Lack of Fusion in Welding: Causes, Detection, Prevention, and Repair

Introduction: The Weld Defect You Can’t See

The weld looked flawless. Smooth ripples, even toe lines, no visible cracks or porosity. The fabricator had every reason to sign off on the job. Then the ultrasonic technician flagged a planar defect running the full length of the joint: a bond that never actually formed. The two plates had touched, melted along the surface, and cooled into what looked like a sound weld. Underneath the cap, the metal never fused.

That scenario plays out in fabrication shops, pressure-vessel yards, and pipeline spreads more often than most people realize. Lack of fusion is one of the few weld defects that can pass every visual check and still leave a structure one fatigue cycle away from failure. It hides under an attractive bead. It doesn’t announce itself with spatter or undercut. It sits there, planar and sharp, waiting for a load to find it.

What Makes Lack of Fusion Different

Lack of fusion — often abbreviated LOF — is classified as ISO 6520-1 reference 401, a planar imperfection. That word “planar” is the whole story. A planar defect is flat and crack-like, with a sharp edge that concentrates stress. Volumetric defects like porosity or slag inclusions are rounded and dispersed; they reduce cross-section but rarely act as crack starters. Planar defects do the opposite. Codes know this, which is why lack-of-fusion defects are treated far more strictly than volumetric defects at every quality level.

That distinction shapes everything downstream — how you inspect, how you qualify a procedure, and whether you repair or scrap. A single porosity pocket might pass acceptance criteria in a given weld class. A lack of fusion of similar size can be a reject, and sometimes a mandatory repair, no matter how clean the rest of the joint looks. ESAB’s welding defects guide is a useful reference for how codes separate these imperfection families.

Talk to any welding inspector and you’ll hear the same thing: visual inspection is the cheapest tool you own, but it has limits. A weld can pass the naked eye and the fillet gauge and still fail ultrasonic testing by a wide margin. That gap between appearance and reality is exactly where lack of fusion lives.

If your visual inspection skills aren’t sharp, you’ll miss the setup problems that create this defect before the arc even strikes. Our guide to good and bad MIG welds walks through what a healthy bead should actually look like, so you can recognize the warning signs that something is wrong beneath the surface.

What This Guide Covers

This article takes lack of fusion from definition to repair. We break down the types — sidewall, inter-run, root, and cold lap — and explain how each one forms. You’ll see the root causes: heat input, travel speed, current, and arc length, plus the fit-up, access, and contamination issues that quietly sabotage fusion. We also compare how MIG, TIG, stick, FCAW, SAW, and aluminum welding each invite the defect in different ways.

From there we cover detection and inspection, including ultrasonic and radiographic methods that find what the eye cannot, per TWI’s overview of sidewall and inter-run fusion defects. Finally, we cover prevention, WPS control, repair, and the acceptance criteria that govern each code.

The takeaway is simple: a good-looking weld is not the same as a sound weld. Learn what sits underneath the bead, and you’ll stop trusting the surface.

What Is Lack of Fusion? Types and How It Differs From Other Defects

Lack of fusion (LOF) is the failure of weld metal to bond with the adjacent parent metal or with a previously deposited weld pass. The filler and the base metal sit physically close, sometimes touching, but the boundary between them never becomes one continuous metallic structure. No fusion means no metallurgical bond — and a joint that is not bonded is not carrying load the way the drawing assumes it is.

The Three Types of Lack of Fusion

International standards recognize three distinct locations where fusion can fail, based on the International Institute of Welding (IIW) reference system:

• Sidewall lack of fusion (IIW 4011): the weld metal does not bond to the vertical face of the joint preparation. The bead sits beside the bevel wall instead of merging into it.
• Inter-run lack of fusion (IIW 4012): a later pass fails to bond to the pass beneath it. In multi-pass welds the layers stack up, but the interface between two runs stays as an unbonded plane.
• Root lack of fusion (IIW 4013): the first pass never fuses into the root face or the root gap, leaving the deepest part of the joint unconnected.

The IIW classifies all three under reference 401. That is deliberately separate from lack of penetration (IIW 402), which is the failure of the weld metal to reach and fill the required root depth. Penetration is about how far the weld goes; fusion is about whether it sticks to what it touches. A weld can achieve full penetration and still suffer sidewall lack of fusion at the same time, which is why the two get confused on the shop floor.

