Welding Hardened Steel: The Complete Guide to Avoiding Cracks and Lost Hardness

The loader bucket had been in the shop for three days before Marcus finally admitted what had happened. A hardened edge on a cutting blade had cracked right down the middle, and the previous crew — confident, fast, and wrong about everything that mattered — had decided the fix was simple: weld it. They ran a hot bead straight down the crack with a big rod, watched the arc eat into the plate, and called it a repair. Twelve hours later, with the bucket sitting cold on the shop floor, the weld split open again with a sound like a rifle shot. The crack ran longer than the original. It had, in fact, been hiding inside the steel all along, waiting for the metal to cool.

That is the story of welding hardened steel in one paragraph: it punishes people who treat it like mild steel. The heat that is your friend on a sawhorse frame becomes your enemy on a quenched-and-tempered plate. The instructions that are optional on A36 — preheat, electrode baking, heat input control, slow cooling — become life-or-death on material that was hardened on purpose.

If you have landed here searching for welding hardened steel, you are almost certainly in one of these situations. You have a cracked hardened part — a cutting edge, a shaft, a mold, a gear tooth — and you need to know whether it can be saved. You have a project that calls for high-strength quenched-and-tempered steel like A514 or T-1 and you want to weld it without destroying the properties you paid a premium for. Or you are a student or a curious fabricator who has heard that “you can’t weld hardened steel” and wants to know which part of that sentence is actually true. This guide answers all three. It covers what hardened steel is, why welding it fails, the exact procedures that make it succeed, the filler metals that keep it alive, and the post-weld steps that determine whether your weld is still there next month.

What “Hardened Steel” Actually Means

Before any advice makes sense, you need a clear picture of what hardened steel is — because “hardened” is not a single material. It is a family of materials that got hard by different routes, and each route demands a different welding strategy. Understand which one you are holding, and half your problems disappear.

Quenched-and-Tempered (Q&T) Structural Steels

These are structural and wear-resistant steels that are heated to full austenitizing temperature, quenched to form martensite, and then tempered to balance strength and toughness. The best-known family is ASTM A514, sold commercially for decades under trade names like T-1®. A514 is a low-alloy steel with a minimum yield strength of 100,000 psi (690 MPa) — roughly two and a half times the yield strength of ordinary A36 structural steel. Related grades include A517 (pressure-vessel service) and tougher HSLA steels like Hardox® and Strenx® in Europe, plus armor plate grades. Because these steels are delivered hard by design, the entire art of welding them is controlling what the arc does to the microstructure the mill built. Some hardening is, of course, lost in the heat-affected zone no matter what you do. The goal is to control how much is lost, and to prevent a crack from forming where that loss happens.

Through-Hardened and Tool Steels

Tool steels such as D2, O1, A2, D6, and H13 are hardened all the way through for tools, dies, punches, shear blades, and machine components. These are high-carbon or high-alloy steels — D2, for example, carries around 1.5 percent carbon — and the carbon content is exactly what makes them weld so badly. High carbon means the heat-affected zone will form hard, crack-prone martensite on cooling, with almost no ductility to absorb the shrinkage stress of the weld. Carbon is the enemy that preheat fights hardest.

Case-Hardened and Surface-Hardened Steels

A lot of “hardened” parts in the real world are only hardened at the surface. Carburized (case-hardened) steel has a thin, very hard carbon-rich outer layer over a tough, weldable core. Induction-hardened shafts and rollers are hardened locally on the surface by rapid heating and quench. Nitrided parts get a hard compound layer at the surface without quenching. All three are traps because the part looks hard and behaves hard, but the dangerous material is only an eighth of an inch deep. Welding ignorance on a case-hardened gear shaft produces exactly the same crack as welding a full tool steel — often worse, because the hard case sits on a soft core, so the crack runs along the boundary.

Work-Hardened and Cladded Steel

Finally, some steel is hard because it was cold-worked or because a hard surface was applied. Work-hardened steel has dislocations and strain from forming, and re-applying heat generally recrystallizes it back to soft. Cladded and hardfaced surfaces — manganese steel, chromium carbide overlays — are a different class again, usually welded with dedicated high-alloy consumables. I will touch on hardfacing later as the opposite of “welding to join.”

The takeaway: when someone says “welding hardened steel,” they usually mean Q&T structural steel, tool steel, or a surface-hardened component. This guide covers all three, with a full section on each. They share the same physics — hard microstructures are crack-prone microstructures — but the practical procedures differ.

Why Welding Hardened Steel Is So Different From Mild Steel

Mild steel forgives you. The arc heats a small patch of A36, you deposit filler, and when the metal cools the microstructure returns to something close to what it was — soft ferrite and pearlite that will not crack under residual stress. Nothing about that chain of events survives contact with hard steel. Three distinct problems appear.

Problem One: Heat Destroys the Hardness

The hardness you paid for is a product of microstructure, and microstructure is a product of heat treatment. Quenched-and-tempered steel got its 100 ksi yield from being quenched to martensite and then tempered at a carefully controlled temperature. Now the arc comes along and re-heats a narrow band of that steel to far above its tempering temperature — often right up to full austenitizing temperature — and then lets it cool on its own. The result, in part of the heat-affected zone, is an overtempered, softened region. You can measure the damage: in A514, heat-affected-zone softening near the weld can drop the local hardness well below the base metal’s, in some regions by tens of hardness points. On a wear plate or a cutting edge, a softened HAZ wears out and deforms long before the rest of the part does. This is the reason engineers say weld-first-then-harden when they can possibly arrange the sequence.

