changes how you plan multi-pass work.

Predictably, preheat runs the other way. If the plate is already at 150°C before you strike an arc, less arc energy is needed to reach the same fusion condition. Preheat does not replace the calculation, but it reduces the arc energy the joint requires.

Too Much Heat, Too Little Heat

Push heat input too high and you get a wide, soft, coarse-grained heat-affected zone, more angular distortion, higher residual stress, more spatter and, in stainless steel, a bigger sensitization window. Cracking risk climbs too, because a wide hot pool shrinks more as it cools.

Push it too low and you get lack of fusion, cold lap, undercut, and in highly restrained joints, cracking from fast cooling and high local restraint. There is no magic number that is always right, only a range set by material, thickness, restraint and position, and that range is written into the procedure. AWS D1.1 Clause 5.7 sets heat input limits for prequalified procedures, and the preheat and interpass values in Table 5.8 assume your WPS stays inside the qualified range. Fillets also carry a minimum heat input requirement, because a pass that fuses inadequately along the leg is not a fillet weld in any meaningful sense. Exceed the maximum and the weld sits outside prequalification and outside code coverage, no matter how good it looks.

Bead Placement, Weave, and Multi-Pass Joints

Stringer beads put each pass in one place, keeping the HAZ narrow. Weaving spreads a pass across a wider footprint, lowering effective heat input per unit length there but widening the HAZ considerably. Weave width has a prequalified limit, and flux core welding on aluminum is a case where weave control gets tight fast, because aluminum’s conductivity and oxide behavior punish a wandering torch.

On any multi-pass joint, interpass temperature is cumulative heat. The second pass starts on a plate the first pass already warmed, so a pass at the same nominal heat input lands on a hotter base and the condition drifts upward unless the interpass cap is respected. Position matters too: vertical-up travel is slower than flat or horizontal, so heat input climbs without the welder changing a setting.

Heat input is the bridge between the machine settings and the metallurgical reality in the joint. Get it right and the rest of the temperature system becomes manageable.

Temperature and Metallurgy: What the Heat Affected Zone Does to Your Steel

Steel does not care how much you welded it. It cares how you heated it and how fast you let it cool. Two welds with the same electrode, amperage, travel speed, and preheat can land in different metallurgical states, and the difference can surface days later as a crack in the heat affected zone, not the bead. The HAZ beside the weld metal is the only zone heated and cooled without being replaced by filler.

The Cycle: Heat, Hold, Cool

For most carbon and low-alloy steels, the story starts when the base metal enters the austenitizing range, roughly above 1,300°F (700°C). Austenite is the only phase from which useful room-temperature microstructures can form, and that fact drives the basic preheat and heat cycle explanation.

Landmark	Approximate value	Why it matters
Austenitizing range	Above ~1,300°F (700°C)	Austenite forms; original structure lost
Ms (martensite start)	Varies with carbon and alloy	Hard untempered martensite forms
Solidus and liquidus	~2,650–2,800°F	Melting range; grains partially melt

In the HAZ, austenite grains grow while held in that range. Large grains mean low toughness, and this is where electric welding temperature control bites. Higher heat input means more time above the austenitizing range; a higher

Temperature and Metallurgy: What the Heat Affected Zone Does to Your Steel

Steel does not care how much you welded it. It cares how you heated it and how fast you let it cool. Two welds with the same electrode, amperage, travel speed, and preheat can land in different metallurgical states, and the difference can surface days later as a crack in the heat affected zone, not the bead. The HAZ beside the weld metal is the only zone heated and cooled without being replaced by filler.

The Cycle: Heat, Hold, Cool

For most carbon and low-alloy steels, the story starts when the base metal enters the austenitizing range, roughly above 1,300°F (700°C). Austenite is the only phase from which useful room-temperature microstructures can form, and that fact drives the basic preheat and heat cycle explanation.

Landmark	Approximate value	Why it matters
Austenitizing range	Above ~1,300°F (700°C)	Austenite forms; original structure lost
Ms (martensite start)	Varies with carbon and alloy	Hard untempered martensite forms
Solidus and liquidus	~2,650–2,800°F	Melting range; grains partially melt

In the HAZ, austenite grains grow while the metal sits in that range. Large grains mean low toughness, and this is where electric welding temperature control bites. Higher heat input means more time above the austenitizing range; a higher interpass temperature re-coarsens the same grains. Grain coarsening cuts HAZ toughness even when hardness tests look acceptable, as explained in this technical look at measuring preheat and interpass temperatures.

