Preheating Before Welding: Why It’s Done and How to Do It

The weld passed visual inspection on Friday. On Monday morning, a sharp ‘tink’ from the fabrication bay announced a crack running out of the heat-affected zone on a 1-inch A572 column flange — welded in a shop that had dropped to 38°F overnight. Nothing about that weld looked wrong. What was missing happened before the arc ever started.

Preheating Before Welding Is Done To… What, Exactly?

The weld passed visual inspection on Friday afternoon. On Monday morning, a sharp tink from the fabrication bay announced a crack running out of the heat-affected zone on a 1-inch A572 column flange — welded in a shop that had dropped to 38°F overnight. Nothing about that weld looked wrong. The bead was straight, the profile was good, the inspector had signed off. What was missing happened before the arc ever started.

So what is preheating before welding actually done to? In one sentence: preheating raises the base metal around the joint so the weld and heat-affected zone cool slowly enough to avoid forming hard, brittle microstructures, to give trapped hydrogen time to escape, and to soften the thermal shock that locks residual stress and distortion into the part. That’s the whole answer in three clauses. Everything else in this article is the working-out.

Those three effects matter because welding is a violent, localised thermal event. You dump thousands of joules into a narrow groove while the surrounding plate — often dozens of times more massive — sits at ambient temperature and acts like a heat sink. The metal that was molten cools in seconds, and the metal just beside it, the heat-affected zone (HAZ), passes through the critical transformation range fast enough to harden instead of soften. Preheat simply gives that cooling a head start in the other direction.

A Working Definition, Borrowed From the Codes

Precise language matters here, because “preheat” gets used loosely on the shop floor. In AWS terminology — AWS A3.0, as quoted in the commentary to AWS D1.1/D1.1M — preheat is the temperature of the base metal in the volume surrounding the point of welding immediately before the arc is started. Note what that definition does and does not say. It measures the base metal, not the flame, not the air, not the electrode. It measures it in the volume surrounding the joint, not just the faying surface. And it measures it immediately before the arc starts, which is why a temperature taken when the torch was pulled away ten minutes ago doesn’t count.

What Preheat Is Not

This is where a lot of good intentions go wrong. Preheat is a thermal tool, and it fixes thermal problems. It is not a substitute for:

• Clean metal. Mill scale, oil, paint, galvanising, and rust all feed hydrogen, oxygen, and porosity into the puddle. A rosebud torch does not remove them — it bakes them in.
• Dry consumables. A warm plate paired with a damp electrode is still a hydrogen delivery system. Low-hydrogen electrodes need proper storage and, for critical work, holding ovens. The selection side of that is covered in this guide to alloy steel welding electrodes.
• Correct heat input. Preheat and heat input are partners, not replacements. A high preheat with a fast, low-energy pass can still quench hard. Conversely, dumping excessive current to “compensate” for a cold joint overheats the HAZ and coarsens the grain.
• Good fit-up and sequence. Preheat doesn’t cure a 3/8-inch root gap or a weld sequence that stacks stress into a restrained corner.

Think of preheat as one leg of a three-legged stool — the other two being low-hydrogen practice and controlled heat input. Remove any leg and the stool falls over, no matter how hot the plate is.

Where This Article Goes From Here

We’ll move from metallurgy to practice. Next we’ll open up the metal itself and look at what happens when the cooling curve passes through the transformation range, and why carbon equivalent and plate thickness decide how much preheat a joint really needs. Then we’ll break the “why” into the four failure modes preheat fights — hydrogen cracking, excessive hardness, residual stress, and cold or damp conditions. From there we get practical: the actual numbers in AWS D1.1, ASME, and API; the equipment options from rosebud torches to induction coils and ceramic pads; how to measure and hold temperature honestly; a material-by-material playbook; the mistakes that keep showing up in failed inspections; and finally the cost and workflow questions, including when skipping preheat is genuinely defensible.

One theme runs through all of it. Preheat is not decoration added to a procedure to satisfy a checklist. It is the cheapest control you have over the metallurgy that decides whether a weld survives its first cold weekend.

What Actually Happens Inside the Metal When You Preheat

Preheat is often called “taking the chill off” steel, but that sells it short. At the microscopic scale, it is a cooling-rate control tool. It changes what the metal becomes after the arc moves on, which is why a weld can look perfect on Friday and crack on Monday.

When the arc passes, a narrow band of base metal beside the weld is heated above the steel’s transformation temperature. The weld pool solidifies, and everything nearby begins to cool. Weld metal and the heat-affected zone (HAZ) do not cool at the same rate or evenly through the section. The outcome depends on how quickly that hot metal crosses the critical temperature range.

A Hot Bearing, a Bucket of Water, and a Very Short Decision

Imagine a bearing heated to 1,500°F. Plunge it into water and it comes out hard, stressed, and brittle. Let it cool in still air and it comes out softer and tougher. Welding is that experiment on a smaller scale, except the quench medium is the surrounding plate. Preheat warms that quench medium first. A warmer plate pulls heat from the weld more slowly because the temperature difference between them is smaller.

