Preheat Temperature for Welding: The Complete Guide

A pressure vessel weld passed visual inspection, passed radiography, and then a week later someone found a hairline crack running through the heat-affected zone. Nobody had welded anything wrong. The joint was simply cold when the arc was struck — and cold joints on susceptible steel are where hydrogen cracking hides. Preheat temperature is the single number that decides whether that happens, and it is one of the easiest things to get wrong on a job site.

Why Preheat Temperature Decides Whether a Weld Holds

Start with the answer, because everything else in this guide hangs off it: preheat temperature is the minimum temperature the base metal must reach before welding begins. Its job is not to make the arc behave or the bead look pretty. It exists to defeat hydrogen-assisted cracking in the heat-affected zone, the defect that turns a good weld into a scrapped assembly days after sign-off. That is why a joint can pass visual inspection and radiography and still be condemned later. Nobody welded it wrong. The joint was simply cold when the arc was struck.

The three conditions that have to line up

According to TWI (The Welding Institute), hydrogen cracking in arc welding needs three factors present at the same time:

• Hydrogen generated by the welding process, carried in by the consumable and by moisture in the coating, on rust or on damp surfaces.
• A hard, brittle structure susceptible to cracking, typically a martensitic or bainitic heat-affected zone in steel with enough hardenability.
• Tensile stresses on the welded joint, from restraint, residual stress or the load itself.

Remove any one of the three and the crack does not happen. That is exactly why preheat works: it slows the cooling rate of the heat-affected zone, so the microstructure forms softer and tougher, diffusible hydrogen has time to escape before the weld locks up, and HAZ hardness drops. TWI also notes that cracking generally occurs only once the temperature has reached normal ambient, which is why defects can appear hours or days after the weld looked perfect.

A shop-floor version of the failure

Picture a structural shop in January. A crew is welding 12 mm base plates to channels for a mezzanine. The lead hand says preheat is for pipe and pressure work, not for “mild structural steel”, and waves it through. The electrodes sat in a bag overnight in the trailer, the plate was swept but never degreased, and the ambient was near freezing. Nothing in that looks like a welding error. It looks like three ordinary Tuesday decisions that, stacked together, removed every condition except the hydrogen. By the time anyone noticed the crack, the joint had cooled to room temperature and the “perfect” weld had already failed.

Code requirement versus shop rule of thumb

A code requirement is a binding minimum an inspector can verify against an approved welding procedure specification. A shop rule of thumb is house practice: usually “preheat is for thin stuff over half an inch” or “we always heat to 200°F because that is what we have always done.” One is engineered, documented and traceable. The other is inherited.

AWS D1.1 has become the industry’s go-to reference largely because its prequalified procedures save the cost and effort of qualification testing, and AWS D1.1/D1.1M puts minimum preheat and interpass temperatures in Table 5.8 for the steel categories it covers. ASME B31.3 does the same job for process piping with its own material grouping. Federal rules such as OSHA 1910.252 set the safety frame but hand you no number. The number comes from the code and the WPS.

So preheat is a controlled variable, not decoration. If the WPS carries no minimum, no interpass ceiling and no measurement method, nobody in the field can prove it was achieved, and the record cannot defend the joint later.

Why the numbers are getting smaller, and when they must not

Modern structural steel has less carbon than the steel most old shop manuals were written for, and low-hydrogen consumables in the H4 and H8 classifications from AWS A5.x limit hydrogen pickup far better than the older high-hydrogen electrodes. Many joints genuinely do not need the old conservative numbers, and a lower or zero preheat can often be justified instead of heating half a tonne of steel for nothing.

Thicker sections, heavily restrained joints and higher carbon-equivalent steels still do. A thick, heavily restrained carbon steel header in freezing weather is the textbook case where skipping preheat is the expensive decision. Preheat is not free either: it adds shop time, fuel gas, electricity, welding hours and fume exposure. Getting the number right is a productivity decision as much as a safety one. On the physics, the costs and field measurement, follow the AWS Welding Digest treatment of preheat and interpass, and the ASME B31.3 requirements for piping work.

What this guide covers

Next we cover what preheat and interpass temperature physically do to the joint, how the codes calculate the required number, how to measure it on site, the mistakes that produce cracked and distorted work, and four worked examples you can copy onto a job folder. One takeaway to keep: the preheat number is a documented, measured variable on the WPS, and it exists to remove one of the three conditions hydrogen cracking needs.

I’ll rewrite this section, keeping the heading, facts, and all links intact while tightening the prose. Let me check the word count first.Still over the 840 limit. Let me tighten more aggressively and measure.962 — need another ~125 words off. Doing a final tightening pass.Still 956. Cutting deeper into prose and table cells.Let me build and measure the final tightened version precisely.

Preheat vs. Interpass vs. Post-Weld Heat Treatment

Three words that sound similar, get used interchangeably in the yard, and mean three completely different things. Get them confused and you will either crack a joint, soften a heat-affected zone, or slow the job down for no reason. Here is the plain-language version.

Term	What it actually is	When it happens	Who sets the number
Preheat	The minimum temperature the base metal must reach before the arc is struck	Before welding and before tacking	The approved WPS, backed by a code table or annex
Interpass temperature	The maximum temperature permitted between successive passes	Throughout the pass sequence	The approved WPS
PWHT / stress relief	A controlled heating and cooling cycle to relieve residual stress and restore properties	After welding is complete	The code, the design, or the equipment manufacturer

Think of it that way. Preheat is a floor, interpass is a ceiling, and PWHT is a separate operation with its own heating and cooling rate limits. As AWS D1.1 and ASME B31.3 both make clear, none of these are optional decoration. They are controlled variables that appear on the welding procedure specification, and the inspector checks them.

Why the interpass limit exists

On a single-pass weld, the story is simple. On a multi-pass weld, every later pass re-melts and re-heats the heat-affected zone of the pass before it. That repeated thermal cycle is why an interpass ceiling exists at all.

Let the interpass temperature run too high and three things happen. Heat accumulates in the joint faster than it can escape, so grains coarsen and the HAZ loses toughness. The repeated reheating softens the HAZ of quenched and tempered steel, which TWI notes is why preheat and interpass temperatures must be limited on those grades and heat input held below the steel supplier’s recommended value. And in the worst case you invite reheat cracking, which is associated with PWHT or high-temperature service, typically in the 350–550 °C range.