The joint then cools. Slow cooling through the transformation range gives ferrite and pearlite, which are soft and ductile. Faster cooling gives bainite, tougher and stronger. Very fast cooling through Ms gives martensite, hard and brittle. Cooling rate is no mystery; it follows from heat input, plate thickness, restraint, and preheat.

Why the HAZ Is the Weak Link

Three things converge in the HAZ. It was melted or partly melted, so it inherited a coarse or untempered microstructure. It sits beside a joint shrinking as the weld metal solidifies, carrying residual tensile stress. And it may still hold hydrogen off the electrode or coating moisture.

A hard, coarse microstructure in tension with hydrogen present is the classic recipe for HAZ hard cracking. That is why preheat is not just a comfort measure. Preheating the plate slows cooling through the transformation range, giving hydrogen time to diffuse out before the HAZ turns hard and crack-prone. More preheat also raises Ms, so transformation happens higher up, where diffusion is still fast.

Martensite, Carbon Equivalent and the Strength/Toughness Trade-Off

Whether you get martensite depends on hardenability, which scales with the carbon equivalent formulas codes use. The higher the carbon equivalent, the deeper a weldment can harden and the more likely a fast-cooling HAZ turns hard. The same equations AWS publishes in the D1.1 structural welding code are the shop tool for predicting this.

More preheat plus lower heat input gives a softer, more ductile HAZ with better toughness. Too little preheat and too much heat input gives a hard, brittle HAZ that is crack-prone and poor at absorbing strain. This logic reverses for quenched and tempered steels, where codes get strict.

Quenched and Tempered Steel: The Opposite Risk

Q&T steel gets its strength from a controlled quench and temper. Overheat the HAZ and you temper it back toward its as-rolled condition, throwing away the strength the customer paid for. So codes set a maximum as well as a minimum preheat, and the WPS controls both numbers rather than leaving them to the welder. Consumable choice matters too; see this guide to alloy steel welding electrodes.

Stainless Steel: Sensitization and Ferrite Balance

Stainless steel brings a different failure mode. Between roughly 800 and 850°F (425–455°C), chromium precipitates as carbides along grain boundaries, robbing the surrounding metal of the chromium it needs to resist corrosion. That is sensitization, and it is why interpass limits are non-negotiable in corrosion-critical service. A dwell in that range, from a high interpass temperature or a slow, wide bead, can produce a part that passes inspection and fails in service.

Wire arc additive manufacturing research on ER316L stainless steel showed the same mechanism: interpass temperature changed the ferrite content of the deposit and drove sigma-phase precipitation, with measurable effects on corrosion behavior, as documented in npj Materials Degradation (2023). Same alloy family, same physics, as the code limits describe.

Nickel Alloys: Liquation Cracking

Nickel alloys, nickel-iron alloys, and many high-chromium stainless grades are prone to HAZ liquation cracking. Low-melting segregation films form at grain boundaries; when restraint pulls the joint while those films are still liquid they open into cracks. The remedy: low heat input, interpass temperature under the limit, and heat kept moving instead of dwelling in one spot.

Measuring Welding Temperature: Tools, Methods and Shop Reality

You can own the best welding machine in the building and still crack a joint, because nobody measured the steel. Preheat and interpass temperature are the numbers a welder must hit and rarely proves, and the gap between guesswork and a logged 165 °C at the joint is where plenty of failed repairs start.

Verifying welding temperature is not complicated science: a small toolbox, a repeatable sequence and a habit of writing things down. A solid welding temperature measurement guide from Xiris covers the same ground, and this section distills what actually matters on the shop floor.

The Tool Hierarchy

Most shops settle on two or three of these.