Problem Two: The Cooling Side Creates Martensite — and a Crack

Here is the cruel symmetry of welding hard steel. On one side of the weld, heat softens an overtempered zone. On the other side, the same heat creates the opposite problem: freshly hard martensite. When a high-carbon or high-alloy steel cools from the weld’s heat, the fast cooling rate — the weld joint acts like a poor man’s quench — transforms the heat-affected zone into untempered martensite, which is hard, brittle, and intolerant of stress. This is the classic recipe for hydrogen-induced cracking (HIC), also called hydrogen-assisted cold cracking, underbead cracking, or simply hydrogen cracking.

Hydrogen-induced cracking needs three things present at the same time: (1) hydrogen from the welding process dissolved into the steel, (2) a hard, crack-susceptible microstructure like martensite, and (3) tensile stress from shrinkage, restraint, or service load. Remove any one of the three and cracking is essentially stopped. That sentence is the entire philosophical basis of every procedure in this article — the preheat is there to slow cooling and avoid the hard structure, the low-hydrogen electrode is there to remove the first ingredient, and the joint design and tacking plan are there to control the third.

Problem Three: Delayed Cracking Happens After You Think It’s Fine

Hydrogen cracking is nearly always delayed. The weld cools, looks perfect, passes visual inspection, and then — hours later — the hydrogen that was trapped in the steel finally concentrates enough stress to crack it. Industry guidance is explicit: hydrogen cracks can initiate and grow after the joint cools below around 200°C, and detection of a serious hydrogen crack often happens 16–48 hours after welding is complete. In design code terms, that is the premise of “delayed inspection”: you cannot trust a weld in hardened steel that was checked only at the end of the shift. The bucket in Marcus’s shop split overnight. It always does.

The Two Kinds of Failure, Side by Side

Welders get confused when hardened steel fails, because it can fail two completely different ways, and the fixes are opposite. The table below is worth pinning above your bench.

Notice what the two headline failures have in common: both are managed with the same levers — preheat, heat input, filler hydrogen level, and cooling. That is why almost every credible procedure for welding hardened steel reads like a check list of those four things. Get them right and you can weld material that shops swear cannot be welded. Get any one of them wrong and the metal will tell you, usually at 2:00 a.m., with a gunshot sound.

Can You Actually Weld Hardened Steel?

The honest answer is: yes — with the right material, the right process, and the right procedure — but there are many situations where you should not, and a few where doing so is outright dangerous. Let me be precise about the boundaries.

High-strength quenched-and-tempered steel (A514, T-1, Hardox): Yes, routinely weldable with qualified procedures. The entire crane, mining, earthmoving, and heavy-equipment industry welds it every day, using preheat, low-hydrogen consumables, and heat-input control. This is the intended, engineered, accepted use of these materials — they are designed to be welded, with discipline.

Repair welding a hardened part to restore a broken tool, shaft, or die: Often yes, but the repair weld will not restore the original hardness in the heat-affected zone, and it may move the failure to the edge of your repair. For many repairable parts the correct play is: weld the geometry back with a tough undermatching filler, then re-harden and retemper the whole part. For other parts — gears that run at peak load, springs, anything fatigue-critical — welding is usually a temporary patch at best.

Welding high-carbon tool steels “as-is” with no procedure: No, not successfully. D2 at 1.5 percent carbon does not want a field weld, period, unless it gets high preheat, controlled cooling, and immediately goes to temper heat treatment afterward. Welding it cold is how you create a guaranteed crack.

Welding on case-hardened or nitrided parts without removing the case: No — you must grind or machine off the hard case before welding, or you are welding onto a material that is already martensitic and brittle. This is the single most common field mistake in the entire topic, and it is responsible for a huge share of “I tried to weld hardened steel and it cracked” stories.

There is also a question about why you are welding. If the goal is to restore strength or hardness, welding is frequently the wrong tool — the alternatives are mechanical repair (shrink-fit sleeves, bolted patches, replacement inserts), braze… no, careful, let me keep it clean: mechanical repair, or complete replacement. If the goal is to deposit wear-resistant material on top of a part, that is hardfacing, covered near the end of this guide. Knowing the goal changes the procedure as much as knowing the material.

Know Your Steel First: Carbon Equivalent and the Mill Sheet

Every good decision about welding hardened steel starts with knowing the carbon equivalent (CE) of the material — the single number that predicts how badly a steel wants to form hard, crack-prone martensite on cooling. CE is not a “nice to know.” It is the number that your preheat chart, filler choice, and qualification status all hang on.

The most widely used form is the International Institute of Welding (IIW) formula:

CE = C + Mn/6 + (Cr + Mo + V)/5 + (Ni + Cu)/15

Compositions are entered as weight percentages from the mill test certificate (mill sheet). The result is read on a scale that experienced welders learn by feel:

• CE below 0.35: low risk; low-carbon structural steel like A36 welds with no fuss, no preheat for moderate thickness.
• CE 0.35–0.45: weldable, but preheat starts to matter as thickness grows; low-hydrogen practice is recommended.
• CE 0.45–0.55: moderate-to-high carbon-manganese and low-alloy steels; preheat, low-hydrogen filler, and controlled cooling become mandatory.
• CE above 0.55: high hardenability; expect serious preheat, extremely low hydrogen control, and often a full qualification program. Most high-carbon tool steels and many hardened hardenable steels land here.