The interpass limit is a ceiling, never a target. A welder who runs the part at the interpass maximum all the way through the sequence is welding outside the intent of the procedure.

Keeping the weldment hot enough

Preheat on paper means nothing if the plate sits on the shop floor for twenty minutes and cools back to ambient before the first pass. Because cracking rarely occurs at temperatures above ambient, TWI points out that maintaining the temperature of the weldment during fabrication is as important as the preheat itself.

So if welding stops long enough for the part to fall below the required preheat temperature, you reheat, re-verify, and re-record. There is no grace period and no judgement call. For crack-sensitive situations such as higher carbon-equivalent steels or highly restrained joints, the holding temperature and period should be increased, typically 250–300 °C held for three to four hours, so diffusible hydrogen has time to escape before the joint locks up.

A worked example

Take a 20 mm ASTM A36 bracket with a multi-pass fillet weld, SMAW with an E7018. Your WPS carries a preheat minimum of 100 °C and an interpass maximum of 200 °C.

• Before tacking: the parent metal is brought to at least 100 °C and verified with a temperature-indicating crayon or thermocouple. Tacks are welded hot, not cold.
• During the sequence: the root pass goes in while the joint is still near 100 °C. Between passes, the welder keeps the area comfortably below 200 °C so the heat does not pile up. 200 °C is the line you must not cross, not the number to aim for.
• If the job stops: the part cools below 100 °C. Before the next pass starts, the joint goes back up to 100 °C and the reading is logged again. Skipping that step is one of the most common causes of a root or HAZ crack found days later.
• After the final pass: if the joint is crack-sensitive, the weldment is held at 250–300 °C for three to four hours as a post-weld soak, then allowed to cool naturally. That soak is not PWHT. It is hydrogen relief.
• If PWHT is required: the part is first allowed to cool, and inspected at ambient temperature, before it goes into the furnace or the local heating blanket.

Where PWHT changes the plan

Stress relief is a controlled cycle with its own heating rate, soak temperature, holding time and cooling rate, and it has to be coordinated with the preheat and interpass limits rather than bolted on afterwards. Where a weldment is heated through the reheat-cracking range, the preheat and interpass values in the WPS need to be consistent with the PWHT chart, otherwise you have built the cracking susceptibility into the procedure.

There is a second, quieter trap here, and TWI warns about it directly. If PWHT is applied immediately after welding, without first allowing the preheat temperature to fall, rejectable defects may not show up until after the PWHT cycle, because inspection is normally carried out at ambient temperature. A defect that was already there, or that was created by the hydrogen held in the joint, stays hidden until the part cools. If the PWHT cycle is long enough, some of that hydrogen can diffuse out. If it is short, it can stay in.

That is the practical argument for cooling the weldment down, inspecting it cold, and only then stress relieving. You find defects while you can still cut them out, and you get the benefit of the hydrogen soak that the cooling period provides.

For quench and tempered pressure equipment, the same logic applies in reverse, and choosing a deliberately undermatching consumable so the HAZ does not need a high preheat at all is a strategy worth understanding before you pick filler metal. Our guide to alloy steel welding electrodes breaks down the matching and undermatching options and what each does to the HAZ.

When the PWHT requirement comes from the design rather than the fabrication code, the ASME B31.1 and AWS D1.1/D1.1M requirements still apply, but the final answer belongs on the WPS and the inspection and test plan, coordinated between the welding engineer, the code consultant and the owner. The AWS D14.8M cold-crack avoidance standard is also worth reading if hydrogen cracking is the live issue on your job.

The Five Factors That Set Your Preheat Temperature

Most welders meet preheat temperature the same way: someone hands them a number and they follow it or ignore it. Swap the electrode, thicken the plate, or turn a fillet into a full penetration butt and that number may no longer be right. Learn the five inputs that move it and you can reason to an answer instead of guessing.

1. Thickness and Restraint

Thickness matters because heat sinks into the steel. A 6 mm plate heats quickly, holds heat close to the joint and cools slowly after the arc stops. A 60 mm section draws heat away from the weld metal almost as fast as you put it in, so the heat-affected zone (HAZ) snaps to low temperature in seconds and passes straight through the martensite transformation range.

A joint free to shrink as it cools can relieve its own residual stress; a joint that is bolted, welded into a rigid frame or closing a loop cannot. TWI’s guidance separates low restraint (thickness under roughly 30 mm) from high restraint (over roughly 30 mm) when setting preheat for austenitic electrodes at 1 to 2 kJ/mm. High-restraint details include a fillet onto a rigid base plate, a box-section closure joint, a nozzle-to-header connection, or a weld into a stiff assembly. A captive joint counts as restrained even when thin.

2. Carbon Equivalent and Hardenability

Carbon content alone is a poor predictor. What matters is how easily the steel hardens, which depends on the alloying elements taken together. Engineers capture this with the carbon equivalent (CE) value, calculated from the IIW, CET or Pcm formulas. A 0.30% carbon steel with strong alloying can be more crack-prone than a 0.40% carbon steel with little else in it, because alloying elements delay the transformation of austenite and let hard martensite form.

TWI notes preheat levels may reach 200 °C when welding thick-section steels with a high IIW CE value, citing EN 1011-2:2001. That sits at the top of the range, not the middle, and it applies where cooling is fast and restraint is high.

3. Diffusible Hydrogen Level of the Consumable

Hydrogen enters the weld from the consumable: moisture in the electrode coating or flux, wet wire, damp storage, contaminated base metal. As TWI puts it, the most effective means of avoiding hydrogen cracking is reducing the hydrogen generated by the consumable, using a low-hydrogen process or low-hydrogen electrodes.

Hydrogen levels are classified under the AWS A5.x filler metal specifications. An SMAW electrode rated H16 carries considerably more diffusible hydrogen than a basic electrode at H8 or lower, and a flux-cored wire at H4 is among the cleanest options. Swapping an H16 electrode for an H4 wire on the same joint can be the difference between needing 100 °C and needing no preheat at all.