• Thermocouples. Type K is the common choice for welding work: wide range, low cost, and available in clamping and surface-mount forms, accurate enough for preheat and interpass verification. Clamp the tip so it sits in the base metal near the joint, not merely on the surface, where it reads skin temperature rather than the heat a few millimeters down where a crack forms.
• Indicator meters and recorders. A simple indicator gives you a live number. A recorder writes a trace of every pass, useful on a long-running joint where interpass temperature creeps upward while nobody is watching.
• Digital pyrometers and infrared thermometers. Fast, non-contact and good for scanning a plate for cold spots. Emissivity-sensitive, so they work best on clean, matte, oxidized base metal next to the joint. Never point one at the arc, and never trust a reading off shiny molten metal.
• Temperature-indicating crayons and sticks. Cheap, single use and genuinely useful: they melt at their rated temperature, giving quick confirmation that the preheat target was reached before the arc was struck. They will not give you interpass temperature, and they lose accuracy on dirty or scaled surfaces, but for a go/no-go check on a 150 °C preheat they are hard to beat.
• Thermal imaging cameras. Overkill for a single fillet, but they map the gradient across a plate in a way nothing else can. Chasing a distorted part or hunting heat pooling? A temperature field image is worth a great deal.

Where You Measure Matters as Much as What You Measure

AWS D1.1 sets the location rule that trips up most shops: minimum preheat must be established at a distance at least equal to the thickness of the thickest member, but not less than 3 in. (75 mm) in all directions from the point of welding, so the coldest point that matters can sit an inch away from the joint. As Lincoln Electric summarizes, the plate has to be warm through, not just warm near the arc.

1. Preheat minimum: measure on the surface of the member at the required distance from the joint, on the side that cools last, usually the far face of the receiving member in a T-joint.
2. Interpass maximum: measure at the weld joint itself, right before tacking or starting the next pass. This is the number most often missed, because the joint is hot and the welder is eager.
3. Starting temperature: read it before you heat anything, so the log shows the plate came in cold.

Heating Methods and Their Trade-Offs

• Oxy-fuel torch. Cheapest and the first thing most shops reach for. It is also the hardest to control, and it leaves oxidation and scale on the joint surface, which then has to be ground or chipped out before welding.
• Resistance heating mat or induction coil. Even, controllable and repeatable, but they need serious electrical capacity. Induction gives localized heat fast; a mat warms a large area evenly and slower.
• Infrared or ceramic fiber heaters. Gentle and controlled, and often used to hold temperature rather than raise it quickly.
• Oven or bulk preheat. For whole parts, large weldments, and anything where uniform temperature through the section matters more than speed. This is where a welding course or your own process engineer earns its keep: it is the method that most reliably satisfies a preheat range instead of just hitting a spot.

The Sequence That Works

Check the starting temperature. Heat to the preheat minimum. Verify with a second instrument if the part is critical, because the torch heating the far side of the plate is not the flame you are watching. Weld, then wait for the joint to cool to the interpass maximum before restarting.

While welding, monitor with crayons at the joint, a pyrometer on the heat-affected zone, or thermocouples wired to a recorder on long joints. Verify instrument accuracy and log preheat and interpass temperatures on the traveler or weld log. Many contracts, code inspection plans and quality management systems require this documentation, and it is far cheaper to produce than to defend.

Check the current edition of AWS D1.1 and your own WPS. Editions change, and the qualified procedure always governs.

Temperature-Related Failures: Cracking, Distortion, Porosity and Hard HAZ

Every weld defect is a message about the thermal cycle behind it: how fast the metal cooled, how much it was restrained, how much hydrogen was present, how clean the joint was. Read the message and you fix causes instead of symptoms.

Hot Cracking: The Crack That Arrives While You Are Still Welding

Hot cracking happens while the weld metal is still partly liquid. As the pool solidifies, the last liquid to freeze forms segregation films of low-melting elements such as sulfur, phosphorus and tin along grain boundaries. These films have a solidus far below the parent metal. For steels, cracking occurs above the solidus of the weld metal, roughly in the 2,650–2,750°F (1,480–1,510°C) range, while that low-melting film is still liquid and being pulled apart by contraction and restraint.

The drivers stack up:

• Restraint. The more the joint resists shrinkage, the more stress builds against those weak films.
• Wide bead and high heat input. A fat, slow bead stays hot longer, so the vulnerable temperature window lasts longer.
• High sulfur and phosphorus. These drive the segregation that creates the films.
• Poor fit-up. Gaps force the root to bridge and stretch the pool, opening the center of the bead as it freezes.