Quenched-and-tempered grades like A514 typically fall in the 0.50–0.75 range depending on the grade — high enough that preheat is non-negotiable. The CE-based approach is so central that AWS D1.1’s own guidance for determining preheat when no table applies is built around it, via the hydrogen-control and hardness-control methods in its annexes.

You also need to identify which steel you hold, for two reasons. First, the proper procedure references a grade: A514, A517, A36, Hardox, D2, H13, 4140. Second, and surprisingly often, the “hardened steel” in someone’s hand is not actually heat-treated at all — it is just mystery steel from a scrapyard with a hard scale on top. Welding that requires a totally different answer. So the first physical act of any hardened-steel weld is not lighting an arc; it is reading the mill sheet, or if none exists, testing: a spark test, a file test, a hardness test, or a known-history check with the supplier. On critical work, send a sample for analysis. Guessing the chemistry of steel you are about to preheat to 400°F and weld with an E11018 rod is like guessing the fuel octane of an engine you are about to rebuild — the cost of being wrong is a cracked part, not a pinging noise.

The Golden Rules: Preheat, Interpass, Heat Input, Cooling

Everything the arc does to hardened steel is decided by temperature and time. The four rules below are not optional flourishes; they are the difference between a weld that survives and a weld that cracks on its own schedule. If you remember nothing else from this article, remember these four.

Rule One: Preheat Before You Strike the Arc

Preheat is the single most important control in welding hardened steel, and it does three distinct jobs at once. First, it slows the cooling rate of the weld and heat-affected zone, so the steel has time to form softer, tougher microstructures instead of untempered martensite. Second, it drives surface moisture out of the joint — moisture is a source of hydrogen, and hydrogen is the crack’s fuel. Third, it reduces the temperature difference between the arc and the plate, which lowers the thermal shock and the residual stress that crack propagation feeds on.

The precise number depends on the steel grade and thickness, not on a rule of thumb. For carbon-manganese steels, the AWS guidance is clear that the temperature must be attained at a distance from the weld at least equal to the thickness of the thickest member — and never less than 3 in (75 mm) from the point of welding. In other words: preheat the whole region, not a fingerprint-sized spot in front of the gun. And the minimum preheat must be maintained from the first tack through the last pass, because “interpass” is just “preheat, enforced continuously.”

Rule Two: Control Interpass Temperature — Upper Limit Included

Everyone remembers the minimum preheat. Half the people forget the maximum interpass. On quenched-and-tempered steels, the interpass temperature has a hard ceiling — for A514 it is 400°F (205°C). Exceed it and each subsequent pass soaks the previously deposited weld metal and the adjacent base metal at a temperature approaching the tempering range of the material, degrading the very microstructure you are trying to preserve. The shop solution is unglamorous but effective: measure before every pass, and if the part drifts too hot, wait. A temperature-indicating crayon, a contact thermometer with a probe, or an infrared gun are all acceptable; a mental “it feels hot enough” is not a measuring instrument.

Rule Three: Keep the Heat Input Inside the Qualified Window

Heat input — the energy the arc pours into the joint per inch of travel — is calculated from amperage, voltage, and travel speed:

Heat Input (J/in) = (Amperage × Voltage × 60) ÷ Travel Speed (in/min)

For quenched-and-tempered steels such as A514, typical qualified heat input runs around 55,000 J/in, plus or minus roughly 20 percent — and your welding procedure specification (WPS) will state the qualified range exactly. Two failure modes live at the two ends of this window. Too little heat per inch means the weld cools too fast and you risk hard, crack-prone microstructures and lack-of-fusion. Too much heat per inch overtemperatures the base metal and widens the soft heat-affected zone. Notice the trap: on hardened steel, “turn the machine down” does not always protect the material. The WPS’s heat-input band is a controlled middle, and production welds that wander outside it are a qualification violation as much as a quality risk.

Rule Four: Cool Slowly, and Don’t Let the Hydrogen Out-Nerve You

Fast cooling is a quench, and a quench is how martensite gets made. So after the last pass, hardened steel wants to cool gradually, not be baked with a water hose or left to face a drafty door. On crack-sensitive steels it is common and correct to hold the joint at or above preheat for a period after welding — typically two to four hours — to let hydrogen diffuse out of the weld zone while the metal is above the dangerous temperature band. The length of the hold grows with the crack sensitivity: for the most sensitive steels, guidance runs to 250–300°C for three to four hours. Some shops wrap the joint in insulation blankets so the cooling rate stays uniform and slow from the last pass to ambient. That blanket is doing engineering, not comfort.

Measuring Temperature for Real: Instruments That Don’t Lie

Hardened-steel procedures are only as good as the temperature measurement underneath them. The AWS-derived standard is a temperature-indicating crayon — a wax stick calibrated to melt within a specific temperature band, drawn across the heated surface. When it melts, you are at or above its rating; when it holds its line, you are below. Crayons are cheap, fast, and defensible, and they are the classic tool for both preheat verification and interpass checks.