For steel, the critical cooling range is roughly 800°C down to 500°C (1,472°F to 932°F). Passage through that window decides much of the final microstructure. Fast cooling can produce hard martensite; slower cooling gives carbon and alloying elements time to form softer ferrite, pearlite, or bainite, depending on the steel. TWI’s overview of heat treatment in welded joints explains how cooling rate and transformation behavior are linked.

Condition	Relative Preheat Demand	What Drives It
Thin section, low CE (roughly ≤ 0.30)	Low	Small heat sink, limited hardenability
Moderate section (about 1/4–1 in), CE 0.30–0.45	Moderate	More heat extraction, higher chance of hard HAZ
Thick section (over 1 in) or CE above 0.45	High	Rapid heat loss plus easily hardened steel
High restraint, repair weld, or cold ambient	High	Stress plus fast cooling

That table is a conceptual map, not a procedure. The actual number must come from the welding procedure specification (WPS) and the governing code.

Why Thick Plate Makes the Problem Worse

Thick steel acts like a heat sink. The arc puts heat into a small volume, but the surrounding mass pulls it away almost immediately. A one-inch flange drains heat far faster than a 3/16-inch sheet, so the weld and HAZ cross the critical range more quickly. That is why required preheat generally rises with thickness for the same grade, and why a heavy column flange can behave very differently from thin plate.

AWS D1.1, the structural steel welding code, covers carbon and low-alloy steels 1/8 in (3.2 mm) and thicker with yield strength up to 100 ksi (690 MPa). That scope includes a lot of work where this heat-sink effect matters. The AWS D1.1 preview document shows the code’s scope and structure.

Aluminum behaves differently. Its thermal conductivity is several times that of steel, so heat drains away faster and distortion becomes a bigger concern. Our aluminum MIG welding guide covers how heat control and preparation change for those jobs.

Hardenability and Carbon Equivalent: The Steel’s Personality

Two steels can have the same thickness and respond very differently to welding. Hardenability is the steel’s tendency to form martensite when cooled. Carbon has the strongest effect, but manganese, chromium, molybdenum, vanadium, and other elements matter too. Engineers combine those effects into the carbon equivalent, or CE. A higher CE generally means a greater risk of hard, crack-sensitive microstructures.

That risk is not theoretical. In an EWI/DOT study of cracked pipeline girth welds, measured carbon equivalents reached CE = 0.66, and repair costs on one offshore project exceeded $10 million. When steel is that hardenable, a preheat number copied from a generic table may not be enough.

The HAZ Is Usually the Weak Link

The HAZ is the band of base metal that got hot but did not melt. Its chemistry is essentially unchanged, yet its microstructure has been transformed. It also sits next to the weld, where restraint and stress concentrate. Near the fusion line, grain growth can occur, and coarse grains generally mean lower toughness. That combination of a hard, coarse, stressed band beside a rigid weld is why cracks so often start in the HAZ rather than the weld metal.

The Failure Triangle: Hard Microstructure, Hydrogen, and Stress

Hydrogen-assisted cold cracking needs three conditions at once: a hard, susceptible microstructure; diffusible hydrogen in the weld; and tensile stress, applied or residual. Remove any leg and the crack cannot complete its triangle. Preheat attacks the first two by slowing cooling and giving hydrogen time to escape, and it reduces the thermal gradient that feeds residual stress.

More Heat Is Not Automatically Better

Because preheat helps, it is tempting to think more is better. It is not. Excessive heat can coarsen grains, reduce toughness, soften quenched-and-tempered steel, and increase distortion. Codes therefore set a minimum preheat and cap interpass temperature. The goal is a controlled thermal window, not the hottest joint in the shop. A good WPS defines both ends; the welder stays inside them.

The Four Problems Preheat Solves: Hydrogen, Hardness, Restraint, and Cold

Ask ten welders why they preheat and you will get ten versions of the same answer: to keep the weld from cracking. That is true, but it is not the whole story. Preheat is one shop-floor action that addresses four different failure mechanisms at once. Understanding them separately is what turns a number on a WPS into a decision you can defend.

1. Hydrogen-Assisted Cold Cracking

Hydrogen is the smallest atom in the weld and the most patient. It enters through moisture on the plate, rust, primer, grease, humid air, or a damp electrode coating. After the arc passes, hydrogen diffuses through the weld metal and heat-affected zone. If the microstructure is hard and tensile stress is present, it collects at imperfections and starts a crack. The failure often appears hours or days later, long after visual inspection, which is why hydrogen cracking has ruined so many Monday mornings.

Preheat attacks the problem from two directions. It slows cooling, giving hydrogen more time to diffuse out of the joint before the weld reaches room temperature. It also dries the surface, driving off condensation and moisture, including the water vapor produced when a fuel-gas flame burns. Preheat helps, but it does not replace dry low-hydrogen consumables. If you are selecting electrodes for a sensitive steel, start with alloy steel welding electrodes and a sealed, dry supply.

2. Excessive Hardness and Loss of Toughness

Cooling speed decides microstructure. A weld that cools quickly through the transformation range can form martensite, a hard, brittle phase that is far more crack-sensitive than ferrite and pearlite. Thick plate makes this worse because it acts as a heat sink, pulling heat out of the weld fast. Preheat raises the starting temperature and reduces the difference between the weld and the surrounding steel, so the joint cools more slowly through roughly 800°C to 500°C, the range where much of that transformation occurs (TWI heat treatment guidance). The result is a softer HAZ, better toughness, and more tolerance for welding stress. Preheat cannot rescue bad heat input, and on quenched-and-tempered steel too much preheat can damage properties, which is why maximum interpass matters as much as the minimum.