Why It Behaves Like a Brazed Joint

When two faces merely stick together, the joint acts less like a welded connection and more like a brazed one. In brazing, filler is distributed between surfaces by capillary action and adhesion, without melting the base metal. A lack-of-fusion boundary is similar: the materials touch, but there is no melting, no grain growth across the interface, and no metallurgical bridge. Under load, that plane is a ready-made path for a crack.

The gap does not have to be large. Internal stresses from cooling and restraint can open a sub-millimeter discontinuity that, on a radiograph or in a section, looks exactly like a crack. Inspectors describe these planar features in TWI’s guidance on sidewall and inter-run fusion, and it is why fusion defects demand the same seriousness as cracking.

The Cold Pocket at Run Corners

Multi-pass welds are not smooth ramps; they have corners where one run meets the next. Those corners sit at the greatest distance from the arc, so the arc energy arriving there is lowest. That creates a cold pocket — a zone that cools before the metal has fully melted and merged. Run corners are the most critical zones for lack of fusion precisely because they sit furthest from the heat source.

Fusion vs. Penetration: A Plain Contrast

Both can hide under a good-looking cap, which is why visual inspection alone is unreliable. Reading good and bad MIG welds teaches you what a healthy bead should look like, but a neat surface is not proof of fusion underneath. As you build skill through deliberate welding practice, you start to feel and hear when the arc is properly wetting the sidewalls — a cue no inspection photo can give you.

Why Lack of Fusion Is Dangerous: Strength, Fatigue, and Real Failures

A properly made weld fails only when the applied load finally outruns its strength. A weld riddled with lack of fusion can let go while the surrounding steel is still loafing along at a fraction of its rated capacity. That gap between “looks strong” and “is strong” is the entire problem — and it is why inspectors treat this defect so seriously.

The danger comes down to geometry. Where metal should have melted together and become one, it instead stops dead, leaving a sharp interface. That shape, not just the missing material, is what turns a small flaw into a serious one.

The Notch Effect: A Crack in All but Name

Where two surfaces never fused, the interface tapers to a razor-thin point. In engineering terms, that geometry is a notch, and at its tip stress piles up to many times the nominal stress carried by the part. The sharp, crack-like geometry acts like a pre-existing crack built into the joint. Because of this notch effect, a crack can propagate from a lack-of-fusion flaw under the smallest applied load — sometimes a load the structure was designed to shrug off many times over.

This is the cruel twist. The flaw can sit dormant for years, passing every visual inspection, until one day a routine overload finds the tip and starts a tear that nothing can stop. Rounded defects rarely behave this way; sharp ones do.

Fatigue: Where Planar Defects Do the Most Damage

Under cyclic or dynamic loading, the unfused interface opens and closes with every load cycle, concentrating strain right at the tip. That is exactly what fatigue cracks want. A discontinuity that might sit quietly under a steady pull becomes a crack starter under vibration, traffic, wave action, or pressure cycling, and the life of the component can drop sharply. A related geometry problem worth understanding is how an overlap welding defect creates a comparable stress-raising notch where weld metal rolls onto base metal without ever bonding to it.

Static Loading and Load-Carrying Capacity

Even without cycling, lack of fusion eats into strength. Studies confirm that lack of fusion considerably lowers the load capacity of a permanently statically loaded structure. Researchers and codes consistently treat lack-of-fusion defects as cracks under permanent loading, potentially leading to final structural failure. Undetected lack of fusion is cited again and again as a cause of welded structure failures, from cranes and bridges to pressure equipment.

Pressure Equipment: Leak Paths and Worse

In vessels, tanks, and pipelines the stakes climb higher. An unfused interface is a ready-made leak path, and if it links up through the wall it can release stored energy all at once in a catastrophic rupture. Even a flaw that only partly penetrates can grow under service stress until it finally breaches the wall. This is why structural welding codes scrutinize fusion discontinuities far more aggressively than rounded defects such as porosity.

What the Codes Say, and What Rework Costs

Because of that crack-like behavior, certification and inspection regimes single these flaws out. AWS D1.1 and the ASME Boiler and Pressure Vessel Code treat fusion discontinuities as planar defects and often judge them rejectable where a rounded indication of the same size would pass. Welders testing for certification learn quickly that lack of fusion is one of the fastest ways to fail a bend or radiograph.

The economics follow the physics. A rejected joint means gouging out sound metal, re-preparing, re-welding, and re-inspecting — plus the schedule slip while a fabrication bay sits idle. Add downstream liability, warranty claims, and possible recalls when a flaw survives into service, and the true cost dwarfs the few seconds of extra heat and travel control it would have taken to fuse the joint properly the first time.