Two instruments do a better job of continuous control. A contact thermocouple pyrometer reads the surface through a probe touching the plate, which is accurate when pressed firmly and given a moment to stabilize. An infrared thermometer reads from a distance, which is convenient for big, hard-to-reach parts — but emissivity, reflectivity off bright steel, and distance all skew the reading, so experienced hands confirm IR readings against a reference or against a crayon on critical joints. Whichever tool you use, the routine is identical: heat the joint well (from the side opposite the weld if possible), verify the temperature at a point adjacent to the joint, and re-verify immediately before each pass for the entire weld. Temperatures wander while you work — preheat that was right at the start of the root pass is not evidence that the fill pass is protected.

Choosing the Right Filler Metal for Hardened and High-Strength Steels

Filler metal selection on hardened steel is a battle between two requirements that sometimes collide: you want enough strength to carry the joint, and you want as little hydrogen as possible feeding the crack. There are also cases where the smartest choice is deliberately lower strength than the base metal. Let me walk through the logic, because the right rod is cheap insurance and the wrong rod is money down the drain with a crack on top.

Low Hydrogen Is the Whole Game

The hydrogen content of a deposit is measured as diffusible hydrogen in milliliters per 100 g of deposited weld metal, and the threshold the industry treats as “low hydrogen” for hardenable steels is 8 mL/100 g or less. For the highest-strength quenched-and-tempered grades, manufacturers often demand even lower — some recommend 5 mL/100 g or less for premium wear steels.

Here is what those numbers look like in ear language, from published consumable guidance:

• Cellulosic electrodes (E6010, E6011): roughly 30–50 mL/100 g. These are the classic “big arc, deep dig” pipeline and hot-pass rods — and they are exactly what you do not use on hardened steel.
• E7018 (the standard low-hydrogen stick): dry and freshly baked, it deposits well under 8 mL/100 g. But it absorbs moisture from humid air fast, and guidance shows a 7018 rod at 0.6 percent coating moisture depositing around 12 mL/100 g — past your low-hydrogen budget. That is why low-hydrogen electrodes are stored in ovens or sealed rod keepers and rebaked if exposed.
• Solid wire GMAW (bare wire fresh from a sealed pack, dry): very low hydrogen, typically far below 8 mL/100 g, and for that reason the mildest-mannered process for many hardenable-steel jobs.
• GTAW (TIG): lowest hydrogen of all when the joint is clean and dry — as low as ~2–4 mL/100 g — which is why TIG is the go-to process for welding thin edges on hardened tooling and for repair where a small, precise, clean deposit is wanted.

The practical translation: dry rods, sealed or baked wire, clean and dry joint surfaces, and no wind pulling your shielding away. Hydrogen also sneaks in through rust, oil, grease, paint, moisture, and even condensation — so joint prep on hardened steel is true prep: grind to bright metal, degrease, and dry before the preheat lamp even starts.

Matching, Undermatching, and the “Softer Is Safer” Rule

For structural joints in high-strength Q&T plate, the rule is to match the strength of the lower-strength member — never to overmatch drastically. When you weld A514 to A36, for example, the correct filler matches the A36 side (typically a 70 ksi class electrode), because overmatching to 110 ksi would force all the strain into the softer, weaker A36 heat-affected zone anyway, and buy nothing but crack risk. When A514 goes to A514, AWS-classified consumables like E11018-M (SMAW), ER110S-1 (GMAW), or the matching flux-cored classes provide the strength and the low-hydrogen pedigree the joint needs.

For repair of hardened or hardenable parts like shafts, dies, and wear plates, the expert move is often deliberate undermatching: deposit a tough, ductile, low-hydrogen weld metal that will yield a little rather than crack the hard parent metal. The goal is not to restore hardness (you cannot, by welding), it is to restore geometry and let the weld flex where the parent metal cannot. This is why so much hardened-plane and gear repair is done with tough low-alloy deposits rather than hard high-carbon ones.

The Austenitic Escape Hatch

One more tool belongs in every hardened-steel welder’s kit: austenitic stainless filler, usually 309L or 307 class, sometimes nickel-alloy wire. Austenitic weld metal dissolves hydrogen readily and holds it in solution, so very little of it diffuses into the hard heat-affected zone — and it is also far more ductile than ferritic deposits. On steels where preheat is impractical or where nothing else will stop the crack, an austenitic deposit is a legitimate, codewise approach that has saved countless hard parts. It is not a substitute for a proper procedure, and it cannot be used everywhere (it changes corrosion and expansion behavior), but when welding conditions are desperate — a repairing a hardened mold on a cold night with no heat source — the austenitic route is the honest professional answer to “what would you do?”

Welding ASTM A514 (T-1 Type) Quenched-and-Tempered Plate: A Working Procedure

A514 is the material most often in people’s hands when they search for welding hardened steel, because it is the premium wear-and-structural plate under bulldozer blades, crane booms, dump bodies, boom truck pads, and armored equipment. Here is a realistic, code-informed working procedure — not a substitute for your WPS, but a map of what a proper one contains.

Qualification Reality Check

First, the thing most hobby welders never hear: A514 and other quenched-and-tempered grades are not on AWS D1.1’s prequalified base-metal list. You cannot just follow a prequalified WPS out of the book. A real production procedure for A514 requires a Procedure Qualification Record (PQR) backed by mechanical testing — tension, bend, and charpy V-notch where the specification needs toughness. Each A514 grade has its own chemistry, so a PQR qualified on Grade B does not automatically cover Grade F. For non-critical repairs, people still weld A514 informally with good practice; just know that on engineered, code-governed work there is no shortcut around the testing.