Code tables reflect this. AWS D1.1/D1.1M:2025 Table 5.8 groups prequalified preheat options by steel category and consumable classification: Groups B and C permit filler metals up to H16, Group D limits classification to a maximum of H8, and Group E covers ASTM A1066 thermo-mechanically controlled processed Grades 50 through 65 with H8 or lower filler metal.

4. Joint Design and Geometry

A single-pass fillet weld in 10 mm plate cools differently from a multi-pass groove weld in 50 mm. Later passes reheat the HAZ of earlier ones, helping hydrogen escape but building heat if the interpass limit is ignored. A V-groove dilutes heat over a wide area and reduces the volume of highly hardened metal at the root. A square butt without preparation concentrates heat in a narrow, deep root pass with high restraint, the hardest case here.

A mill edge is sharp and may melt unevenly; a bevelled or prepared edge gives the arc a cleaner path and usually reduces the preheat needed for sound fusion. A fillet weld on a rigid base carries high restraint through the throat of the joint, so it behaves more severely than its thickness suggests.

5. Heat Input and Welding Process

Heat input controls cooling rate and hydrogen exposure. A slow, wide weave lays a broader, flatter bead with a lower local cooling rate and gives hydrogen longer to diffuse out; a fast stringer concentrates heat into a narrow region that cools quickly. The same energy per metre, spread differently, produces different microstructures.

Process changes hydrogen pickup and deposition behaviour. SMAW with basic electrodes is a standard low-hydrogen choice. GMAW with solid wire and submerged arc welding with some flux-cored wires can achieve very low hydrogen levels, though deposition rates and shielding conditions vary. FCAW and SAW at high deposition rates deliver heat quickly, affecting both the HAZ and the interpass temperature.

Factor	Effect on preheat
Thickness	More as thickness grows
Restraint	More in high-restraint details
Carbon equivalent	More for higher CE steels
Consumable hydrogen	Less with H4 or H8 consumables
Joint geometry	Multi-pass and square butt need more
Process and heat input	Wide weave beats fast stringer

Why This Matters in Practice

Every factor can move the answer, and they interact, which is why preheat sits in the welding procedure specification (WPS) rather than the welder’s judgement on the day. If plate arrives at 38 mm instead of 25 mm, if a fillet becomes a full penetration butt, if the job moves from SMAW to FCAW, or if the electrode is swapped because the specified one was out of stock, the preheat stated in the WPS may no longer be justified.

Any such change should trigger a check back to the WPS and the engineer who approved it. A documented deviation is a decision somebody has to own. Making it quietly at the torch is how cold cracks end up in the heat-affected zone a week after the inspector walked away.

Read these five factors and the code tables stop looking like magic. Next, we see how AWS D1.1 Table 5.8 and ASME B31.3 turn those inputs into a number for a work order.

Reading the Codes: AWS D1.1 Table 5.8 and ASME B31.3

Every preheat number on a job site came from a code table. The hard part is knowing which table applies to your joint, which category your steel falls into, and which edition your contract calls out.

AWS D1.1: the prequalified route

Most structural fabrication in North America falls under AWS D1.1/D1.1M, Structural Welding Code – Steel (the current edition is D1.1/D1.1M:2025 with Amendment 1, published through the AWS codes and standards program).

The key idea in D1.1 is prequalification. Clause 5.7 sets the rules, and the preheat and interpass temperature for a prequalified joint is not left to engineering judgment. It comes from Table 5.8, “Prequalified Minimum Preheat and Interpass Temperature.” If your material, joint detail, process and filler metal sit inside a prequalified combination, the table gives both the minimum preheat and the maximum interpass temperature. You do not get to lower them because the plate looks thin.

Table 5.8 is organised into five steel categories. Finding yours is the first step:

Category	What it typically covers	What the category controls
A	Lower-strength carbon steel; permits SMAW electrodes that are not low-hydrogen classification	No filler metal hydrogen ceiling, so more conservative preheat
B	Common structural steel, including ASTM A572 Grade 50 up to 55 ksi SMYS	Low-hydrogen filler permitted under defined conditions; the usual home for structural work
C	Higher-strength steels, 60 to 70 ksi SMYS	Tighter limits on processes and heat input since the HAZ hardens readily
D	A limited list of steels, such as ASTM A913 at 50 to 65 ksi SMYS	Filler metal capped at H8 or lower diffusible hydrogen, restricted joint details
E	ASTM A1066 TMCP steel	Its own set of preheat, interpass and filler metal conditions

A category is more than a preheat number. It also fixes which filler metals, joint details and processes stay inside the prequalified framework. Substitute a different electrode or change a joint detail and you may fall out of the prequalified combination entirely, which is where why SMAW consumables and process choices start to matter as more than a preference.

Two gaps to know about. First, Table 6.9 covers quenched and tempered steels and special steels such as ASTM A1043; those are not prequalified and must be qualified by test. Second, to use preheat or interpass lower than Table 5.8 allows, go to Clause 6.8, “Essential Variables,” Subclause 6.8.4, “Preheat and Interpass Temperatures,” which points to Annex B and other recognised methods. Note that 6.8.4 also states qualification testing is still required even when you calculate a lower number. The calculation reduces your testing burden; it does not remove it.

ASME B31.3: Table 330.1.1 and the P-numbers

Process piping has its own logic. ASME B31.3, Process Piping, in its 2024 edition running to 592 pages, sets preheat in Table 330.1.1 and groups materials by P-number. The commonly reproduced values give P-No. 1 carbon steel a minimum preheat of 50 °F (10 °C) unless wall thickness exceeds 1 in. (25 mm) and carbon content exceeds 0.30%. Thinner or lower-carbon material gets the ambient minimum, which shows why the same steel can need nothing at 10 mm and something meaningful at 30 mm.

Higher P-numbers behave differently. Low-alloy grades such as P-No. 3, 4 and 5 pick up substantially higher preheat because their hardenability is far higher, and Cr-Mo creep-resistant grades carry heat input and interpass ceilings rather than a simple preheat number. Austenitic stainless P-numbers sit at the other end, where the driving concern shifts from hydrogen cracking to distortion, interpass temperature and interpass cooling rate. Power piping runs through the parallel ASME B31.1, and the two do not give identical answers.