3. Residual Stress and Restraint

Every weld shrinks as it cools. In a restrained joint, that shrinkage is fought by the surrounding structure, and the resulting tensile stress can be high enough to drive cracking even in a sound microstructure. Preheat reduces the thermal gradient between the hot weld and the cold plate, so contraction is more gradual and peak residual stress is lower. It also gives the welder a wider working window.

Restraint is where code minimums can be misleading. A heavily restrained repair or a thick, boxed-in connection may need more preheat than the table value because restraint adds another leg to the failure triangle. AWS D1.1 also requires the preheat to be maintained for a distance at least equal to the thickness of the thickest part, and never less than 3 in (75 mm), in all directions from the weld (AWS D1.1). A narrow hot strip at the joint does not satisfy that intent.

4. Cold, Moisture, and Frost at the Joint

Cold, damp conditions add two problems. First, a cold plate accelerates cooling, pushing the weld toward a harder microstructure. Second, low temperature encourages condensation and frost on the steel, and that water is a direct hydrogen source. Preheat brings the base metal above the dew point, dries the joint, and removes the thermal shock of starting an arc on near-freezing steel. The flame itself can add moisture, so the heat must be applied long enough to drive water off, not just color the surface. A warm, dry joint is one less variable working against you.

These four problems explain why “preheating before welding is done to” has more than one ending. A single 250°F preheat may dry the joint, slow the cooling, soften the HAZ, and reduce residual stress all at once. That is also why copying a number without understanding the reason behind it is risky. The number serves the mechanism, not the other way around.

The Numbers: What AWS D1.1, ASME, and API Actually Require

Ask ten welders where the 150°F on a WPS came from and you’ll get ten shrugs. Nobody invented that number at the bench. It traces back to a table, a formula, or a qualified procedure test — and knowing which one matters when an inspector asks why you did what you did.

How a WPS Gets Its Preheat Number

There are only three routes. Route one is a prequalified table lookup, which is what most shop WPSs use because it’s fast, conservative, and easy to defend. Route two is an engineering calculation based on chemistry and hydrogen level, permitted by code annexes. Route three is procedure qualification testing, where you prove a number works by welding and breaking specimens.

The WPS governs the job. Code tables set a floor, never a ceiling. If a drawing or specification says 250°F minimum and the table says 50°F, you weld at 250°F.

Inside AWS D1.1 Table 5.8

AWS D1.1 organizes its prequalified minimum preheat and interpass temperatures by steel category (A through G), thickness range, and welding process. To read it, you find the category of your base metal, locate the thickness of the thickest part at the joint, note your process, and read the cell. That cell gives you both the minimum preheat and the minimum interpass temperature — they’re the same number in the prequalified tables.

One source of endless confusion: older editions numbered this table 3.2. If you’re holding a pre-2015 copy and the reference doesn’t line up with the AWS D1.1/D1.1M preview, that’s why. Same table, new address.

The categories also carry a hydrogen story. Higher-strength steels get pushed into stricter categories where the code assumes you’re running drier consumables. The 2025 edition added Category G for A913 Gr 80, limited to H4 electrodes — just 4 mL of diffusible hydrogen per 100 g of deposited metal (AWS Welding Digest, June 2025). As yield strength climbs, allowable hydrogen falls.

Condition	Prequalified minimum
A572 Gr 50 or A992, FCAW, over 1-1/2 in through 2-1/2 in thick	150°F preheat and interpass
A913 Gr 50 and Gr 65 (Category D), ambient above 32°F, H8 or drier electrode	No preheat required
A913 Gr 50 and Gr 65 (Category D), ambient below 32°F	70°F preheat
A913 Gr 80 (Category G, new in 2025)	H4 electrodes only
Any steel below 32°F ambient	Heat to at least 70°F before welding

The Alternative Route: Annex B and Pcm

AWS D1.1 Annex B offers an alternative preheat method built on four inputs: the composition parameter Pcm, the filler metal’s diffusible hydrogen level, thickness, and restraint. Instead of looking up a conservative category, you calculate from actual heat chemistry. Run low-Pcm steel with H4 or H8 consumables at modest thickness and restraint, and the calculated minimum often lands below the prequalified table — sometimes far below.

The catch is real chemistry. Pcm needs the mill certificate, not a guess that “it’s probably A36.” Substituted steel without paperwork will sink the calculation. If you’re chasing hydrogen levels, alloy steel welding electrodes selection and safety covers how H4, H8, and H16 classifications differ in practice.

ASME, API, and the CE/CET Methods

ASME B31.3 sets minimum preheat by material group and thickness in its own table, and Section VIII Division 1 handles preheat alongside its postweld heat treatment rules. API 1104 for pipeline work leans harder on the qualified procedure, with ambient and moisture controls layered on top rather than one universal table.