Typical Preheat and Interpass for A514

The commonly cited AWS D1.1-style preheat table for low-hydrogen welding of A514 runs by single-plate thickness at the joint:

Read that table with both hands on it. The minimums climb with thickness because thick sections cool faster and concentrate more restraint. The 400°F maximum is not a suggestion — it protects the temper. And these values assume low-hydrogen practice; non-low-hydrogen processes need more, which is a good reason to never reach for a cellulosic rod on this material at any thickness.

Heat Input and Technique

Hold the qualified heat-input window, roughly 55,000 J/in ±20 percent as a typical anchor. Keep the bead profile controlled, avoid leaving craters (crater cracks are a favorite failure of hard steel), and let each bead fill the joint without bulging. Out-of-position welding on A514 multiplies every risk — vertical and overhead welds cool differently and are harder to keep in the heat-input band — so the simplest controls are the ones most worth making: right joint prep, tight fit-up, no bridging where a gap should be, and no arc strikes anywhere off the joint. Every stray arc strike on hardened steel is a tiny, crack-prone quench site with a hard HAZ and a thumbnail-sized chance of becoming a longer crack later.

Repairing Tool Steel and Hardened Tooling

Tool steel is where “welding hardened steel” stops being a fabrication question and becomes a rescue mission. Dies, punches, shear blades, knives, molds, broaches, and machine ways get damaged in service, and the question is whether a weld can bring them back. The honest answer is: sometimes, under the right discipline, for the right damage — and never with the mindset of a structural weld.

First, Grind Away the Damage — and the Decarb

Every real tool-steel repair starts with removing the cracked, fatigued, or decarburized material. Welding over a crack or over the soft oxidized layer on top of a hot-worked die is how you turn a small repair into a bigger crack. Grind or machine to bright, sound metal and open the defect into a shape a weld can fill — a “U” or “V” excavation with rounded corners at the bottom. Square-bottomed grind-outs concentrate stress into the repair zone, and on hardened steel the sharp corner becomes a starter notch.

Preheat Your Tool Steel Like a Heat Treat, Because That’s What It Is

Tool steel preheat is not the 50°F “warm it up” of structural A514 — it is a real heat-treat-minded event, and it is often far higher than instinct suggests. Hot-work tool steels such as H13 commonly call for preheat in the 350–450°F (175–230°C) range before welding, with tempering immediately after. Cold-work D2, at about 1.5 percent carbon, demands the same respect and more: high preheat, low-hydrogen welding, and a post-weld tempering cycle before the part goes back into service, because the as-welded structure in the HAZ is untempered martensite that will crack as it sits. When a shop is honest with itself, tool-steel repairs get treated as a miniature heat-processing operation: preheat, weld hot and controlled, temper promptly, slow-cool, inspect after a delay.

Filler Chemistry Follows the Tool, Not the Machine

The filler for a tool-steel repair is selected by what the tool does after the repair. For a hot-work die made of H13, the repair is typically made with matching H13-type filler — often a TIG wire or a thin coated electrode — so the deposit has the same tempering response and the whole tool can go back to heat treat as one coherent piece. For cold-work tools where matching chemistry isn’t practical, repair welds frequently use either a tough low-alloy build-up underneath with a hard-facing cap, or a dedicated repair electrode that tolerates the carbon pickup from the base metal. And here too the austenitic route appears: 309-type filler is a documented rescue filler for welded repairs on hardenable steels where preheat and post-heat are impossible, because it swallows hydrogen and stays ductile.

Note the word “tempering” in that paragraph. A tool-steel weld does not end when the arc breaks. There is widespread agreement across the trade that repair welds on hardened tooling must be followed by a post-weld temper (or a full anneal-reharden-temper sequence where the part’s properties must be restored). Weld, then temper, then inspect after the delayed-crack window — that is the complete cycle. Weld and leave it is how dies explode in service.

Case-Hardened, Induction-Hardened, and Nitrided Parts

The single most dangerous “hardened steel” in the real world is the one that is only hard on the surface — because it looks like a normal part, and it will crack like a tool steel the moment you weld it cold and dry. Shafts, gear teeth, cam lobes, spindles, rollers, and bearing surfaces get carburized, induction-hardened, or nitrided in massive quantities, and every one of them is a trap.

The rule is short and absolute: weld only on the soft core, never on the hard case. The hard case is a thin layer — commonly 0.02 to 0.06 in (0.5–1.5 mm) for carburizing and induction hardening, thinner still for nitriding — of high-carbon or high-nitrogen martensite that is precisely the kind of microstructure hydrogen cracking loves. The accepted procedure for welding a case-hardened shaft or gear is to grind or machine away the hard case in the joint area before welding, verify with a file or hardness test that you have reached the soft base metal, and then weld that core with a low-hydrogen, often undermatching filler. Weld after, and you are welding to a brittle shell sitting on a soft core — the crack will follow the shell boundary and take the whole repair with it.

There is a second, quieter reason the case must go: the hard case also shrinks the load path. A weld rooted in the hard shell carries almost no load because the shell is brittle and the core is soft; the joint fails in fatigue even when it does not crack immediately. Removing the case restores a weldable, load-carrying connection that behaves like the ductile material underneath — which is what you actually want a repair weld to do.