Root Causes: Heat Input, Travel Speed, Current, and Arc Length

Every case of lack of fusion traces back to one physical fact: the sidewall or the previous run never reached fusion temperature before the molten metal arrived to cover it. Steel doesn’t bond by contact alone. The parent metal has to reach its melting point so grains can grow across the boundary and form a genuine metallurgical bond. If the arc doesn’t push enough energy into that boundary, the liquid pool just sits on top of cold, solid steel. You end up with a bead that looks joined but is really resting in place.

Heat input is the arc energy delivered per unit length of weld, usually expressed in kilojoules per millimeter (kJ/mm). Thermal efficiency is the fraction of that energy that actually reaches the workpiece; the rest is lost to radiation, spatter, and the surrounding air. Arc processes like stick, MIG, and TIG run roughly 0.6 to 0.85 efficient, while submerged arc welding, with the arc buried under flux, approaches 1.0 because almost nothing escapes.

Heat input is computed from real measurements, not dial settings:

Heat input (kJ/mm) = (Voltage × Current × 60) ÷ (Travel speed in mm/min × 1000)

The voltage is true arc voltage, and the current is measured amps at the arc, per pass. Nominal figures lie. A machine set to 200 A may deliver 175 A under load, and the voltage printed on a procedure may not match what the arc actually sees. Travel speed carries the largest error of all, because it’s both the hardest value to hold steady and the one sitting in the denominator, so small drift swings heat input dramatically.

Once you understand that equation, the causes of lack of fusion line up neatly:

• Travel speed too high. Energy is spread over more length, so energy input per unit length drops and the parent metal never melts. This is the most common single cause.
• Current too low. Less energy enters the joint, and the arc lacks the punch to dig into sidewalls. Undershooting the qualified amperage is a classic trap.
• Arc length too long. The arc fans out, voltage rises, and heat spreads across a wider area instead of concentrating into the root and sidewalls.
• Wire feed speed and voltage mismatch. The arc shifts out of a stable spray or globular transfer into erratic running that never wets the edges.
• Wrong wire extension (stick-out). Too long a stick-out preheats the wire resistively and consumes energy before the arc. Contact tip condition matters too, since a worn tip arcs off-center and misdirects heat.

All of these reduce real energy at the fusion boundary. Cold wire extension, undersized leads, and a weak power source can quietly cut your usable amps even when the display reads correct. That’s why properly sized welding leads and clean connections belong in any root-cause investigation. And when you do raise the current, stay inside the qualified range for the electrode you’re running, whether that’s 3/16 E7018 amperage for stick or a spray-transfer setting on MIG.

The Counterintuitive One: Too Slow Can Also Cause Lack of Fusion

You’d think slowing the travel speed guarantees fusion. Past a point, it does the opposite. A too-low travel speed creates an oversized molten pool that runs ahead of the arc, riding over the edges before they’ve reached temperature. The arc then plays over liquid metal instead of the cold sidewall, and no bond forms. This is why welding practice on scrap plate matters more than dialing numbers: you learn to feel when the pool is running ahead of you.

Two more quiet culprits: a machine with poor arc stability at low current, and a gun angle that pushes the arc toward the puddle instead of the joint. Both cut real energy at the sidewall without changing any dial.

In short, both extremes fail. Travel too fast starves the joint of energy; travel too slow floods it and lets the pool beat the arc to the edge. The qualified window exists to keep you out of both traps.

Joint Preparation, Fit-Up, Access, and Contamination

Some lack-of-fusion causes live in the machine settings, but plenty live in the joint itself. If the groove is too narrow, the included angle too tight, the root gap wrong, or the fit-up sloppy, no amperage setting will save you. Defect references such as TWI’s guide to lack of sidewall and inter-run fusion list poor joint design and fit-up as primary contributors, and they are right.

When the Groove Itself Causes the Defect

Geometry controls where the arc goes. When a groove is too narrow, the arc tends to climb and attach to whichever sidewall is closest. That wall melts; the arc never reaches across the gap, so the opposite wall stays cold and unfused. The result is lack of sidewall fusion on one side even though the weld looks full and tidy.

A narrow groove can also flood. Molten metal fills the gap faster than the walls reach bonding temperature, so the puddle bridges the opening before the base metal melts. You end up with a cold, unbonded interface hidden under an apparently sound bead: the classic attractive weld with nothing behind it.