Outside the US, AWS D14.8M and ISO/TR 17844 provide calculation methods using carbon equivalent (CE) and the CET formula to estimate the minimum preheat that avoids hydrogen cracking. These are the tools an engineer reaches for when the table is too blunt for the joint in front of them.

Putting the Numbers to Work

Picture a 1-3/4 in A992 column flange welded with FCAW in January. The table says 150°F. The shop reads 29°F at 6 a.m. So the flange gets heated to at least 70°F before anything else, then brought to 150°F minimum, and held there through every pass. A 1/4 in A36 stiffener welded in July with the same process? Much lower number — and that difference is exactly why blanket rules like “we always preheat to 200°F” waste money while “we never preheat” cracks columns.

Code Minimums Are Floors, Not Targets

Three things sit above the table: the WPS, the project specification, and the engineer’s judgment on restraint, section thickness, and repair welds. Any of them can raise your number. None of them can lower it without a documented alternative method. Read the table to understand the floor. Read the WPS to know what you actually have to hit.

How to Preheat: Torch, Induction, Resistance, Ovens, and Blankets

Once the WPS gives you a number, the next question is practical: what do you use to bring a thick joint to temperature without turning one corner into a hot spot? AWS D1.1 sets the temperature, not the hardware, so several methods are legitimate. Each has a personality, and each can create a new problem if used carelessly.

Oxy-Fuel Torches: Cheap, Portable, and Easy to Overdo

A rosebud tip on an oxy-fuel torch, fed from a properly rated fuel-gas supply, is the oldest and cheapest option. It is portable, needs no power, and works in places a heater cart will never fit. It is also the least even. A single flame puts a steep gradient into the plate, and the operator’s wrist becomes the temperature control.

The hidden issue is water. Fuel-gas combustion produces water vapor, which can condense on cooler plate several inches from the flame until the metal is hot enough to drive it off. That moisture can sit exactly where you plan to weld. Keep the flame moving, use a soft bushy flame rather than a sharp jet, and give the plate time to dry before welding.

Flame impingement is the other trap. Holding a hot flame on one spot can gouge, scale, or locally soften the steel long before the surrounding plate is warm. Distance and motion matter as much as tip size.

Induction Heating: Fast and Repeatable

Induction heating uses an alternating magnetic field to heat the steel itself. It is fast, repeatable, and easy to control, which makes it a favorite for pipe and thick plate. Coils can be shaped around a joint, and the heat stays where you put it. The limits are equipment cost and coil fit: a coil that does not follow the joint will leave cold bands.

Electrical Resistance Pads: The Shop and Field Workhorse

Electrical resistance heaters use ceramic pads wired to a controller. They are common in both shop and field because they wrap around complicated shapes, deliver a wide even band, and hold a set temperature far more consistently than a flame. They are slower to start and need careful cable management, but for code-quality preheat they are hard to beat.

Furnaces and Ovens: Whole-Part Heating

Furnaces and ovens heat the whole part, not just the joint. That is the right choice for castings and complex fabrications, because slow, uniform heating avoids the unequal expansion that can itself crack a part. The trade-off is size and logistics: the part must fit, and it must be moved hot.

Insulating Blankets and Thermal Covers

Insulating blankets and thermal covers do not create heat. They hold it. Used after preheating, they slow cooling and help maintain interpass temperature; used during heating, they reduce losses on large parts. They are cheap insurance, but they are not a substitute for a heat source.

Heat-Soak Time, Heating Rate, and Heating Both Sides

A surface reading is not a through-thickness reading. Thick plate needs heat-soak time so the core catches up with the face, and the joint should be heated from both sides when access allows. TWI guidance for locally applied preheat calls for the heated zone to extend at least 75 mm from the weld, and preferably to be measured on the face opposite the one being welded.

Heating rate matters too. Thick sections, castings, and complex parts should be brought up slowly and uniformly, because a fast local rise creates its own thermal stress. Follow the WPS or an approved procedure rather than racing the clock.

Choosing a Method: A Quick Comparison

Use this as a starting point, then let the WPS and the part geometry decide.

Method	Best for	Strength	Watch out for
Oxy-fuel torch	Field work, small joints, repairs	Cheap, portable, no power	Uneven heat, combustion moisture, flame damage
Induction	Pipe, thick plate, repetitive work	Fast, controllable, repeatable	Coil fit, equipment cost
Resistance pads	Shop and field fabrications	Uniform band, good control	Slow startup, cable and electrical safety
Furnace or oven	Castings, complex parts	Whole-part uniformity	Size limits, handling hot parts
Insulating blankets	Holding temperature, slow cooling	Low cost, simple	No heat input on their own

The best method is the one your crew can apply evenly, measure honestly, and repeat on every joint. Fancy equipment does not fix a rushed preheat.

Measuring and Holding Preheat and Interpass Temperature

Everything before this point is preparation. The number on the WPS only becomes real when someone proves the steel is at temperature, at the joint, at the moment the arc starts. A rosebud torch roaring over a joint proves heat went somewhere. It does not prove the plate hit 150°F, stayed there through the root, and never climbed past the maximum on the cap.

Measurement is also where paperwork meets reality. An inspector can watch you weld but cannot see whether the plate was 40°F short. The instrument, the location, and the timing of the reading are the evidence.