One caution: B31.3 preheat values have been reprinted, summarised and paraphrased for decades, and the table is revised between editions. Treat any number you find on the internet as a prompt to open Table 330.1.1 in your own licensed copy, not as an answer.

When the table does not cover your case

When your joint does not fit a prequalified combination, or the material is not listed at all, three things apply in order of authority:

1. The WPS governs. Once the procedure is qualified and approved, the preheat and interpass temperatures written on it are what the welder and inspector work to.
2. Qualification testing is required, and in B31.3 work that means qualification under ASME Section IX as well as the preheat rules in the B31.3 chapter.
3. The owner specification can be stricter than the code. A refinery, shipyard or power plant may impose higher preheat, tighter interpass limits, or additional heat input caps. The code sets the floor; the contract sets the job.

Read the table, find the category, note the filler metal restrictions that come with it, then write the result into the WPS where the welder can see it and the inspector can check it.

Beyond the Lookup Table: Carbon Equivalent and Alternative Preheat Methods

Nearly every preheat number you use comes out of a table, and that is not a criticism. Tables are fast, safe and let the shop floor get on with the job. But a table has to assume the worst case for everything it cannot see: the highest carbon equivalent the grade could realistically have, the highest diffusible hydrogen the consumable might carry, the most restrained joint anyone might build. Run the numbers properly and the answer often drops a long way.

Three formulas, one idea

Carbon equivalent is a single number that sums up how easily a steel hardens in the heat-affected zone. Three versions are in common use, and each weighs the same idea differently:

Formula	Origin	Expression
CE (IIW)	International Institute of Welding	C + (Mn/6) + (Cr + Mo + V)/5 + (Ni + Cu)/15
CET	Developed for C-Mn steels, widely used in Europe	C + (Mn + Mo)/10 + (Cr + V)/5 + Ni/15
Pcm	Used in AWS D1.1 Annex B	C + (Si + Mn + Cr + Cu)/20 + (Ni + Mo + V)/15 + (B/5)

Read the coefficients as a summary of what each alloying element does to hardenability:

• Carbon is the dominant term. More carbon means more hard, brittle martensite or bainite forming in the HAZ, exactly what hydrogen cracks love.
• Manganese adds hardenability and strength, and works with carbon to raise the crack risk in C-Mn steels. TWI’s guidance on hydrogen cracking in C-Mn steels is worth reading alongside any calculation.
• Nickel improves both hardenability and toughness, so a modest but real contributor.
• Chromium and molybdenum have strong hardenability effects. Molybdenum also slows tempering, which is why Cr-Mo creep-resistant grades get their own heat input limits.
• Vanadium hardens efficiently in small amounts and carries a full 1/5 weight in the IIW and CET formulas.
• Copper contributes a smaller hardenability effect and appears in CE and Pcm.
• Boron is the outlier: a few parts per million can dramatically raise hardenability, which is why it gets its own term in Pcm.

One warning before you compute anything: carbon equivalent means nothing in isolation. TWI points to BS EN 10277-5:2008 (formerly BS 970 Part 3) as a good example of quenched and tempered steel families with high carbon content and high hardenability. A CE value tells you the steel’s potential, not the family, the strength or the design rules that apply to it.

AWS D1.1 Annex B and AWS D14.8M

Annex B of AWS D1.1/D1.1M, titled Guidelines on Alternative Methods for Determining Preheat, is a normative annex, so it is part of the code rather than optional commentary. It offers two methods that set both preheat and interpass requirements from three inputs: steel composition and thickness, filler metal diffusible hydrogen level, and an assumed level of joint restraint.

AWS D14.8M:2009 (ISO/TR 17844:2004 IDT), Standard Method for Avoidance of Cold Cracks, gathers four recognised methods in one place: two based on European practice (CE and CET), one based on a Japanese standard (CEN), and one derived from Annex B of AWS D1.1 using Pcm. It is published in SI units only, so shops that think in Fahrenheit convert on the fly.

The AWS Welding Digest makes the case plainly: these four methods have been relied on globally for decades, and running all of them where they apply gives a strong cross-check on the preheat and interpass levels you have chosen.

So why are they under-used? Two honest reasons. They look complicated on paper, and using them means qualification testing, documentation and an engineer who owns the result. In exchange you get lower preheat, less fuel, less distortion and faster cycle times. If you are choosing a consumable for a heavy carbon steel joint, read how alloy steel welding electrodes are classified by hydrogen level first, because the hydrogen content of the rod is one of the three inputs the calculation hangs on.

A worked example

Take a 50 mm (about 2 in.) thick structural steel butt joint with a nominal CE of 0.38, welded with a low-hydrogen consumable rated for the lowest diffusible hydrogen category. Read conservatively from the standard table, assuming worst-case composition, higher hydrogen and fully restrained conditions, and the job may be quoted at 200 °C minimum preheat. Run the Annex B or D14.8M calculation with realistic chemistry and a moderate restraint assumption, and the same joint can land in the 100 to 150 °C band, with a correspondingly lower interpass ceiling. Those figures are illustrative rather than code text; the real number has to come out of the annex procedure and be backed by qualification testing, but the size of the gap is real and repeatable.

Preheat Temperature by Material: Carbon Steel, Alloy Steel, Stainless and Aluminum

Preheat is not one number with minor variations. It shifts from family to family, because each has its own cracking mechanism, tempering behaviour and reasons a heat-affected zone turns brittle. The AWS Welding Digest guide to preheat and interpass temperature frames the same point from the code side: the answer comes from material, thickness, restraint and consumable together.

Carbon and carbon-manganese structural steel

For everyday structural grades, preheat is often not needed on thin material. TWI’s guidance is a workable rule of thumb: when welding steels with up to 0.2% carbon, a preheat would not normally be required, whereas above 0.4% carbon a minimum temperature of about 150 °C is needed to prevent HAZ cracking when using austenitic electrodes at 1–2 kJ/mm heat input.

Published ranges put ASTM A36 around 50–100 °C depending on thickness, and ASTM A572 Grade 50/65 slightly higher at roughly 75–125 °C. Read them as orientation, not instruction. AWS D1.1 Table 5.8 and your approved WPS always govern. One related detail: ASTM A1066 thermo-mechanically controlled processed steel in Grades 50 to 65 sits in AWS D1.1 Group E, where filler metals classified H8 or lower are permitted.