The bevel angle and root face (land) set how far the arc can reach. Too heavy a land or too tight a gap blocks penetration; too wide a gap invites burn-through. Misalignment adds another problem. A high-low step creates a shoulder the arc cannot cross, so one side of the joint never fuses.

Fit-Up Checks Before You Strike an Arc

• Included angle and bevel machined to the drawing, not eyeballed
• Root gap and land measured, not guessed
• Joint faces square and aligned, with high-low within tolerance
• Backing or backing gas confirmed for full-penetration work
• Groove wide enough for the nozzle, cup, or electrode you plan to run

Tubing and piping jobs punish sloppy fit-up hardest, so the fit-up and welding guide for tube fittings is worth studying for how small gaps and mismatched bores drive fusion defects.

Consumable and Nozzle Access

Consumables and nozzles need physical room. In MIG/MAG and FCAW, a narrow groove may not provide clearance for the shielding gas nozzle diameter. If the nozzle is jammed against a sidewall, gas coverage collapses and you lose torch angle control. Matching wire and cup size to the groove, and keeping the contact tip clean and correctly sized so wire feed stays smooth, matters as much as voltage. Wire too large for the prep forces excessive weaving, and weaving with poor aim is exactly how welders miss sidewalls.

Stick welding has the same limit in bigger packaging. A 4 mm or 5 mm electrode simply cannot follow a cramped groove; the flux drags the sidewall and the arc wanders. Drop to a smaller diameter, or open the prep.

Contamination: The Invisible Bond Breaker

Even with a perfect groove and ideal heat, contamination kills bonding. Rust, mill scale, oil, paint, and moisture all sit between the molten weld and the base metal. Insufficient cleaning of oily or scaled surfaces contributes directly to lack of fusion. The arc may burn through light surface rust, but the oxide remains and the interface never truly joins. Guides like weldability-sif’s lack-of-fusion reference flag this as one of the most common and most avoidable causes.

Oily or painted surfaces are the worst offenders, because the arc burns the contamination into the joint instead of removing it. Degrease first, then grind or wire-brush back to bright metal. A quick visual check before welding, plus a clean dedicated brush for stainless, prevents hours of rework.

Between passes, slag and oxide films are the quieter problem. A slag or oxide film between runs prevents metallurgical bonding even when thermal conditions are correct. Multi-pass welds are especially vulnerable because every layer adds a new surface to clean. Chip, brush, and where needed grind each run before the next.

Process-by-Process Comparison: MIG, TIG, Stick, FCAW, SAW, and Aluminum

Lack of fusion is characteristic of gas-shielded arc welding processes with consumable electrodes, but every process carries its own version of the risk. The same defect looks different depending on how the arc, the filler, and the shielding meet the base metal. Here’s how the major processes compare.

GMAW/MIG and MAG

Gas metal arc welding is the classic home of lack of fusion. With a continuously fed consumable electrode, the process rewards a welder who holds a steady travel speed and torch angle, and punishes one who doesn’t. Push the travel speed too high and the arc races ahead of the molten puddle, so weld metal bridges over cold, unmelted sidewalls. Wire feed issues, a worn contact tip, or the wrong polarity all shrink the heat that actually reaches the joint root. If your beads look good but won’t tie in, check your contact tip and travel consistency before blaming the machine.

GTAW/TIG

TIG is slow, precise, and easier to control, until the operator gets too conservative. Low amperage, insufficient filler addition, and narrow bevels leave sidewalls cool and unwetted. On thinner material a welder may run just enough current to look tidy while the toes never properly fuse. Polarity matters here: choosing the right DCEN, DCEP, or AC setup directly affects how much heat reaches the work. For carbon and stainless steel, see our breakdown of AC versus DC for TIG welding steel. Too tight a joint gap is another common culprit, because the arc can’t reach both faces.

SMAW (Stick)

Stick welding brings two special risks: arc blow and slag entrapment between passes. Arc blow pushes the arc off course, especially at high currents or on magnetic materials, so one sidewall never fuses. Slag left behind between passes acts as a barrier that new metal simply flows over, so clean every pass thoroughly. On cellulosic and downhill procedures the risk climbs, because there is a high chance of lack of fusion in downhill welding as the puddle runs away from the heat. Correct amperage for the electrode, such as a 3/16 E7018 setting, keeps the arc driving into the root rather than skating across it.