Pick the Right Instrument for the Job

Tool	How it works	Best for	Watch out for
Temperature-indicating crayon (Tempilstik)	Mark melts at its rated temperature	Quick field checks, tack welds	Surface only; scale gives false marks
Contact pyrometer	Spring-loaded probe touches the steel	Shop and field compliance checks	Must touch bare metal; hold until stable
Thermocouple	Probe attached or clamped to the part	Critical joints, thick sections, logged records	Setup time; placement must be deliberate
Infrared thermometer	Reads emitted radiation	Fast screening, hot spots	Emissivity errors on shiny or scaled steel

Temperature-indicating crayons are the simplest tool in the box. Mark clean base metal near the joint and watch for the mark to melt within seconds. They are cheap, pocket-sized, and never drift out of calibration. Their weakness is resolution: the mark tells you the surface reached its rated temperature, not how far past it you went. Grind or wire-brush first, because mill scale insulates the mark from the steel.

Contact pyrometers are the everyday workhorse for code work. A spring-loaded probe pressed against clean metal reads in seconds, provided you hold it until the number stops climbing. A probe resting on loose scale measures the scale, not the plate.

Thermocouples clamped to the part are the gold standard for thick sections and critical repairs, because they catch the slow equalization a spot check misses.

Infrared thermometers are tempting: instant, and safe at a distance. The problem is emissivity. Shiny mill scale, rust, and oxidation change how much radiation a surface emits, so the same 200°F plate reads differently depending on its finish. That is why many inspectors refuse an IR gun as sole verification. Screen with it, then confirm by contact.

Where and When to Measure

Measure base metal, not the bead or the puddle. On accessible joints, check both faces at least one material thickness or 3 inches (75 mm) from the joint edge, on the side ahead of the arc. On thick plate, take several readings along the joint length. One happy number at one end does not describe a twelve-foot flange.

AWS D1.1 Clause 5.6 requires preheat and interpass temperature to be checked just before the arc is initiated for each pass. Timing matters. A reading taken the moment the torch comes off proves the surface is hot while the through-thickness steel is still catching up.

A practical field rule is to wait about one minute per inch (25 mm) of thickness before checking on the same side, so heat equalizes instead of rewarding you for a hot skin.

Holding Interpass Without Crossing the Ceiling

On most WPSs the minimum interpass temperature equals the minimum preheat. A separate maximum must also be listed and enforced just as seriously. Too much heat can coarsen grains, reduce toughness, and in quenched-and-tempered steel soften the properties preheat was protecting.

Short joints are the usual offenders. A long seam has enough surrounding mass to shed heat between passes; a six-inch stiffener weld does not, and interpass can climb past the maximum in two or three passes. Reduce current, spread the passes, allow air cooling, or change the sequence. Check between passes instead of assuming the number still holds.

If welding stops and the joint drops below the minimum preheat, it must be reheated before welding resumes. A lunch break in a cold bay is an interruption like any other. For related background, see TWI’s guidance on heat treatment of welded joints.

Aluminum sheds heat from the joint fast, which changes how preheating before welding is done and how often you re-check it. Our aluminum MIG welding guide covers the practical side.

Documentation and a Field Checklist

Note: Inspectors want a record, not a recollection. Log the instrument ID and calibration date, the reading, its location on the joint, the pass number, the time, ambient temperature, and who took the measurement. A photo of the gauge against the marked joint is cheap insurance.

The gauge does not make the weld better. It makes the weld honest.

Preheat by Material: Carbon Steel, HSLA, Cast Iron, Stainless, and Aluminum

Steel is not one material, and neither is cast iron or aluminum. A preheat that protects a thick carbon steel column can ruin the temper of a quenched-and-tempered plate, while an austenitic stainless joint may need no preheat at all. The practical rule is simple: identify the alloy, read the procedure, then choose the heat.

Carbon and Low-Alloy Structural Steel

This is where most code tables live. For prequalified welds, AWS D1.1 Table 5.8 sets minimum preheat by steel category, thickness, and process. Those values are deliberately conservative. For joints outside the table, Annex B methods use composition, diffusible hydrogen, thickness, and restraint to calculate a number. Carbon equivalent is the shorthand: more alloying elements mean more hardenability and usually more preheat. Thick sections make it worse because they pull heat away from the arc. Quenched and tempered grades add another rule: control heat input as well as preheat so the original temper is not destroyed. The WPS governs every time.

HSLA and High-Strength Grades

HSLA and high-strength low-alloy steels offer better strength-to-weight ratios, often with lower carbon. Lower carbon does not mean no risk. These grades are sensitive to hydrogen and to excessive heat. Over-preheating can soften the heat-affected zone and reduce toughness, and too much heat input can do the same. Use low-hydrogen consumables such as H4 or H8, keep them dry, and follow the WPS. For filler selection, see our guide to alloy steel welding electrodes.

Cast Iron

Cast iron is a different animal. It is brittle, and the heat-affected zone can crack if it cools quickly. TWI lists typical preheat from about 100°C to 400°C depending on iron type and allowable HAZ hardness. Lincoln Electric recommends 500–1200°F for many cast iron repairs and warns not to exceed about 1400°F, where cracking risk rises. Heat the whole part evenly where possible, weld in short passes, and cool slowly under a blanket. White cast iron is generally considered unweldable because of its massive iron carbides.