Appearance gives the diagnosis away. Hot cracks open immediately after welding, while the metal is still warm. They sit in the center of the bead, where the last liquid froze, or along the fusion line where the pool boundary solidified last. Edges are smooth and rounded, sometimes with blue or brown heat tint nearby. A crack still warm to the touch is almost certainly hot.

Cold Cracking: The Crack That Shows Up Overnight

Cold cracking is hydrogen-assisted and delayed. It commonly appears hours to days after the arc is off, when the heat-affected zone has hardened into a brittle, untempered microstructure while residual tensile stress is still locked in. Three conditions must exist together:

1. A susceptible microstructure. A hard, martensitic HAZ with low toughness. Low carbon equivalent helps; high restraint on thick plate makes it worse.
2. Diffusible hydrogen. Hydrogen from moisture, rust, paint, oily steel or cellulosic electrodes. Cellulosic electrodes (E6010, E7010, E8010) bake into the weld and release hydrogen late in the cooling cycle, exactly when the steel is least able to tolerate it.
3. Tensile stress. Residual stress from shrinkage, structural restraint or a rigid fit-up.

The timing is the tell. Cold cracking frequently shows up in the morning, in a shop where the part was welded the previous shift and the welder was not there to see it happen. Preheat and interpass control reduce the risk, but the preventive triad is straightforward: low-hydrogen filler, adequate preheat and low restraint.

Distortion: When the Part Moves Before You Measure It

Uneven heating and uneven contraction bend metal. The joint pulls toward the hot side while the cold side holds it back, and the result is angular distortion, bending, buckling or twist. Out-of-flat on a flange is often a heat-input problem wearing a machining problem’s clothes.

Preheat is a double-edged tool. It lowers the thermal gradient and cuts distortion from a cold start, but a large preheated zone means more of the part is expanding and contracting. Preheat plus high heat input on a large assembly can distort more than cold welding at low heat input on the same part. The trick is to heat only the region you need, keep the heated zone small, and balance the passes around the joint so shrink stresses cancel each other out.

Distortion also has a code dimension. PWHT for stress relief is no longer something a shop adds on its own. As The Fabricator’s overview of recent AWS D1.1 changes explains, the revised Clause 7.8 language requires PWHT for stress relief to be specified in the contract documents or approved by the engineer, and performed per a WPS qualified for PWHT. If nobody specifies it, the default is no stress relief. Plan for it in the procedure, not after the part is warped.

Porosity, Spatter, Undercut and the Rest of the Thermal Family

• Porosity comes from gas: moisture on the joint, contaminated shielding gas, a leak in the hose, or a long arc that pulls nitrogen and moisture into the pool. Short arc, clean steel, dry gas, correct flow rate.
• Spatter usually means excess voltage or too long an arc, so the bead gets wide and droplets are thrown rather than deposited.
• Undercut shows up as a groove along the toe of the bead. Travel speed too high, work angle wrong, or the arc blowing the puddle backward all contribute.
• Lamellar tearing splits plate layers parallel to the rolling direction in thick, restrained joints. The thermal cycle plus stress raises the normal to the plate surface past the through-thickness strength.
• Sensitization attacks stainless steel in the 800–850°F (425–455°C) range, so controlling interpass temperature and heat input protects corrosion resistance, not just toughness.
• HAZ liquation cracking is a nickel alloy problem, where a wide, hot bead leaves low-melting constituents in the HAZ that crack under restraint.
• Slag inclusions in SMAW are usually a cleaning problem between passes, not a heat problem.

None of these are exotic. They are ordinary thermal-cycle mistakes with ordinary causes.

Where the Code Comes In

Two clauses of AWS D1.1 matter most here. Clause 5.7 sets heat input limitations, and those limits must be observed. The same clause allows changes to preheat and interpass temperature when qualified by testing, which gives you a legitimate path to correct a problem on a specific job rather than guessing on the shop floor.

Once you can name the failure mode, the fix is usually a temperature fix: more preheat, less heat input, lower hydrogen, less restraint or better fit-up. That is the whole game.

Matching Temperature to Your Material and Process

The practical question at the bench isn’t what the four temperatures are, it’s what numbers to use for the material in front of you and the process you have. Charts can get you in the ballpark, but they will never override your qualified welding procedure specification (WPS) or contract documents. Treat them as a sanity check, and verify against the governing procedure before striking an arc.