For nitrided parts the situation is even stricter: the compound layer at the surface is extremely hard and essentially unweldable, and it must be ground off completely, with the chemical depth (case depth) verified before any filler is laid down. And for parts hardened in service by wear and work — like manganese steel crusher elbows — you are not dealing with a heat-treated case at all; you are dealing with work-hardened austenitic steel that welds by a completely different rulebook (dedicated high-manganese filler, low heat input, no preheat — because for manganese steel, preheat is precisely what you must not do).

Welding Hardened Steel vs. Hardfacing: Two Different Games

People searching for “welding hardened steel” frequently mean something else entirely: “how do I make steel harder by welding?” That is hardfacing — depositing a wear-resistant overlay on top of a part — and it deserves a sharp distinction, because the two disciplines have opposite goals.

Welding hardened steel is a joining or repair operation. Your goal is a sound, crack-free joint that carries load, in a base metal that is already hard. You fight the base metal: you preheat it, cool it slowly, and cross your fingers about its microstructure.

Hardfacing by contrast is a surface engineering operation. Your goal is to lay down a deliberately hard, wear-resistant deposit — chromium carbides, tungsten carbides, high-chromium iron, manganese steel — on a part that will rub, scrape, impact, or slide. You choose filler chemistry whose hardness you want, whether the base metal is hard or not. The two operations overlap in technique (both demand clean joints and controlled heat) but diverge completely in intent, filler selection, and acceptance criteria. A weld that shows a hardfacing rod was used where a structural rod was needed is a repair that will crack; a structural rod used where a wear surface was needed is a part that will wear out.

The reason the distinction matters in this article is simple: almost every hardfacing job involves a hard or high-carbon base metal, so the base-metal side of the job still follows the preheat, low-hydrogen, and cooling rules in this guide — and only the deposit chemistry changes. You can follow every hardfacing overlay chart in the world and still crack the part underneath if you skipped the preheat for the base metal. Many welders who “welded hardened steel and it cracked” were actually hardfacing without base-metal discipline.

The Most Common Mistakes People Make Welding Hardened Steel

None of the failures discussed here are exotic. They are variations of the same handful of mistakes, made by experienced welders who simply treated hardened steel like mild steel. If you avoid these seven, you are already ahead of most of the industry.

1. Welding cold, with no preheat. The classic. The part looks fine, the arc starts, and the weld cools like a quench. Untempered martensite forms; a delayed crack follows. Preheat is not a luxury on hardened steel; it is the whole point.
2. Using a high-hydrogen rod. E6010, E7010, or an unbaked E7018 read “welding rod” to many people, and they are 30–50 mL hydrogen bombs on a crack-prone joint. Hardened steel wants the driest low-hydrogen consumables you own — baked, rod-kept, or sealed-wire — anything under 8 mL/100 g.
3. Welding a case-hardened part without removing the case. The most expensive “I didn’t know” mistake in the trade. A carburized gear or induction-hardened shaft welded cold will crack along the hard case boundary and take all your filler with it.
4. Ignoring the max interpass. Preheating, then letting the part climb to 500°F during multi-pass welding, overtempers the steel exactly like the arc did in the first place. The 400°F ceiling on A514 exists for a reason.
5. Aiming for hardness with the filler. On hardened steel, more than anywhere else, the weld should be a tough hinge, not a harder-than-everything plug. Matching or undermatching the filler to the lowest-strength member is the correct play.
6. Inspecting too early. Hydrogen cracks are delayed by hours. A weld that passes at the end of the shift is not proof of anything; the 16–48 hour window is exactly when the damage shows up. If the job is critical, schedule the real inspection after the delay.
7. Leaving arc strikes and craters. Every scratch of the arc off the joint is a quench site. Every crater is a crack starter. On hardened steel these details stop being cosmetic and start being structural.

A Field-Ready Checklist for Welding Hardened Steel

When the part is on the bench and your hand is on the torch, the following sequence is the difference between a repair that holds and a repair that announces itself with a rifle crack at midnight. Print it, laminate it, and defy anyone to skip a line.

1. Identify the material. Mill sheet, grade mark, or — in its absence — spark test, file test, hardness test, and history. Know whether you are holding A514-type Q&T plate, a through-hardened tool steel, a case-hardened component, or work-hardened manganese. The wrong identification means the wrong procedure, and the wrong procedure costs a part.
2. Decide whether welding is even the right answer. Structural Q&T plate: proceed. A repairable tool with a post-weld temper available: proceed with discipline. A fatigue-critical gear or spring or a case-hardened part whose case you cannot remove: seriously consider replacement or mechanical repair instead.
3. Verify the procedure. On code work, confirm the WPS and PQR cover the grade and thickness. Confirm the preheat minimums, the interpass ceiling, the heat-input band, and the filler classification (including the hydrogen designator) before you strike an arc.
4. Remove the case or the damaged material. Grind away any hard case, decarburized skin, cracked material, or contamination. Open cracks into rounded U/V excavations. Verify with a file that you have reached weldable base metal.
5. Clean and dry the joint. Degrease, grind to bright metal, and keep moisture out. This step is hydrogen control as much as it is cosmetics.
6. Set up low-hydrogen consumables. Baked low-hydrogen electrodes from an oven or rod keeper, sealed wire, dry TIG rod. No cellulosic rods, no damp 7018, no oil-contaminated filler.
7. Apply preheat across the joint zone — not just a spot under the arc — to the specified minimum, measured at the right distance (at least the thickness of the thickest member, and never less than 3 in from the weld). Maintain it through every tack and pass.
8. Weld inside the heat-input window. Steady travel speed, controlled amps and volts, fill the crater on every bead, no arc strikes off the joint. Watch the interpass temperature and let the part cool if it pushes the ceiling.
9. Cool slowly and hold where specified. Post-weld hold at preheat for two to four hours on crack-sensitive steel, then slow cool under a blanket or in still air. Never quench the weld.
10. Temper or post-heat if the material requires it. Tool steels and many hardened repairs want a post-weld temper; Q&T structural steels like A514 want their hydrogen-depleting cool-down instead of a stress-relief soak (because stress-relief PWHT destroys their temper). Your procedure tells you which.
11. Inspect after the delay. Visual, and where the job warrants it, NDT — scheduled 16 to 48 hours after welding, when hydrogen cracks actually reveal themselves. A weld inspected only at quittin’ time has not been inspected.