Stainless Steel

Austenitic stainless steel such as 304 or 316 generally does not use preheat. The concern is interpass temperature, not initial heat: too hot and you can lose corrosion resistance or promote sensitization. Ferrite content in the weld also matters for cracking resistance, especially in thicker sections. Martensitic and duplex grades have their own rules, so do not apply the austenitic recipe to them. Follow the WPS and control heat input.

Aluminum

Aluminum conducts heat away fast, so thick plate can be difficult to fuse. Preheat may be needed for plate roughly 3/4 in and thicker, but it should generally not exceed 350°F. The goal is to drive off moisture and reduce the chill, not to cook the alloy. Too much heat can soften heat-treated grades and make the weld worse. Clean the surface first; oxide and hydrocarbons cause porosity. Our aluminum MIG welding guide and aluminum cleaner guide cover prep and process in more detail.

Material-by-Material Summary

Material	Typical preheat	Why	Watch out for
Carbon and low-alloy structural steel	Per WPS or code table; often 50–300°F depending on thickness and CE	Slows cooling, limits HAZ hardness, helps hydrogen escape	Treating a table value as a universal recipe
HSLA and quenched-and-tempered steel	Modest, per WPS; heat input limits usually apply	Hydrogen control without damaging toughness or temper	Over-preheat and excessive heat input
Cast iron	100–400°C typical; Lincoln cites 500–1200°F	Even heating and slow cooling prevent HAZ cracking	Above about 1400°F; white cast iron is generally unweldable
Austenitic stainless steel	Generally none	Preheat is not the primary control; interpass and ferrite are	Overheating, sensitization, and applying steel rules blindly
Aluminum	Only for thick sections; generally ≤350°F	Reduces chill and drives off moisture	Softening heat-treated alloys; poor cleaning

Common Mistakes, Troubleshooting, and Safety on the Preheat Bench

Most preheat failures we see in fabrication shops are not dramatic cases of a crew ignoring the WPS. They are quieter: a joint heated only at the face, a temperature checked once before the root pass, a flame allowed to sit too long on one edge. Inconsistent or uneven thermal control causes more cracked welds than a completely skipped step. Here are the mistakes worth policing, the diagnostic path when a weld still fails, and the safety hazards that come with putting heat into steel.

Eight Mistakes That Keep Showing Up

1. “It’s only a fillet.” Fillet welds cool just as fast as groove welds, and often sit on thicker, more restrained members. A small weld in a stiff joint can crack without any visual warning. The process still needs the preheat your procedure calls for, whether you are running flat or vertical; even a simple all-position fillet deserves the same thermal discipline.
2. Heating only the joint face. A hot strip along the weld seam leaves the surrounding plate cold and acts like a heat sink. Heat a band wide enough to cover the heat-affected zone and several inches beyond it, on both sides where access allows.
3. Measuring once at the start. Preheat is verified immediately before welding, and interpass temperature is verified before each pass. One reading at 7 a.m. says nothing about the joint at 2 p.m.
4. Letting the part cool between passes. A lunch break, a fit-up fix, or a long repositioning move can drop the joint below minimum. Recheck and reheat before restarting.
5. Using a wet torch or a condensation-prone setup. Water drips, damp blankets, and hoses lying in puddles add moisture exactly where you are trying to keep hydrogen out. Dry the equipment and the surrounding steel.
6. Over-preheating quenched and tempered steel. More heat is not safer heat. Excessive preheat or interpass temperature can soften the HAZ and reduce strength even when the weld looks perfect, as heat-treatment guidance from TWI explains.
7. Letting a flame overheat one edge. A rosebud parked in one spot creates a steep thermal gradient, which drives distortion and can leave one side of the joint far hotter than the other. Keep the torch moving in a broad pattern.
8. Assuming a warm shop day equals compliance. Air temperature is not base-metal temperature. A comfortable 75°F afternoon does not guarantee a thick column flange has reached 150°F, and shade, wind, and cold soak can keep steel well below ambient.

Troubleshooting a Preheat-Related Failure

When a weld cracks, work through the evidence in a fixed order rather than guessing.

1. Identify the crack type and location. A transverse crack running out of the HAZ hours or days after welding points toward hydrogen-assisted cold cracking. Toe cracks, centerline cracks, and lamellar tears have different causes.
2. Check hardness. Portable hardness readings across the HAZ show whether the steel cooled too quickly and formed hard martensite. High readings support a thermal-control problem, but they are a clue, not a complete diagnosis.
3. Audit hydrogen sources. Look at electrode storage, flux condition, shielding gas, surface moisture, paint, oil, and humidity. A perfectly preheated joint can still crack if the consumable is damp.
4. Evaluate restraint. Thick members, tight fit-up, repair welds, and welds surrounded by existing structure all raise stress. Preheat helps, but joint detail and weld sequence may also need revision.
5. Confirm WPS compliance. Ask how temperature was measured, where, when, and by whom. “We preheated it” is not a record. A code minimum that was met on paper but not held through the cap pass is still a failure.