Pressure equipment steels and low-alloy grades

Piping and pressure equipment push the numbers up. Indicative preheat bands run about 150–200 °C for P11, 200–250 °C for P22, and 200–300 °C for P91 and P92, with P-No. 1 carbon steel sitting at the lower end of that scale.

Cr-Mo creep-resistant steels work differently: the HAZ must stay fine-grained or the metal loses the creep properties it was bought for, so these grades carry their own heat input ceilings and interpass limits. On P91 and P92, preheat is only half the job; the interpass cap and heat input limit matter just as much. For consumables, our guide to alloy steel welding electrodes covers matching and storage discipline. For piping, the controlling table is in ASME B31.3.

Quenched and tempered steel

Q&T steel sets two demands in direct opposition: enough preheat to avoid hydrogen cracking in a hardenable base, and not so much that the HAZ overtemperes, softens and loses the strength the plate was heat treated for. TWI lists the practical countermeasures against fabrication hydrogen cracking as an undermatching filler metal, increasing preheat, or increasing heat input, and notes that thin sections of Q&T steels present no significant welding problems. Because these grades fall outside ordinary preheat tables, qualification by test is normally the route.

Stainless steels

Most austenitic stainless grades need no preheat. Type 304, 316 and their variants are not hydrogen-cracking sensitive the way carbon steel is, and published WPS templates typically list preheat as not required. Ferritic and martensitic grades are the exception and behave far more like carbon steel.

The real controls on stainless are heat input, interpass temperature and interpass cooling rate. Keep heat input and interpass modest and you limit distortion while avoiding sigma-phase embrittlement in high-chromium grades and 475 °C embrittlement in ferritic and duplex families. Duplex 2205 is typically welded with low or no preheat and an interpass limit of 150 °C or less, per the WPS. Because austenitic stainless shrinks more than its parent metal, a higher preheat is sometimes used deliberately to reduce distortion, and TWI recommends austenitic nickel alloy electrodes for highly restrained joints, since nickel expands less.

Aluminium alloys

Aluminium is the family where preheat is generally avoided. It conducts heat several times faster than steel and melts at a much lower temperature, so heating the base metal invites burn-through and distortion, and preheat can lower HAZ strength. The controls are interpass temperature and heat input instead. If the process is new to you, our aluminium MIG welding guide covers the setup that keeps the puddle where you want it.

Material or family	Indicative preheat (published references)	What governs it
ASTM A36 carbon steel	About 50–100 °C, thickness dependent	D1.1 Table 5.8 category, thickness, restraint, hydrogen
ASTM A572 Gr 50/65	About 75–125 °C	Higher carbon equivalent, same factors
ASTM A1066 TMCP (D1.1 Group E)	Reduced values commonly permitted	Filler metal classified H8 or lower
P-No. 1 / P11	About 150–200 °C	B31.3 Table 330.1.1 plus the WPS
P22	About 200–250 °C	B31.3, with heat input ceiling
P91 / P92	About 200–300 °C	Heat input and interpass ceilings
Quenched and tempered grades	Limited; qualified by test	Cracking risk vs HAZ softening
Austenitic stainless (304, 316)	Not generally required	Heat input, interpass, cooling rate
Duplex 2205	Low or not required	Interpass 150 °C or less per the WPS
Ferritic and martensitic stainless	Often required	Behaves like carbon steel
Aluminium 5xxx / 6xxx	Generally avoided	Interpass temperature and heat input

Measuring and Verifying Preheat on the Job

The theory only matters if the steel actually reaches the number on your WPS. Preheat verification is where most shops get sloppy, and it is the single easiest step to document properly. Get this right and you have a clean audit trail; get it wrong and a later crack has no trail back to you at all.

The good news is that the measurement hierarchy is not complicated. You are choosing between four families of equipment, and each one trades cost against accuracy and, more importantly, against how badly it can mislead you in a real shop.

Temperature-Indicating Crayons and Melt Sticks

These are the cheapest and most common option, and for a lot of field work they are the right answer. You rub or press a stick onto the metal at the temperature point. Once the surface reaches the stick’s calibrated melt point, the mark liquefies or the stick loses its shape, and you know the surface has arrived at that temperature.

They are single-use, they only tell you about one set point at a time, and they need a reasonably clean surface to give a readable mark. But they have one big advantage over infrared: they are not fooled badly by surface condition. An infrared thermometer reads the surface it sees, and a scaled, oxidised, rusty or painted surface radiates differently from clean steel, which throws the reading out. A crayon sits on that same surface and reacts to the actual local temperature regardless of colour or oxidation.

That is why, on a weathered plate or a mill-scale surface, crayons and melt sticks are frequently more trustworthy than a spot gun pointed at the same spot. Multiple sticks at different set points also give you a range rather than a single number.

Contact Thermocouples with a Meter

A thermocouple welded to, clamped to, or magnetically attached to the parent metal, wired into a temperature meter, is the most accurate option you will find on a job site. It gives you a live reading and, if you log it, a continuous record of what the joint temperature did over the whole weld.

The catch is placement. The thermocouple belongs on the parent metal a couple of inches back from the joint, not on the weld, and away from the direct path of the flame or arc heat. Put it in the flame path and you are measuring the torch, not the steel. Put it on the joint itself and you are measuring the fusion zone, not the base metal the code is talking about.

Infrared Thermometers and Thermal Imaging Cameras

These are fast, non-contact, and the only sensible option for mapping the temperature field across a whole assembly or a large plate. A thermal camera is genuinely valuable for finding cold spots near a joint that a single spot reading would miss.

They are also the easiest tool to misuse. Every reading depends on emissivity, spot size, distance, and angle. Point a cheap IR gun at a shiny mill-scale surface from six feet away and the number can be off by a hundred degrees or more. Fix the technique: use the same emissivity setting, the same distance, and the same angle every time, hold the gun square to the surface, keep the spot smaller than the area you care about, and calibrate against a contact thermometer on the same part when you can. TWI’s guidance on hydrogen cracking in steels is a good reminder that the whole thermal history of the joint matters, not just the number at the start.