That checklist is undramatic on purpose. Every line is boring, and every line is load-bearing. Marcus’s bucket skipped lines one through eleven and paid for the whole list at once.

After the Weld: Tempering, Post-Heat, and the Inspection Timing

Post-weld treatment is where amateur and professional diverge completely, so it deserves its own section. Two phrases get mangled constantly on this topic, and the difference between them is money in the bank.

Post-Weld Heat Treatment (PWHT) vs. Post-Weld Temper

Stress-relief PWHT is the classic slow bite-him: heat a weldment to roughly 1,100–1,250°F (595–675°C), hold, and cool slowly to relax residual stress — usually then and there, the correct move for carbon and low-alloy structural steels. On hardened steel, it is frequently the wrong move. Quenched-and-tempered grades like A514 were tempered by the mill at a lower temperature than standard stress-relief practice, so running a conventional stress-relief soak on a finished A514 weldment overtempers the base metal and can drag a 100 ksi plate down toward 80 ksi territory in the heated zone. That is why the professionals’ instruction on A514 is blunt: stress-relief PWHT is not a routine option; it is either prohibited or must be re-qualified with the PWHT baked into the PQR. For hardenable tool steels, the mirror-image logic applies but with a different answer: those steels need a post-weld temper — a controlled, time-at-temperature tempering cycle — to convert the brittle untempered martensite of the HAZ into something tough before the part goes under load.

So the mental model is: do not assume “stress relief” helps hardened steel. On Q&T structure, the correct post-weld therapy is hydrogen-depleting slow cooling at preheat and a controlled cool-down, not a soak. On tool steel, the correct therapy is a tempering cycle matched to the tool’s service requirements. Same five letters, opposite intent.

The Delayed-Crack Window Is a Schedule, Not a Superstition

Because hydrogen cracking is delayed, inspection scheduling is part of the welding procedure. On hardened and hardenable steels, guidance consistently puts the meaningful inspection time hours after completion — commonly 16 to 48 hours, sometimes longer for the most sensitive materials and for high-CE vintage steels. If your inspection is worth doing at all for a fatigue- or safety-critical joint, it is worth doing after the crack had its chance to appear. This is not paranoia; it is physics. The hydrogen that arrived with the weld needs time to concentrate, the stress needs time to do its work, and the martensite needs to be cold enough to be brittle before it cracks. None of those happen while the weld is warm and everyone is watching.

Frequently Asked Questions About Welding Hardened Steel

Every question below is one I have been asked, or one that shows up verbatim in welding forums, with the honest, procedure-first answer.

Can you weld hardened steel at all?

Yes, with a properly qualified procedure — but “weld it” and “restore it to original hardness” are not the same sentence. Quenched-and-tempered plate welds well when preheated and welded low-hydrogen within a heat-input band. A hardened tool repaired by welding will have a softened heat-affected zone and a hard, crack-prone zone unless you preheat, weld carefully, and temper afterward.

Why does hardened steel crack after welding?

Almost always hydrogen-induced cold cracking: hydrogen from the arc, a hard martensitic heat-affected zone from fast cooling of hardenable steel, and tensile stress from shrinkage. Remove any one of the three — usually by preheating to slow the cooling and using dry low-hydrogen consumables to cut the hydrogen — and the crack cannot form.

What happens to the hardness when you weld hardened steel?

Two things at once. Near the weld, part of the heat-affected zone gets overtempered and softer than the base metal; another part gets heated above its critical temperature and quench-cools into harder, brittle untempered martensite. Both are damage. This is why welding hardened steel is discouraged anywhere the joint will carry the part’s full service load.

Do you need to preheat hardened steel before welding?

For nearly every genuinely hardened steel: yes. It slows cooling, dries the joint, reduces thermal shock, and — critically — is what prevents the hard, crack-prone microstructure and gives hydrogen time to escape. The exception that proves the rule is work-hardened manganese steel, which must not be preheated and welds to its own rulebook.

What rod do you use to weld hardened steel?

Low-hydrogen first, and strength-matched second. For A514 the classes are E11018-M stick, ER110S-1 wire, or equivalent low-hydrogen flux-cored classes; for joining hard plate to mild steel, an undermatching 70-class filler; for tool-steel repairs, typically matching TIG filler or a tough low-alloy build-up with a hard cap; and as a last-resort rescue filler, austenitic stainless (309-type). Never cellulosic.

Can you weld hardened tool steel and keep it hard?

You cannot keep the heat-affected zone hard by welding — the weld was already going to soften it. What you can do is weld the tool, then re-harden and temper the whole tool (or just temper, if the repair doesn’t demand full re-hardening). That is why tool-steel repairs go through the heat-treatment shop, not just the weld bay.