FCAW

Flux-cored welding behaves much like MIG for fusion, but the flux and polarity change the picture. Getting DCEN versus DCEP right for flux core is essential: the wrong polarity softens penetration and encourages cold lap. Gasless varieties are especially sensitive because the self-shielded slag can sit between the metal and the arc. Our guide to gasless MIG and flux-cored welding without gas covers the fundamentals. Remember that self-shielded wire runs on DCEN, a detail that trips up welders used to standard MIG.

SAW (Submerged Arc)

Submerged arc welding deposits metal fast, and that high deposition rate is exactly what makes it vulnerable. Deep, narrow grooves plus high current and travel speed produce sidewall fusion problems that hide under the flux blanket. Automated and robotic submerged arc welding setups demand tight control of joint fit-up, wire position, and flux handling, because a machine won’t feel the puddle cooling off. A drifting wire position is one of the most common causes of inter-run fusion defects in SAW.

Aluminum

Aluminum raises the stakes because its oxide layer melts at a far higher temperature than the metal beneath it. In arc welded aluminium the rules are blunt: under BS EN ISO 10042:2005, long imperfections are not permitted at any of the three quality levels. Because aluminium conducts heat away so quickly, it’s also prone to cold starts and poor tie-in. DC welding of aluminium and correct AC balance both help, but technique remains the deciding factor.

Resistance and Gas Welding

It isn’t only arc processes. Resistance welding can suffer lack of fusion when current, time, or electrode force is wrong, leaving a weak bond at the interface. Gas welding is vulnerable too, particularly when the flame is too small or the travel too fast, so the puddle never wets both edges. The underlying causes stay the same across all of them: enough heat, in the right place, for long enough to melt both faces.

Common Mistakes Welders Make (and How to Fix Each One)

Most lack-of-fusion failures trace back to a handful of habits that feel efficient in the moment, and each one has a concrete fix. None require new equipment, just better habits. Here are the eight we see most often.

1. Aiming the Arc at the Pool Instead of the Leading Edge

When you drop the arc into the molten pool, the arc energy never reaches unmelted parent metal, so the sidewalls stay cold and unbonded — a classic sidewall fusion defect. Fix: aim at the leading edge of the pool, where the arc digs into the joint faces. Verify by watching the puddle wash into both walls, then macro-etch a test coupon to confirm fusion.

2. Not Weaving or Dwelling at the Sidewalls

A straight stringer bead can leave the toes starved of heat. Weaving with a brief dwell at each sidewall lets the pool wash into the parent metal and greatly improves sidewall fusion, as TWI explains. Fix: add a controlled weave with a short pause at each wall. Verify with a side-bend or macro test; the fusion line should show a continuous bond, not a sharp notch.

3. Estimating Travel Speed Instead of Measuring It

Too fast and the arc outruns the melt; too slow and you overheat the puddle. Both wreck fusion. Fix: clock your travel against the WPS range using a stopwatch over a marked length. If your bead overlaps you’re moving too slow; if you see undercut at the toes, you’re too fast. Verify by checking bead width and heat input against the procedure.

4. Running an Arc That’s Too Long

A long arc spreads heat and loses penetration, leaving cold lap at the root. Fix: tighten the arc until you hear a crisp, consistent crackle. Verify with a stable, narrow bead and a penetration check on a coupon.

5. Skipping Interpass Cleaning

Slag, spatter, and oxides trapped under the next pass become instant lack-of-fusion planes. Fix: wire-brush or grind every pass, especially in flux-cored work. On stainless and aluminum, use a dedicated stainless brush to avoid cross-contamination. Verify by inspecting the cleaned surface; it should be bright metal with no slag islands.

6. Using an Oversized Electrode or Nozzle in a Tight Groove

If the electrode or nozzle can’t reach the joint root, the arc can’t either. Tight V-grooves and narrow gaps are the classic trap. Fix: switch to a smaller diameter or a longer-reach consumable, and confirm the wire aims where it must. Verify by watching the arc reach the root, and by checking gas coverage with a fresh contact tip if the nozzle is marginal.

7. Fighting Arc Blow Instead of Fixing It

Arc blow from an uncontrolled current-return clamp position deflects the arc and causes lack of fusion. Fix: reposition the current return clamp, switch to AC in MMA, or demagnetize the steel. Verify by watching the arc stabilize, then re-inspect the joint.