Dedicated Preheat and Interpass Monitoring Systems

For code work, there is a fourth option: a purpose-built preheat monitoring system that logs temperature at one or more points and records the curve, including the interpass peaks and the cooldown. On pressure equipment, structural work with an owner’s inspector on site, or any job where the documentation has to survive an audit, this is the cleanest route. It turns preheat verification from an argument into a data file.

Where You Take the Reading

AWS D1.1’s preheat verification guidance for the 2025 edition calls for measuring preheat on the base metal surface at a minimum of 3 in. (75 mm) from each edge of the weld joint. That is the convention: the temperature of the parent metal near the joint, not the pool and not the far side of the plate.

Surface readings also lag the core temperature. That cuts both ways. A low surface reading on a cold part is a clear failure, because the core is even colder still. But a high surface reading on a heavily oxidised or scaled surface can be misleading, particularly with infrared, because the surface reads hotter than the metal beneath it. Preheat must be reached before the arc is struck, and held through the weld, including when welding resumes after an interruption, per ASME B31.3 preheat practice.

One more habit from TWI is worth internalising: because cracking rarely occurs at temperatures above ambient, maintaining the weldment temperature during fabrication matters as much as hitting the preheat number. Do not let the joint cool back down between passes or during a stoppage and then weld on.

Safety and Containment

Preheat means heat. Heat means fire risk, and it means gas. Under 29 CFR 1910.252, all hollow spaces, cavities or containers must be vented to permit the escape of air or gases before preheating, cutting or welding, and purging by inert gas is recommended. A sealed cavity heated to 150 °C will do damage well before it reaches its set point.

The Documentation Chain

Preheat belongs on the inspection and test plan, not in someone’s memory. The record should carry:

• Preheat and interpass temperatures required, and the governing clause or table
• Measured values, with the time, the location and the method used
• Welder and inspector sign-off
• Confirmation that the temperature was reached before the arc was struck
• Any reheat and re-verification after an interruption

When a weld is stopped long enough for the part to cool below the required temperature, the sequence is always the same: reheat, re-verify, re-record. Some codes also require a post-weld soak to let hydrogen escape before the joint is allowed to cool, so check the WPS before the torch comes back on.

If you are shopping for temperature-measurement equipment, our Aluminum MIG welding guide covers the measurement discipline needed for aluminium work, where temperature control is even tighter. We also keep a welding equipment guide that covers preheat gear and monitoring options alongside the rest of the kit, and you should read our all-position welding rod guide before changing consumables, since a change in electrode classification can change your hydrogen level and therefore your preheat requirement.

Affiliate disclosure: some of the links above may earn us a commission if you buy through them. That does not change our advice, and we only point you toward tools we would consider ourselves.

Common Preheat Mistakes and How to Troubleshoot Them

Most preheat failures are not exotic metallurgy. They are Tuesday afternoon decisions made in a hurry: skipping the preheat because the plate looked thin, pointing the torch at the wrong end of the joint, letting a multi-pass weld go cold between passes, or opening a box of electrodes that have been sitting in a damp trailer. Understanding the nine mistakes below will save more rework than any amount of theory.

Mistake 1: “The plate is only 3/8 inch, I don’t need preheat”

Thickness is the easiest number to find and the most tempting to lean on, but it is only one input. The same 10 mm plate can be a relaxed single-pass fillet in a shop frame or a highly restrained closure joint in a stiff assembly, and those two joints do not share a cracking risk. Restraint raises both the cooling rate and the tensile stress that pulls the HAZ apart. Add a consumable that introduces more diffusible hydrogen and you have changed the answer again.

There is a hard limit on how much preheat can rescue a bad consumable choice. TWI is clear that preheating alone cannot offset a consumable that introduces too much hydrogen, and that the most effective means of avoiding hydrogen cracking is to reduce hydrogen generation in the first place by using a low-hydrogen process or low-hydrogen electrodes. Preheat manages hydrogen; it does not create less of it.

Mistake 2: Preheating the wrong location

Heating the torch side of the joint, or a convenient flange a metre away, warms the steel the welder can see rather than the steel that matters. The joint itself can still be colder than the required minimum at the moment the arc is struck. Preheat is only effective when the heat is put into the base metal immediately adjacent to the weld, the region where the HAZ will form. Far away, you are simply wasting gas and time.

Mistake 3: Measuring the surface instead of the parent metal near the joint

A surface reading on cold, oxidised plate tells you very little about the temperature a few millimetres below. And a surface reading taken where the flame is licking tells you even less. Measurements need to be taken on the parent metal adjacent to the weld edge, on the surface, with the check taken far enough from the edge that the torch heat is not the thing you are reading.

Mistake 4: Letting the part cool between passes

Multi-pass welds lose heat to every pass. If the interpass temperature falls below the preheat minimum, the joint must be re-preheated before the next arc is struck. Crews routinely weld a root, take a coffee break, and fill with a cold joint without realising they have undone the preheat. The rule is simple: before the next pass, measure. If the part has dropped below the required minimum, reheat and re-record.

Mistake 5: Overheating

More heat is not safer heat. Excessive preheat or excessive heat input produces coarse grain in the HAZ, more distortion, more spatter, higher residual stress and a real risk of reheat cracking in susceptible steels. This is the trade-off TWI describes for reheat cracking: reducing heat input will almost certainly require a higher preheat temperature to avoid hydrogen-assisted cracking. The two controls pull against each other, which is why both need to come from the procedure rather than from instinct.

Stainless steel adds its own failure mode. Running excessive interpass temperatures on austenitic stainless promotes distortion and can encourage intergranular corrosion problems, and controlling the interpass cooling rate matters as much as the peak temperature. When preheat is used on stainless it is usually a deliberate distortion-control measure, not a hydrogen-cracking measure.

Mistake 6: Damp electrodes, wet storage, oily or rusty plate

Moisture defeats the purpose of preheat entirely, because the preheat cannot remove hydrogen once it is already inside the deposited metal. Electrodes left out of a rod oven, flux-cored wire stored in a damp shed, plate that has been rained on, and rust that has lifted the mill scale all feed hydrogen or prevent fusion. Preheat will not rescue a joint contaminated with paint, oil or shop dirt either; that metal needs cleaning, not more heat. Good preparation practice for these surfaces is covered in our guide to cleaning aluminum before welding, and the same principle of removing contamination first applies to steel.