Typical Shop Starting Points

Most shops rely on reference values like those discussed by Xiris on measuring preheat and interpass temperatures to build that ballpark. The pairings below reflect commonly published starting points, not code requirements.

Material family	Typical process	Typical preheat range	Typical max interpass	What drives the decision
Carbon steel (A36, A572, etc.)	SMAW	about 150–200°C (300–400°F)	about 300°C (570°F)	Thickness, carbon equivalent, restraint
Low-alloy / HSLA steel	GMAW	about 100–150°C (200–300°F)	about 250°C (480°F)	Hardenability and HAZ cracking sensitivity
Quenched-and-tempered steel	GMAW / SMAW	Minimum and maximum both specified	Often 150–250°C	Overheating destroys strength
Austenitic stainless (304, 316)	GTAW	about 100–200°C (200–400°F)	about 300°C (570°F)	Sensitization and grain growth
Duplex stainless (2205, 2507)	GTAW / SMAW	about 100–200°C	about 250–300°C	Ferrite-to-austenite balance and cooling rate
Aluminum alloys	GMAW / GTAW	about 60–120°C (140–250°F)	about 150°C (300°F)	High thermal conductivity and oxide film
Nickel alloys	GTAW / SMAW	about 80–150°C (175–300°F)	about 200°C (390°F)	Liquation cracking and strain-age cracking
Cast iron	SMAW / GTAW	about 200–310°C (400–590°F)	Work with the WPS	Cooling gradient and graphite flake cracking

Keep scale in mind when applying these. A value that fits a thick, restrained carbon steel joint can be wildly excessive for a thin stainless bracket. For aluminum, even a modest preheat can be meaningful due to how quickly it pulls heat from the puddle.

How the Process Impacts Heat

The material sets the target, but the process dictates how easy it is to hit. Thermal efficiency, defined in EN 1011-1 and cited in AWS material, shows how much of the electrical energy actually reaches the joint.

Process	Thermal efficiency	Thermal character	Best fit
SMAW	0.8	Sluggish but forgiving with high heat per pass	Thick carbon steel, field work, out-of-position
GMAW	0.8	High deposition with high heat input	Production steel fabrication, long joints
FCAW	0.8	High heat and penetration, similar to GMAW	Heavy structural and ship work
GTAW / PAW	0.6	Lowest heat input with precise control	Thin sections, stainless, aluminum, nickel alloys
SAW	1.0	Highest deposition rate and heat input per pass	Long, thick seams where productivity is key
Pulse and advanced processes	Varies	Controlled pulses to limit total heat	Thin or heat-sensitive alloys, distortion control

With SAW at 1.0 efficiency and GTAW/PAW at 0.6, process selection can be just as impactful as preheat. For thick, dirty, or out-of-position steel, why stick welding still wins on thick, dirty, out-of-position steel often comes down to practicality over peak efficiency.

Material-Specific Considerations

Aluminum: Conducts heat roughly ten times better than steel. This makes cold lap and poor fusion common if prep is lacking. Higher amperage, faster travel, and thorough cleaning are required. For a starting point, reference the practical guide to aluminum MIG welding and use a proper aluminum cleaning routine.

Stainless steel: Its biggest threat is chromium carbide precipitation between 800–850°F (425–455°C), which robs corrosion resistance. For this reason, capping the interpass temperature is often far more critical than a high preheat minimum.

Duplex stainless: Maintaining a balanced ferrite and austenite count relies heavily on controlled heat input and cooling rate. Keeping passes tight and allowing enough time between them is a procedural requirement for most critical duplex joints.

Cast iron: A preheat between 200–310°C (400–590°F) helps shrink the steep temperature gradient that would otherwise promote cracking through its brittle, non-ductile matrix.

Quenched-and-tempered steel: These grades demand both a minimum and maximum preheat. Too cold risks a hard, crack-prone HAZ, while too hot overtempers it and causes a measurable loss of the strength it was treated for.

The Cost Factor

Preheat isn’t free. Oxy-fuel, induction, and resistance heating all add fuel, electricity, and clock time. As noted by the AWS Welding Digest on preheat and interpass temperature, alternative heating methods, cleaner low-carbon steels, and H4/H8 low-hydrogen filler metals can cut those costs, but only when applied correctly. For any quote, those numbers belong in the estimate just as much as they belong in the weld book.