8. Blaming the Operator When Fit-Up or Design Is the Problem

Gaps, misalignment, and inaccessible joints defeat even good technique. Fabricators who burn hours chasing fusion on a joint the design never made accessible are solving the wrong problem. Fix: correct fit-up and access, or redesign the joint so the arc can reach the root. Verify with a pre-weld fit check before the first pass.

Individually these mistakes look small. Together, they’re the leading recipe for the defect. Skill is the common denominator behind every fix above, and deliberate, structured welding practice builds the arc control and timing that keep these habits from forming in the first place.

Detection and Inspection: How to Find a Defect You Can’t See

Lack of fusion is a nightmare for inspectors because it doesn’t announce itself. The bead looks sound, the profile is uniform, and there’s no open crack or porosity to catch your eye. The bond simply never formed — and that flat, unbonded plane sits inside the metal waiting for a load to pull it apart. Finding it takes the right method.

Why is it so slippery? Because lack of fusion is planar — a two-dimensional flaw squeezed into a three-dimensional joint. A void or inclusion gives an inspector a round reflector to bounce sound off; a tight, unfused plane may give only a weak, ambiguous return. Even a large lack of fusion can produce a surprisingly faint indication, which is why untrained eyes and outdated techniques overlook it constantly.

Why the “Easy” Methods Fall Short

Start with what doesn’t work. Visual testing (VT) only sees the surface, so a buried planar defect is invisible to it. Liquid penetrant (PT) and magnetic particle (MT) find surface-breaking flaws well, but they only catch lack of fusion that has reached the surface. For inter-run and sidewall fusion that hides mid-wall, they’re the wrong tool.

Radiography (RT) is often the default for volumetric inspection and it does catch larger root and sidewall lack of fusion and slag inclusions. But RT has a well-known blind spot: if the unfused plane runs parallel to a beam, the X-rays pass right through the gap with little differential absorption, and the defect disappears. When the plane is unfavourable to the beam, radiography can miss a planar defect entirely.

Ultrasonic Testing: The Go-To for Planar Defects

Where RT struggles, ultrasonic testing shines. Ultrasonic inspection is normally more effective than radiography in detecting lack of inter-run fusion, because ultrasound reflects off the flat, mismatched interface that radiography happily transmits through. Probe choice matters:

• Angle probes (shear-wave) are efficient for sidewall lack of fusion, admitting sound at an angle to hit the vertical bond line.
• A normal probe with a ground face is used for inter-run lack of fusion, where the flaw lies roughly parallel to the surface.

Modern practice leans on codes. As TWI explains, planar flaws are technique-sensitive and demand the right probe and scan pattern. Two standards anchor the work: ISO 17640 governs manual ultrasonic testing of welds, while ISO 13588 covers automated phased-array UT. Phased array (PAUT) is now the leading route for planar defects — it steers and focuses the beam electronically, sweeping multiple angles from one position and mapping the flaw in detail. Acceptance is then judged against levels in ISO 11666.

Advanced UT techniques such as those presented at NDT.net’s world conference also refine sizing of planar discontinuities that RT under-reports.

When Only Destructive Testing Confirms

Here’s the hard truth: pure lack of fusion is a structural defect that in some configurations cannot be detected by NDT at all — only microscopic inspection reveals the un-bonded interface. The only reliable way to detect some lack-of-fusion defects is destructive testing, such as the bend test and macrographic examination. Cut a coupon, polish the section, and etch it; the unfused line shows up under the microscope. That’s why bend tests anchor both procedure qualification and welder testing — they physically open the joint and expose the bond.

The same logic applies to related surface profiles like the overlap welding defect, where excess metal masks an underlying fusion problem until a section is cut.

The Double-Examination Rule

For critical hardware, one method isn’t enough. Pressure vessels frequently demand double examination — RT plus UT — because the two techniques compensate for each other’s blind spots. RT catches volumetric flaws and root geometry; UT catches the planar fusion defects RT slides past. Codes such as ASME B1.11 and AWS D1.1 reflect this layered philosophy. The takeaway for any fabricator: match the method to the flaw. If lack of fusion is your concern, standard radiography alone won’t give you peace of mind.

Prevention, WPS Control, and Repair: Acceptance Criteria Across Codes

Prevention is not one trick. It’s a stack of controls, and lack of fusion shows up whenever any one of them slips. Get the heat input high enough, keep the arc short, and the joint edges will actually melt.