Mistake 7: Mixing up consumables

Substituting an H16 electrode where the WPS calls for a low-hydrogen filler changes the hydrogen content of the deposit, which changes the cracking risk, which can invalidate the preheat basis of the procedure. The filler metal classification is part of the qualified procedure, not a convenience choice. If a different classification is used, the procedure needs to be requalified before it goes into production code work.

Mistake 8: Ignoring the environment

Outdoor welding in wind, high humidity or rain, and welding base metal that has been sitting at ambient temperature all raise the effective heat requirement above what the table minimum suggests. Convection strips heat out of the joint faster, and cold steel absorbs more of the arc energy before it reaches fusion temperature. Practical guidance is to apply stricter preheat control in these conditions, not the bare minimum from AWS D1.1 Table 5.8, which assumes controlled shop conditions. If the fabricator chooses to qualify outside the prequalified Chapter 5, Part F route, the preheat values fall under the qualification requirements of Clause 6.8.4.

Mistake 9: Confusing preheat with post-weld heat treatment

Preheat happens before the arc. PWHT happens after the weld is finished, usually in an oven, with its own heating and cooling rate limits. A post-weld soak held at temperature to let hydrogen escape is a third thing again, and it is not a substitute for preheat. Teams that treat the two as interchangeable either overheat the joint before welding or believe a scheduled oven run has covered the cold-cracking risk that preheat was there to address.

Troubleshooting table

Symptom	Likely cause	Corrective action
Crack in the heat-affected zone, appears hours or days after welding	Hydrogen-assisted cold cracking: insufficient preheat, high restraint, or high diffusible hydrogen from the consumable	Verify the preheat record, switch to a low-hydrogen consumable, requalify the procedure, and review the restraint level
Root crack at the stop or restart of a bead	No crater fill or the arc was struck on cold metal after a pause	Require crater fills, and re-establish preheat before restarting after any interruption
Excessive distortion after welding	Overheating, too much heat input, or preheat applied to one side only	Reduce heat input, distribute preheat evenly around the joint, and check the interpass ceiling
Coarse grain or poor bead profile, heavy spatter	Excessive preheat temperature or excessive heat input	Lower preheat to the specified minimum, reduce travel speed or weave width, and add stringer beads where the WPS allows
Root oxidation, porosity or lack of fusion	Preheat applied to the wrong area, contaminated or wet plate, or base metal still cold under the slag	Move the heat to the joint, clean the surface, dry the consumables, and re-measure at the correct distance from the weld edge
Distorted stainless weld with corrosion staining	Excessive interpass temperature or slow interpass cooling	Lower the interpass limit, add a cooling pass between passes, and control heat input

The crack is a chain of events

A crack discovered a week after inspection is never a single failure. It is a sequence: cold base metal, a consumable that gave up too much hydrogen, a restraint that pulled the HAZ apart, an interpass excursion, a contaminated surface, and a cooling rate that turned the HAZ hard. That is why the weld log is the best diagnostic tool you have. Preheat readings, consumable lot numbers, electrode oven records, ambient temperature, inspection results and repair history, read in sequence, will tell you where the chain broke. Without those records, the same crack will happen again on the next job and nobody will be able to say why.

Building a Practical Preheat Plan: Four Worked Examples

Four scenarios, worked as a welding engineer would. Every number is illustrative, drawn from published code guidance rather than quoted code text, so treat this as a method to copy, not a substitute for your own WPS and code edition. New to preheat calculations? Having the case for SMAW and its consumable handling fresh helps, since two examples turn on electrode classification.

Example 1: A Shop-Fabricated Structural Bracket in A36

A fabricator builds a mounting bracket from 20 mm (roughly 3/4 in.) ASTM A36 plate with 6 mm fillet welds, SMAW on an E6013 electrode. The bracket is free to deform as it cools, the shop is heated, and no impact testing was requested.

Under AWS D1.1/D1.1M, A36 sits in the lower-strength bucket, putting it in Category B territory for this fillet size and thickness. A non-low-hydrogen electrode there still carries a modest preheat, typically the 50 to 150 °F (10 to 65 °C) band in a qualified WPS. With 20 mm material, low restraint and a 20 °C shop, no preheat may be the practical answer.

The WPS controls: where it lists a minimum preheat, the welder obeys it even in a warm shop; where it lists none, the welder records ambient and base metal temperature, checks the plate is dry and free of scale and oil, and welds. Log grade, thickness, WPS number, electrode lot, preheat and method, and sign-off. Swap electrodes and the drier can drive hydrogen into a joint running with no preheat: good weld on bracket one, cracked bracket on bracket ten.

Example 2: A 25 mm A36 Plate Butt Weld, Outdoors in Winter

Same material, but a 25 mm (1 in.) plate-to-plate butt weld in a moderately restrained connection, on a winter site with base metal near 5 °C and a stiff wind across the joint. More thickness speeds cooling through the joint, more restraint raises residual stress, and the cold plate has no thermal headroom.

The material may still be Category B, but the code’s guidance and the WPS set the number, and a restrained 1 in. butt weld is not the low-restraint case TWI describes for material under roughly 30 mm. A sensible plan starts near 150 °F (65 °C) minimum, with an interpass ceiling held from first pass to last.

The interpass limit is what people break: a slow torch and a long puddle raise the pass temperature, the second pass goes down on a red, oxidised surface, and the extra heat adds distortion the fit-up cannot absorb. Re-establish preheat if the arc is off long enough for the part to drop below the minimum, then re-measure and re-record, and use interpass monitoring or frequent contact readings rather than eyeballing bead colour. As TWI notes on hydrogen cracking, the classic cold-crack site is the HAZ of a hard, restrained joint, exactly this geometry.

Example 3: An ASME B31.3 Process Line in P-No. 1 Carbon Steel

A 12 mm wall process line in P-No. 1 carbon steel, schedule piping, root pass by SMAW with fill and cap by GMAW. This is the ASME B31.3 world, and the logic differs from D1.1.