Heat, Fume and the Welder: Safety Tied to Temperature

Two arcs run in a welding bay: the thermal arc and the exposure arc. The heat you control at the joint sets the fume leaving the puddle and how hard the welder’s lungs work at shift’s end. Temperature control is health control.

What Welding Fume Actually Contains

Welding fume is a fine aerosol of metal oxides condensed from vaporized base metal, filler metal, coatings and plating. The OSHA fact sheet on controlling hazardous fume and gases during welding is blunt about chromium: in stainless steel, nonferrous alloys and chromate-coated or chrome-plated materials, welding converts it to hexavalent chromium, Cr(VI), which is highly toxic, damages eyes, skin, nose, throat and lungs, and causes cancer. It falls under 29 CFR 1910.1026 (general industry) and 1926.1126 (construction).

Manganese and iron oxide follow, and the limits show how far the enforceable standard sits from the health-based one.

Contaminant	OSHA PEL	NIOSH REL	ACGIH TLV
Hexavalent chromium	5 µg/m³ (action level 2.5)	0.2 µg/m³	0.2 µg/m³
Manganese fume	5 mg/m³ ceiling	0.2 mg/m³	0.02 mg/m³ respirable
Iron oxide fume	10 mg/m³	5 mg/m³	Not listed

Read the manganese row twice: a shop can sit under the enforceable OSHA ceiling and still be twenty-five times over the NIOSH limit and past the ACGIH threshold. The legal limit is not the safe limit.

Manganese is why NIOSH and CDC treat welding fume as a neurological hazard too. Prolonged exposure can cause manganism, a Parkinsonian syndrome marked by tremors, slowed movement, rigidity and poor balance, and the damage can persist after exposure stops. AWS training resources cover this in welder qualification prep.

Ventilation Does the Heavy Lifting

NIOSH’s engineering controls database names local exhaust ventilation the primary control for welding, and its criteria document reports roughly a 40% increase in relative risk of lung cancer among welders. Dilution ventilation cannot keep pace with a rising fume plume. The workable setups:

• Local exhaust hoods close to the arc, pulling the plume before it spreads.
• Flexible arm hoods repositioned as the joint changes.
• Enclosed booths with negative pressure, standard for high-volume stainless and galvanised work.
• Scrubbers or filtered units where capture alone cannot clean the breathing zone.

Engineering control outranks respirators, so the respirator is the last line, not the plan.

Heat Stress on the Welder

Now add the welder’s own body heat: flame-resistant clothing over a hot plate, an arc, a stagnant booth, a July afternoon and an oxy-fuel torch running for preheat. OSHA’s heat guidance stresses hydration, rest, shade and, critically, acclimatization of new workers. Heat illness is preventable, and still a leading cause of workplace death.

Preheat makes it worse: a large plate at 150 °C radiates heat as well as conducting it, and cooling the plate between passes does not cool the person. Morning scheduling, induction or resistance heating for bulk preheat instead of an open torch, and hydration breaks built into the weld sequence are low-cost responses.

PPE That Matches the Heat

Heat protection and fume protection are different kit, and shops often blur them.

• Helmet shade matched to process and amperage: a shade picked for thin material is under-tinted on a hot 300 amp SAW pass, forcing squinting and strained eyes. The advantages of stick welding guide covers SMAW shade selection.
• Flame-resistant clothing, sleeves down: spatter and radiant heat burn ordinary clothing on the forearm.
• Respirators where ventilation cannot hold the breathing zone below the action level, or where a fit test and written program exist. On stainless, galvanised or chromate-plated work, assume respirators until a measurement says otherwise.
• Eye and face protection: safety glasses under the helmet for grinding, a face shield for nearby cleanup.
• Insulated gauntlets: a plate at 300 °C burns through a glove in seconds.

Preheat burns are a leading cause of first aid in fabrication shops, and they happen during handling, not welding, so heated plate should be marked, barricaded and cooled in a controlled area. Less fume, less distortion and shorter cycle time is the argument that wins the budget.

On aluminum the fume story is lighter than stainless, but the thermal logic holds: the metal pulls heat out of the joint fast, so clean oxide removal and preheat limits in the aluminum flux core welding breakdown matter more.