The Prevention Playbook

Every reliable fix traces back to the same fundamentals:

• Heat input: enough current and voltage to melt both sidewalls and the root, not just lay metal on top.
• Short arc, moderate travel speed: a long arc spreads heat thin; rushing the travel leaves cold lap behind.
• Torch and electrode angle: aim so the arc reaches the joint root and both edges.
• Weaving and dwell: pause at the sidewalls where heat input limits allow, letting each edge fuse before moving on.
• Joint prep and access: wider bevels, proper root gaps, and enough room for the torch angle all reduce sidewall fusion defects.
• Bright metal: grind off mill scale, rust, paint, and oil so the arc isn’t fighting contamination.
• Interpass cleaning: slag and oxide between passes are classic causes of inter-run lack of fusion.
• Arc blow control: adjust ground placement, reduce current, or switch polarity to stop the arc wandering off the joint.

Formatting details matter here. Good technique on clean, well-fitted joints is what separates a passing weld from a reject, and it’s built through deliberate welding practice. Polarity choices can also change penetration and fusion behaviour, whether you’re running TIG polarity or flux core DCEN vs DCEP. For high-deposition work such as robotic submerged arc welding, the same rules scale up, with parameters dialled in tightly because the arc is buried and hard to watch.

How a Qualified WPS Locks It All Down

A qualified Welding Procedure Specification encodes these variables as ranges you must stay inside: amperage, voltage, travel speed, polarity, electrode angle, preheat, interpass temperature, and cleaning requirements. The WPS is the contract that turns good intentions into repeatable fusion. Deviate and you’re welding outside qualification. In-process control matters because a welder watching the puddle can catch a cold edge seconds before it becomes a permanent defect, while the WPS alone can’t.

Repair: Excavate, Re-weld, Reinspect

When lack of fusion is found, superficial fixes don’t work. Remedial action normally requires removal by localised gouging or grinding back to sound metal, followed by re-welding per the agreed procedure, then reinspection with the same method that found it. Volume or surface methods such as UT, phased array, or radiography confirm the excavation reached sound metal before re-welding begins. Remove the full length of a planar defect; chasing only the tip leaves a crack-like flaw in place. Overlap defects are repaired in much the same way.

Acceptance Criteria Across Codes

Codes differ, but the logic is consistent. Let’s compare:

Note the split in AWS D1.1: cyclically loaded structures are held to tighter limits than static ones because fatigue cracks initiate at unfused planes. Because fatigue life depends on the unfused plane’s length and orientation, cyclic service codes treat any planar fusion defect far more seriously than a rounded indication. The general rule across all codes: cracks and lack of fusion must always be repaired, while minor porosity or spatter may be acceptable at defined levels.

FAQ and Key Takeaways: Mastering Lack of Fusion

Here are the questions welders and inspectors ask most, followed by the core lessons that tie this guide together.

Frequently Asked Questions

What is lack of fusion? A planar defect where the filler metal and the parent metal, or two adjacent passes, never bonded because the surface wasn’t molten when the metal arrived.

Can you see it on the surface? Rarely. Buried inter-run and sidewall lack of fusion hides under a sound-looking bead, so it often turns up only during volumetric inspection.

Why is it so dangerous? It produces a sharp notch effect that concentrates stress and starts cracks under fatigue, making it one of the most serious weld defects a joint can carry.

How do you detect it? Buried planar lack of fusion relies on ultrasonic or phased array testing, because the defect is often oriented so radiography misses it. Surface-breaking flaws are found with dye penetrant or magnetic particle methods.

Is it the same as lack of penetration? No. Lack of penetration sits at the root where the joint isn’t filled; lack of fusion is a missing bond along sidewalls or between passes. Both are planar, and both demand the same respect.

Which processes are most prone? It appears across MIG, TIG, stick, FCAW, and SAW, but most often when travel speed outruns heat input or the arc misses the sidewall. Aluminum is especially unforgiving, so dial in how polarity shapes that heat with DCEN, DCEP, and AC in TIG welding.

Key Takeaways

Every cause in this guide traces back to one idea: the metal must reach melting point before more metal arrives. Prevention depends on heat input, travel speed, arc aim, joint preparation, and cleanliness. Get those five right and you rarely create the defect at all.

Your Next-Job Checklist

1. Verify parameters against the WPS before the first pass.
2. Measure travel speed instead of guessing it.
3. Watch the pool and the toes — if they aren’t wetting in evenly, slow down.
4. Clean between every pass; remove slag, spatter, and oxide.
5. Control arc blow with proper grounding and shorter arcs.
6. Inspect with the right method for a planar defect, not just a visual once-over.