P-No. 1 carbon steel falls in the group where the published preheat table gives a modest minimum, commonly reproduced as 50 °F (10 °C), unless wall thickness exceeds 1 in. (25 mm) and carbon exceeds 0.30%, which pushes the requirement up. Twelve mm wall sits well under that, so the minimum stays low. The requirement is held through the weld and re-established when welding resumes after an interruption, and the qualified WPS sets the specific minimum and interpass the shop floor works to. Mixing processes is normal, but the WPS must cover the SMAW root and GMAW fill and cap as a qualified combination, or you qualify it as one.

Preheat sits on the inspection and test plan; temperature, method, time, welder and inspector go on the weld log. An overnight stop means preheat is re-established and re-verified before the arc is struck again, and that is recorded too. Most B31.3 documents are only as good as the traceability behind them.

Example 4: A Repair Weld on a Q&T Pressure Vessel Component

A nozzle-to-shell repair on a quenched and tempered pressure vessel component. Here the rules run the other way.

Q&T steel is hardened by quench and strengthened by temper. Welding the HAZ risks losing the temper, but preheating hard enough to fight hydrogen cracking pushes the HAZ toward martensite, which over-tempers during PWHT and softens the joint below its specified minimum strength. So preheat has to be limited, not maximised, and heat input capped to keep the HAZ narrow. Filler metal typically undermatches the parent strength on purpose, so the weld is not the weak link.

Nor is this a case where the procedure engineer simply picks a number: under ASME B31.1 and the vessel codes, repairs and alterations often fall outside the original manufacturer’s data, and welding may not proceed without approval from the manufacturer, the code holder or the authorised inspector. As TWI explains on reheat cracking, PWHT on Q&T steel needs controlled heating and cooling rates with the holding temperature kept below the original parent tempering temperature, or the quench and temper is undone. One more wrinkle: PWHT may run straight after welding without letting the preheat fall, but inspection is normally only possible at ambient temperature, so rejectable defects often surface only after the cycle completes.

The Preheat Planning Checklist

• Material and thickness — grade, spec, actual thickness, CE/CET/Pcm where applicable.
• Restraint — low or high, and why.
• Process and consumable — classification, diffusible hydrogen (H4, H8, H16).
• Code and clause — code, edition, table.
• Table lookup — preheat, interpass, category or P-number.
• Annex B alternative — considered or rejected, with reasons.
• Interpass ceiling — on the WPS and the shop floor copy.
• Measurement — crayons, thermocouples or IR, where and who signs.
• Re-preheat rule — time limit after interruption, re-establishment, re-verification.
• PWHT and soak — post-weld cycle, hydrogen soak, coordination with preheat and interpass.

One method runs through all four examples: identify the material, judge restraint and hydrogen level, look up the category or P-number, then decide whether the tabulated answer is realistic or a lower value is justified through qualification. Write it down once, and the shop floor argument becomes a lookup instead of an opinion.

Frequently Asked Questions

What is the minimum preheat temperature for welding carbon steel?

For thin, low-restraint carbon steel, no preheat may be required at all under AWS D1.1 — TWI notes that steels up to about 0.2% carbon generally need none. As thickness, carbon equivalent, restraint and hydrogen level rise, the minimum climbs: commonly reproduced values from ASME B31.3 Table 330.1.1 give P-No. 1 carbon steel a 50 °F (10 °C) minimum unless wall thickness exceeds 1 in. (25 mm) and carbon exceeds 0.30%. Always take the controlling number from the approved WPS and the governing code.

How do I calculate carbon equivalent for my steel?

The IIW formula combines carbon, manganese, nickel, chromium, molybdenum, vanadium, copper and boron into a single hardenability index, and AWS D1.1 Annex B, AWS D14.8M (ISO/TR 17844) and EN 1011-2 each use slightly different variants (CE, CET, CEN, Pcm). Because the formulas differ, always state which one you used and why, and check the governing code or WPS for the acceptable method rather than picking whichever gives the lowest number.

Can I use a lower preheat temperature than the code table says?

Yes, but only through a defined route. AWS D1.1 Annex B and the D14.8M methods let you calculate lower preheat and interpass values for specific combinations of composition, thickness, hydrogen level and restraint, and that can save substantial time, fuel and money. The catch is that qualification testing is still required whenever you go below Table 5.8, and the engineer takes on the documentation.

Why does my crack show up a day after the weld looked perfect?

This is the signature of hydrogen-assisted cold cracking. TWI explains that cracking generally occurs only when the temperature has reached normal ambient, so the defect appears after the joint has cooled and stresses have redistributed. Look at three things: was preheat actually applied and verified, was the consumable within its hydrogen classification, and were the electrodes and the joint free of moisture, oil and rust.

What is the difference between preheat and interpass temperature?

Preheat is a minimum that must be reached before the arc is struck; interpass is a maximum that must not be exceeded between passes. Keeping the interpass under control matters because later passes reheat the earlier heat-affected zone, and excess heat causes grain growth, distortion, HAZ softening in quenched and tempered steel and reheat cracking in Cr-Mo alloys.

Do stainless steels and aluminium need preheat?

Most austenitic stainless steels need no preheat, and aluminium alloys are generally welded without it — preheat on aluminium risks burn-through because of its high thermal conductivity and low melting point, and can reduce the strength of the heat-affected zone. The exceptions are the ferritic and martensitic stainless grades, where preheat limits and heat input control are commonly specified to avoid cracking.

Is an infrared thermometer good enough to check preheat?

It can be, but only with discipline. Infrared readings are sensitive to emissivity, so they become unreliable on oxidised, rusty, painted or otherwise dirty surfaces, and readings taken at a distance or at an angle to the spot can be badly wrong. Temperature-indicating melt sticks or crayons, or a properly attached thermocouple on the parent metal, are generally the more trustworthy choices on real shop steel.

How long must I keep the part warm after welding?

For susceptible steels, TWI advises maintaining the preheat temperature for a period of typically two to three hours to allow hydrogen to diffuse away from the weld area, and in crack-sensitive situations increasing that to 250–300 °C for three to four hours. If your code calls for post weld heat treatment instead, that usually replaces the soak — but coordinate the two, because PWHT can introduce its own reheat-cracking risk.