The Practical Playbook: A Step-by-Step Welding Temperature Checklist

Temperature control is not one decision at the torch. It is a sequence of decisions made before, during and after the arc strikes. Print this and tape it to the booth.

Before the Arc

Most weld cracks are lost before the arc ever lights, right here in the prep.

1. Identify the material grade and the real thickness, not the nominal print dimension. Measure the thickest member at the joint, since that number drives the preheat rule.
2. Find the preheat and interpass numbers that apply. The traveler, the WPS and the code inspection plan govern; if they disagree, contract documents win.
3. Confirm the electrode classification and diffusible hydrogen level. E7018 is low hydrogen; E6010, E7010 and E8010 are cellulosic with far higher diffusible hydrogen. On restrained, high-carbon joints, cellulosic rods without adequate preheat grow cold cracks, and the rod from the oven has to match what the WPS calls for.
4. Check the starting plate temperature with an instrument, not your palm. Plate at 40°F and plate at 120°F are different welding problems.
5. Clean the joint. Scale, oil, paint and moisture change the picture, and on aluminum so does oxide. Oxide removal is not optional; see our guide to preparing aluminum for welding.
6. Plan tack and sequence. Tacks count toward heat input, shrink the plate and start the distortion.
7. Set up heating equipment and instruments. Torch, mat or induction heater ready; pyrometer or thermocouple calibrated; a spare instrument for critical parts. Unmeasured preheat is guesswork, and guesswork is what the Duluth bracket had.

During Welding

• Measure and log the preheat at the required distance. AWS D1.1 requires the minimum preheat to be established at a point at least equal to the thickness of the thickest member and not less than 3 in. (75 mm) from the point of welding in all directions. Measure on the base metal at the joint, not across the fire.
• Hold the arc length and voltage the WPS specifies. A wandering arc inflates voltage, spreads the heat and wrecks your heat input.
• Control travel speed deliberately. Heat input is inversely proportional to travel speed, so creeping along the joint quietly doubles the heat you deliver.
• Keep interpass temperature under the cap, checked before tacking the next pass. On stainless and nickel this is a hard limit, because the metallurgical damage is done once you exceed it.
• Do not weave wider than the qualified width. Weaving lowers heat per unit length but widens the heat-affected zone, and a wider weave takes the joint outside its qualification.
• Check the root and clean between passes. Slag, spatter and dross left in place are inclusions waiting to happen, and they also tell you the arc was too long.

After Welding

Do not walk away when the arc goes out. Delayed cracking often appears overnight or on the next shift. Cold cracking needs time, and the night supplies plenty.

• Watch for delayed cracking. Leave the joint visible, and treat any new crack or tear as a stop-work event until it is evaluated.
• Hold or slow-cool when the WPS requires it. If yours calls for a delayed start or slow cool to keep hydrogen moving out of the HAZ, set a timer and honor it.
• Plan the NDT so a rejected weld is found before the assembly ships.
• Record the temperatures used: preheat reading, interpass readings, instrument used, time and welder. Most code inspection plans expect it, and it is your only defense when a joint is questioned months later.

Mistakes to Avoid

• Skipping preheat on thick or restrained joints because “it doesn’t look cold.”
• Measuring temperature on the arc-lit surface, where the reading is meaningless.
• Letting the torch cool the joint before striking the arc, undoing the preheat.
• Exceeding the maximum interpass temperature on stainless or nickel alloys.
• Running cellulosic electrodes (E6010, E7010, E8010) on restrained high-carbon steel without a qualified low-hydrogen alternative.
• Welding with damp electrodes; moisture and rutile coatings on cellulosic rods feed hydrogen into the HAZ.
• Confusing “warm to the touch” with “code-compliant.” A plate can feel cold and be fine, or feel warm and be short.
• Relying on the WPS without confirming the material grade matches the joint.
• Mixing processes under one WPS. A WPS written for FCAW does not cover GMAW.
• Heating the plate unevenly with a torch, buying a little crack resistance and a lot of distortion.

Conclusion

Electric welding temperature is not one number. It is a system: arc temperature, base metal temperature before you start, the interpass ceiling, and the heat input tying them together. Control the system and the weld passes; miss one and the joint may still fail months later, exactly like that split fillet in Duluth.