If the code minimum was genuinely met and the joint still cracked, the answer is rarely “add another hundred degrees and hope.” An engineer may increase preheat modestly, switch to a lower-hydrogen consumable, revise the weld sequence, or specify a controlled cooling or post-weld treatment. The fix should follow the diagnosis.

Safety on the Preheat Bench

Preheating turns a welding job into a hot-work operation before the arc even starts. Burns are the obvious risk, but fire is the bigger one: blankets, rags, wood, and plastic wrap near a 400°F joint can smolder for hours. Follow the hot-work permit, maintain the required fire watch, and keep a suitable extinguisher within reach. In general industry, OSHA 29 CFR 1910.252 sets the fire prevention and protection requirements for welding, cutting, and brazing.

Surface contamination matters for health as well as quality. Oils, paint, marking compounds, and galvanized coatings can release fumes when heated, so clean the joint before applying heat and ventilate the area. In confined spaces, preheat equipment adds heat, exhaust, and electrical risk to an already hazardous location. Resistance heating pads bring their own electrical hazards: inspect leads, keep connections dry, use ground-fault protection, and never let workers handle damaged cables.

Good preheat practice is boring by design: even heat, verified temperature, dry consumables, and a written record. The dramatic failures tend to come from the shortcuts that seem harmless at the time.

Cost, Field Versus Shop, and When You Can Legitimately Skip Preheat

Every preheat conversation eventually hits the same wall: the schedule. Heating a 2-inch flange to 250°F, holding it there through six passes, and waiting for soak time is not free. So the honest question is not whether preheat costs money — it does — but whether skipping it costs more.

What Preheat Actually Costs

The direct costs are fuel or electricity, the labor of the person holding the torch or pads, and the clock time the joint spends sitting at temperature instead of being welded. AISC’s Steel Interchange is blunt about it: preheating and maintaining interpass temperature can consume considerable shop or field time, and that time carries added labor plus fuel or electrical expense. On a large fabrication, preheat can be a genuine schedule driver.

The indirect cost is the one nobody budgets for: a crack found after galvanizing, a repair weld in a restrained joint, or a column flange that fails inspection. That is where the cost argument collapses. Preheat is cheap insurance priced against a very expensive claim.

Field Versus Shop: Where the Money Goes

In the shop, preheat is usually cheap per joint. You have power, a furnace or ceramic pad set, and the part sits on a bench where it can be heated evenly and checked with a contact pyrometer.

In the field, everything multiplies. Wind strips heat off the surface, blankets are awkward, ambient temperature swings, and the crew is working at height or in a confined space. Field preheat is where shortcuts appear — a single pass with a rosebud, a measurement taken once, a joint that cools while the welder repositions. Budget the time honestly and the shortcut loses its appeal.

When Lower or No Preheat Is Defensible

Code tables are conservative by design, written to protect the worst reasonable combination of thickness, chemistry, and restraint. Modern steelmaking has moved the goalposts: cleaner, lower-carbon steels paired with H4 and H8 low-hydrogen consumables can justify lower preheat through alternative calculation methods rather than blanket table values. Choosing the right consumable matters as much as the torch — our guide to alloy steel welding electrodes covers selection and storage.

Lower or no preheat is defensible when all of these are true:

• Thin material with low hardenability and low carbon equivalent.
• Low-hydrogen consumables, properly stored and dry.
• Dry surface and dry atmosphere, no frost or condensation.
• Ambient comfortably above the code floor.
• Alternative calculations (Annex B in AWS D1.1, or the D14.8M / ISO/TR 17844 methods) show a lower permissible value than the table.
• Low restraint and a joint that is not a repair.

When You Must Not Skip

• Thick sections. The plate is a heat sink; the HAZ quenches itself.
• High restraint. Restrained joints convert shrinkage into tensile stress.
• High carbon equivalent. More alloy means more martensite for the same cooling rate.
• Cold, damp, or windy conditions. Below 32°F, the steel itself must be brought up before welding.
• Repairs. Repair welds are inherently restrained and often on unknown or aged steel.
• Tack welds that will not be remelted. AWS D1.1 requires preheat for these, with higher minimums for small tacks because they cool so quickly.
• Any WPS that states a value. The WPS governs. Code minimums are floors, and the engineer may specify more.

A Simple Decision Walk-Through

1. Read the WPS. If it gives a preheat number, that number is the answer — stop here.
2. If not, identify the base metal, its category, thickness, and carbon equivalent.
3. Look up the code minimum for that combination and process.
4. Check ambient temperature, wind, and surface moisture against the code floor.
5. Decide whether an alternative calculation is warranted and approved.
6. Compare the resulting number to ambient. If ambient is higher and the surface is dry, document it and weld.
7. If anything in the “must not skip” list applies, heat it — and record what you did.

Skipping preheat is a decision with a paper trail, not a shortcut taken at 6 a.m. because the torch was cold. When the justification is written down, the choice is engineering. When it is not, it is a gamble with someone else’s structure.

Putting It Together: A Practical Preheat Workflow and Final Thoughts

Everything in this guide collapses into one repeatable routine. Whether you are welding a quarter-inch bracket or a two-inch restrained moment connection, the sequence is the same. Only the numbers change.