Lack of fusion rewards prevention, not detection. A few disciplined habits — honest settings, measured travel, and clean metal — keep a sound weld sound from root to cap. Build your skills with deliberate welding practice, and learn to spot trouble early by comparing good and bad MIG welds.

Finally, an inspection-ready shop starts with the space itself. Set up clean extraction, stable power, good light, and room to move with our guide to planning and setting up safe welding facilities. Master the fundamentals, apply the checklist, and lack of fusion becomes a footnote on your next job instead of a finding on your inspection report.

Frequently Asked Questions

What is lack of fusion in welding?

Lack of fusion is a planar defect where deposited weld metal fails to bond metallurgically with the parent metal sidewall or with a previous weld pass. It can occur at the sidewall, between runs, or at the root, and it is classified as ISO 6520-1 reference 401.

What is the difference between lack of fusion and lack of penetration?

Lack of fusion is the absence of a bond between weld metal and base metal or a previous pass, while lack of penetration is the failure of the weld to extend fully through the joint thickness at the root. In the IIW classification, lack of fusion is reference 401 and lack of penetration is reference 402.

Can visual inspection detect lack of fusion?

Usually not. Because lack of fusion is buried and planar, a bead can look completely sound while the defect sits underneath. Visual inspection during welding can catch the conditions that produce it, such as the pool running ahead of the arc, but post-weld visual inspection alone is not reliable.

Which NDT method is best for detecting lack of fusion?

Ultrasonic testing, especially phased array, is the leading route for planar defects. Radiography can find larger root and sidewall lack of fusion but often misses inter-run defects, while penetrant and magnetic particle testing only detect defects that reach the surface.

What causes lack of fusion?

The core cause is too little energy at the bond line when metal arrives, meaning the sidewall or previous run never reached fusion temperature. Contributing factors include low current, excessive travel speed, long arc length, wrong torch angle, narrow joint preparation, contamination, and magnetic arc blow.

Is slow travel speed safer for avoiding lack of fusion?

Not always. While too fast a travel speed causes insufficient melting, too slow a speed floods the groove with an oversized molten pool that runs ahead of the arc and also prevents the sidewalls from melting. The qualified window for travel speed matters in both directions.

How is lack of fusion repaired?

Repair normally involves removing the defect by localized gouging or grinding down to sound metal, then re-welding following the approved welding procedure specification and re-inspecting. The repair must comply with the acceptance criteria of the governing code.

Are any lack-of-fusion defects acceptable under welding codes?

It depends on the code and quality level. Under ISO 5817, incomplete fusion is not permitted at stringent or intermediate levels and only permitted as intermittent, non-surface-breaking defects at the moderate level. AWS D1.1, API 1104, and other codes set their own limits.

Does lack of fusion affect fatigue life?

Yes, significantly. Unfused interfaces act as sharp stress concentrators and crack initiation sites, so they are especially dangerous in cyclically loaded structures. Even a large lack-of-fusion defect can give weak inspection indications, which is why codes treat planar defects so strictly.

How can I prevent lack of fusion on the next job?

Verify parameters against the WPS, measure rather than estimate travel speed, keep a short arc, aim at the leading edge, dwell at the sidewalls, clean between passes, prepare bright-metal surfaces, and control arc blow with return-clamp placement or by switching polarity.

Authoritative Sources & Further Reading

• https://www.twi-global.com/technical-knowledge/job-knowledge/weld-defects-imperfections-in-welds-lack-of-sidewall-and-inter-run-fusion-041
• https://www.ndt.net/article/wcndt00/papers/idn403/idn403.htm
• https://www.ndt.net/article/ecndt2006/doc/P213.pdf
• https://www.ndt.net/article/wcndt2008/papers/72.pdf
• https://pubs.aws.org/Download_PDFS/B1.10M-B1.10-2009PV.pdf
• https://pubs.aws.org/Download_PDFS/b1.11-2000PV.pdf
• https://pubs.aws.org/Download_PDFS/d1.1-d1.1mPV.pdf
• https://www.osha.gov/welding-cutting-brazing
• https://www.cdc.gov/niosh/topics/welding/
• https://esab.com/gb/eur_en/esab-university/articles/welding-defects-guide/
• https://weldability-sif.com/knowledge-centre/guides/defects/lack-of-fusion-and-penetration/