Start with the WPS, the traveler and the pyrometer, and treat the numbers as numbers, not impressions. Go to the source for more: AWS education programs cover metallurgy, inspection and codes, the AWS Welding Digest article on preheat and interpass temperature is the shortest read on the code side, and OSHA’s welding fume fact sheet with NIOSH’s welding health and safety resources cover the health side of the heat. The AWS D1.1 text governs; the hosted 2000 edition is background reading only.

Frequently Asked Questions

How hot does an electric welding arc actually get?

Arc column temperatures are commonly cited around 6,000–10,000°F (3,300–5,500°C) at the core, which is far hotter than the melting point of the electrode or the workpiece. That heat is concentrated in a tiny column, so the base metal only warms modestly. What matters for weld quality is the temperature of the steel near the joint, not the temperature of the arc.

What temperature should I preheat mild steel to before welding?

For ordinary structural carbon steel, many codes and prequalified WPSs call for a minimum preheat around 50°F (10°C) for material up to about 3/4 in. thick and 70°F (21°C) above that, while higher-strength and heavily restrained joints need considerably more. Always follow the WPS and the contract documents for the specific grade, thickness and joint. Low-hydrogen electrodes such as E7018 reduce the need for preheat on thin, unconstrained sections.

What is the difference between preheat temperature and interpass temperature?

Preheat is the minimum temperature the base metal must reach before the arc is struck. Interpass is the maximum temperature the joint may reach between passes. Preheat speeds up heating; interpass limits stop the joint from getting so hot that the HAZ over-coars, over-softens, or — in stainless and nickel alloys — loses corrosion resistance or becomes crack-prone.

How do I calculate heat input when welding?

Use H = (60 × V × A) / (1000 × travel speed), with travel speed in mm/min for kJ/mm or in in/min for kJ/in. A 24 V, 110 A pass at 140 mm/min gives about 1.13 kJ/mm (28.7 kJ/in). Then note that ASME Section IX reports raw arc energy while ISO 15614-1 and EN 1011 apply a thermal efficiency factor (0.8 for SMAW/FCAW/GMAW, 0.6 for GTAW/PAW, 1.0 for SAW), so the same pass can legitimately show two different numbers.

Why did my weld crack hours after I finished it?

That timing is the signature of hydrogen-assisted cold cracking. The weld cooled into a hard, brittle heat-affected zone while residual tensile stress and diffusible hydrogen from the electrode coating were still present. Prevention is the classic triad: low-hydrogen consumables, adequate preheat for the material and thickness, and low joint restraint. Verify the electrode’s storage and drying, and check your preheat records.

Do stainless steel and aluminum need preheat?

Usually not for the same reason steel does. Stainless steel is used for corrosion resistance and temperature limits rather than to avoid hydrogen cracking, though interpass caps of roughly 150–300°C protect grain size and corrosion properties. Aluminum is a very different case: it conducts heat away rapidly, so high amperage and fast travel are needed to maintain a proper puddle, with any preheat limited to a low 60–120°C range to avoid distortion.

Can I weld when the plate is cold or the weather is below freezing?

Cold steel is one of the most common causes of unexpected cracking, because the cooling rate through the transformation range increases sharply. Many specifications set a minimum starting metal temperature of50°F (10°C) or 70°F (21°C) for common carbon steels and prohibit welding below0°F (-18°C) in some cases. Always check the contract, and preheat to the required minimum before striking the arc.

Is a thermocouple or an infrared thermometer better for checking preheat?

A thermocouple (usually Type K) is the most accurate and repeatable choice and is what you want for code compliance and documentation. Infrared thermometers are fast and convenient for spot checks and gradient mapping, but they need correct emissivity and must never be aimed at the arc or at shiny molten metal. Temperature-indicating crayons remain a cheap, effective way to confirm a preheat target was reached.

Does using more preheat always make a better weld?

No. More preheat slows cooling and reduces cracking risk in hardenable, restrained joints, but it also enlarges the heated zone, increases distortion, adds labor, fuel and electricity cost, and can overheat the HAZ of quenched-and-tempered steels. Carbon steel tolerates some overshoot above the minimum, but Q&T steels need both a maximum and a minimum preheat that are followed closely.