1. Read the WPS and the drawing. The WPS controls. Code tables only set floors. Note the preheat minimum, the interpass maximum, heat input limits, and whether postheat is called for.
2. Identify the base metal and its carbon equivalent. Match the steel to the correct category before you touch a torch.
3. Check ambient and surface conditions. Below 32°F, steel must be brought to at least 70°F before welding, per AWS D1.1. Frost, dew, or rain means dry the joint and re-check it.
4. Clean and dry the joint. Preheat never compensates for oil, heavy mill scale, or damp electrodes.
5. Select the heating method to suit geometry and location: rosebud, induction coil, ceramic pads, or furnace.
6. Heat evenly, on both faces where accessible, and allow soak time so through-thickness temperature catches up with the surface.
7. Verify at the correct location and distance — typically at least t or 3 in back from the joint, ahead of the arc, using calibrated contact instruments rather than infrared guesswork.
8. Record it. Preheat and interpass records are part of the quality documentation inspectors review alongside the WPS and welder qualifications.
9. Weld without letting temperature drop. A long conversation between passes is how a joint drifts 80 degrees out of spec.
10. Watch maximum interpass. The minimum protects the metal; the maximum protects toughness and, on quenched-and-tempered steels, the temper.
11. Cool or postheat as specified. Postheat (hydrogen bake-out) is typically done by raising preheat by about 100°C and holding for several hours on crack-sensitive thick joints.

When aluminum enters the mix, the logic shifts from hydrogen cracking to moisture and heat sink, and the preparation steps in our aluminum MIG welding guide pick up where this workflow leaves off. Hot work permits, fire watches, and fume control belong to the same routine — see OSHA 1910.252 for the baseline requirements.

So, preheating before welding is done to slow the cooling rate of the weld and heat-affected zone, which suppresses hard martensitic structures, gives diffusible hydrogen time to escape, and softens the thermal gradient that drives residual stress. It is a measured, documented, repeatable part of the process — not a warm feeling on the plate.

Keep the economics in view. A weld that cracks after inspection can cost far more than the fuel and labor saved by skipping preheat: cut-out, re-weld, re-test, schedule slip, and sometimes a reputation. Preheat is cheap insurance against a failure that shows up when nobody is watching.

Final Preheat Checklist

Frequently Asked Questions

What is preheating before welding done to?

Preheating is done to slow the cooling rate of the weld and heat-affected zone. That single change suppresses hard, brittle microstructures, gives diffusible hydrogen time to escape, and reduces the thermal shock and residual stress that lead to cracking.

What temperature should I preheat steel to before welding?

Use the value on your WPS first. If the WPS is prequalified to AWS D1.1, the minimum comes from the preheat table based on steel category, thickness, and process — for example 150°F for A572 Gr 50 or A992 from over 1-1/2 in through 2-1/2 in with FCAW. Code minimums are floors, not targets, and highly restrained or repair welds often need more.

Can I skip preheat if the weather is warm?

Sometimes, but warmth alone is not the test. Low-carbon thin steel welded with low-hydrogen filler in dry conditions may need little or no added heat, but thick sections, high carbon equivalent, high restraint, and damp surfaces still call for preheat. If the steel is below 32°F, AWS D1.1 requires heating to 70°F.

How do I measure preheat temperature correctly?

Measure at least the thickness of the thickest part, and never less than 3 in (75 mm), away from the joint in all directions, on the side being welded or the opposite face after a soak delay. Contact thermocouples and calibrated contact pyrometers are the most reliable; infrared guns can be badly thrown off by surface finish and are often not accepted for code compliance.

Does preheating prevent all welding cracks?

No. Preheat addresses hydrogen-assisted cold cracking, hardness, and thermal stress, but it cannot fix dirty metal, wet electrodes, wrong filler, excessive restraint, or bad bead shape. Solidification cracking, lamellar tearing, and shrinkage cracking have different causes and different fixes.

What is the difference between preheat, interpass, and postheat?

Preheat is the base metal temperature before the first arc. Interpass is the temperature of the weld and surrounding metal between passes, with both a minimum (usually equal to preheat) and often a maximum. Postheat is a deliberate hold after welding, typically raising preheat by around 100°C for several hours, to let hydrogen diffuse out of crack-sensitive joints.

Do you preheat stainless steel and aluminum?

Austenitic stainless steel is usually not preheated, because the concern is interpass temperature control and sensitization rather than hydrogen cracking. Aluminum is preheated mainly for thick sections and to drive off moisture, and the temperature is generally kept at or below about 350°F.

Can I preheat with a propane or oxy-acetylene torch?

Yes, and many shops do, but a torch is the least even method and burning fuel adds water vapor that can condense on cooler plate. Heat broadly around the joint, keep the flame moving, allow soak time, and verify with a gauge or crayon rather than judging by color.

How long should preheat be held before welding?

Long enough for the heat to soak through the thickness so the reading is real, not just surface-hot. A common field rule is to wait about one minute per inch (25 mm) of thickness before checking on the same side, then hold at or above the minimum until the weld is complete.

Does preheating increase welding cost?

It does add labor, fuel or electrical cost, and time. AISC notes that holding preheat and interpass can consume significant shop or field time, which is why alternative calculation methods and low-hydrogen consumables are worth considering. The comparison, though, is against rework, inspection failure, and delayed cracking.