Can you weld case-hardened steel?

Only after removing the hardened case in the joint area down to the weldable core. Weld the case and you are welding to a brittle shell that will crack along its own boundary. Grind, verify with a file, weld the core, and — when the part must keep its surface — re-case or re-harden the affected area afterward.

What is the 350 HV number I keep hearing about?

It is the widely used acceptance threshold for heat-affected-zone hardness in the weldability world: below about 350 Vickers hardness, hydrogen cracking is generally not expected. Above it, you are in the risk zone that preheat, low hydrogen, and delayed inspection exist to manage. The exact critical hardness depends on hydrogen level and composition — with very low hydrogen consumables the tolerable hardness rises, roughly into the 400 HV neighborhood — but 350 HV is the industry’s working line in the sand.

Is it better to weld first and harden after?

Where the part’s design allows it, dramatically yes. Welding done on annealed, soft base metal dodges nearly every failure in this article — no hard HAZ, no martensite quench, minimal hydrogen risk — and then the finished weldment is hardened by the heat-treatment shop with the weld inside the process. This is standard practice for many hardenable components and custom tooling. Welding hardened steel is what you do when the sequence can’t be arranged; it is never plan A.

How long do I wait to inspect a weld on hardened steel?

Plan the inspection 16 to 48 hours after welding for hydrogen-crack-prone materials, and longer where the specification or history demands it. A same-day visual sign-off is a warm-metal inspection; the crack is a cold-metal event.

What is the difference between welding hardened steel and hardfacing?

Welding hardened steel joins or repairs a hard parent metal; the steel’s hardness is the obstacle. Hardfacing deposits a deliberately hard, wear-resistant overlay; the hardness is the goal. They share base-metal discipline (preheat, low hydrogen, controlled heat) but diverge in filler selection and in what “success” means — a sound joint versus a wear-resistant surface.

Final Thoughts: Respect Is a Procedure List

Let me end where I started — with the loader bucket. When I last saw that shop, the failed repair was freshly ground out, and the corrected procedure was written on a whiteboard above the bench in four lines: Identify. Preheat. Low-hydrogen. Delay the inspection. The man doing the repair was not a hero and not a genius; he was just tired of the rifle-shot sound of steel that was welded without respect.

That is the whole secret of welding hardened steel. It is not a mystery material that defies the arc; it is a material whose microstructure answers to temperature and hydrogen, and nothing more. Give it a real preheat, feed it dry low-hydrogen metal inside a controlled heat-input window, cool it slowly, and treat the 400°F ceiling as a wall — and there is almost nothing in the hardened-steel family you cannot weld. Skip a step, and the steel will not argue. It will simply wait until 2:00 a.m., go cold, and crack — taking a few weeks of your schedule with it.

If you are new to welding and you have read this far because a hardened part is your first serious project, don’t start there — practice the fundamentals on mild steel first. When you do move up to high-strength plate, the practical mechanics of running low-hydrogen electrodes at the right amperage are worth reviewing before the preheat lamp comes on. And whether you are welding a hardened edge or a wear overlay, weld in a shop that respects the smoke: hardened and high-alloy steels intensify fume exposure, and the same discipline you owe the metallurgy you owe your lungs.

Bore through the material sheet, honor the temperatures, and the steel will let you win. Every hardened-steel crack you have ever heard about started as a skipped step. The fix is right there, in the procedure.

Authoritative Sources (Outbound References)

• AWS D1.1/D1.1M:2025, Structural Welding Code — Steel (qualification, preheat and interpass rules, quenched-and-tempered steel provisions): https://pubs.aws.org/Download_PDFS/D1.1-D1.1M-2025WebPV.pdf
• TWI Global, Job Knowledge — Defects: Hydrogen Cracks in Steels — Prevention and Best Practice (hydrogen cracking, preheat, post-heating, austenitic consumables): https://www.twi-global.com/technical-knowledge/job-knowledge/defects-hydrogen-cracks-in-steels-prevention-and-best-practice-046
• U.S. Department of Transportation / U.S. DOT ROSAP, Development of Heat-Affected Zone Hardness Limits for In-Service Welding (350 HV critical hardness, hydrogen effects): https://rosap.ntl.bts.gov/view/dot/34520
• SSAB Welding Handbook (high-strength Hardox/Strenx welding: hydrogen control, preheat, CET carbon equivalent): https://www.ssab.com/-/media/Handbooks/Welding%20Handbook/ssab-welding-handbook-ed2.pdf
• OSHA 29 CFR 1926.353, Ventilation and Protection in Welding, Cutting, and Heating: https://www.osha.gov/laws-regs/regulations/standardnumber/1926/1926.353
• CDC/NIOSH Pocket Guide to Chemical Hazards — Welding Fumes: https://www.cdc.gov/niosh/npg/npgd0666.html
• Lincoln Electric, public reference library on low-hydrogen electrode storage and handling (SDS guidance referencing ANSI Z49.1 and NFPA 51B): https://www.lincolnelectric.com/assets/US/EN/MSDS_lib/ZLE_SDS_NA-EN-200000000146.pdf
• Miller Electric, Welding Safety & Health Guide (fume extraction and respiratory protection for high-alloy/hardening steels): https://www.millerwelds.com/-/media/inriver/miller-safety-guide-vol-ii-interactive.pdf