Lack of Penetration in Welding: Causes, Detection, Prevention & Repair

Table of Contents

Why a Weld That Looks Perfect Can Be Dangerously Weak

You finish a root pass with a steady, even bead. The toes tie in cleanly, the ripple pattern looks textbook, and there is no undercut, porosity, or spatter to apologize for. Then the radiograph comes back with a single dark line running along the bottom of the joint, and the weld you were proud of never fused through the root at all.

That dark line is lack of penetration. Its greatest danger is how little it shows on the surface: the weld face can look flawless while the root sits completely unfused.

Lack of penetration (often shortened to LOP) is classified in ISO 6520-1 under reference 402 as a planar discontinuity that forms at the root of a joint. “Planar” is the key word. Unlike a gas pocket, which is rounded and fairly benign, lack of penetration is a flat, two-dimensional flaw that behaves like an internal crack. It hides inside the joint, oriented across the load path, invisible from the side you spent an hour making look good. You cannot see it from the cap, and a casual visual pass will clear it every time.

A Flaw With Crack-Like Geometry

Not all weld imperfections carry equal weight. Porosity and slag inclusions are volumetric, occupying space in three dimensions. Planar flaws are different. After cracks, incomplete fusion and incomplete joint penetration are widely regarded as the most serious weld discontinuities because their flat, crack-like geometry acts as a severe stress concentrator. Under cyclic loading or a sudden impact, that sharp tip becomes exactly where a real crack begins. That is why inspectors treat it with the same suspicion as a crack.

AWS A3.0 defines incomplete joint penetration as a joint root condition in a groove weld in which weld metal does not extend through the joint thickness. In plain terms: the arc melted the joint walls, but the molten metal never bridged all the way through. It is a fusion failure, not a filling failure. The gap at the root stays there, waiting.

Quick Note: The same imperfection can be acceptable in one joint and rejectable in another. Before you condemn a root, read the welding symbol and the welding procedure specification (WPS); full penetration is a requirement you design in, not a default.

Why the Face Lies to You

The visible bead tells you what the arc did at the surface, not at the bottom of the joint. Full penetration depends on root gap, land, heat input, travel angle, and arc position, variables the finished cap completely conceals. Even something as ordinary as contact-tip-to-work distance shifts where the arc energy lands. A welder can produce a beautiful cover pass over a root that never fused, because the two things are measured at different depths.

What This Article Covers

The rest of this guide walks through the whole picture. We will define LOP and its terminology, break down the root causes, and explain the structural and safety stakes. Then we compare inspection methods, cover prevention through fit-up, parameters, and technique, and look at how the problem differs across SMAW, GMAW/FCAW, GTAW, and SAW. You will learn how LOP differs from lack of fusion, overlap, and high-low, plus how to repair it without making things worse. Detection, prevention, and repair each get their own focus. Along the way we lean on solid fundamentals, the kind you build through deliberate welding practice rather than guesswork.

What Is Lack of Penetration? Definitions and Terminology

Lack of penetration (LOP) is a discontinuity in which the weld metal fails to extend through the full thickness of the joint. The arc melted the sidewalls and the cap, but the root faces never fused. A thin, unbonded land of base metal remains at the bottom of the joint, running along the length of the weld like a crack waiting for a load. That unbonded plane is exactly what inspectors are hunting. It hides until something pulls the joint apart.

The Four Names for One Problem

Walk into any fabrication shop and you will hear the same flaw called by several names. They are close cousins, but they are not perfectly interchangeable.

Lack of penetration — the umbrella term for weld metal that did not reach the intended depth of the joint.
Incomplete penetration (IP) — the broad synonym, common in older code text.
Incomplete root penetration — the flaw specifically at the root, where the two root faces stay unbonded.
Incomplete joint penetration (IJP) — the term used for groove welds, describing penetration less than the full joint thickness.

The words matter because a drawing, a WPS, and an inspection report must mean the same thing. Standard terminology comes from AWS A3.0, AWS D1.1, and the ISO family. Some pipeline and structural codes prefer the phrase inadequate penetration, and AWS A3.0 sorts out the formal definitions so two parties are not arguing about vocabulary instead of the weld. In ISO 6520-1, which classifies imperfections by number, lack of penetration sits under reference 402. Reference 4021 then narrows the idea to something very specific: a single-side butt weld welded from one side only, where the welder simply cannot reach the back of the joint.

Quick Note: The preferred word for any weld discontinuity is imperfection, not defect. A discontinuity only becomes a rejectable defect when it exceeds the acceptance criteria of the applicable code. Same flaw, different verdict, depending on the standard you are welding to.

Why the Same Root Image Gets Judged Differently

Here is where terminology turns into real money on the shop floor. The acceptability of incomplete penetration depends entirely on the joint design.

In a complete-joint-penetration (CJP) groove weld, the whole point is fusion through the full thickness. Any incomplete penetration is a rejectable defect, because the design assumed the full section would carry the stress. In a partial-joint-penetration (PJP) weld, limited penetration is intentional and written into the design. The joint is sized for the throat the engineer specified, so a shallow root is not a flaw at all.

This is why the same dark line can pass on one drawing and fail on another. An inspector is not judging the image in isolation; they are judging it against the joint’s design intent. Weld it as a PJP and a shallow root is exactly what the engineer ordered. Weld it as a CJP and that same root is a serious, rejectable flaw.

Under ISO 5817, lack of penetration is graded in the same strict planar-imperfection rows as lack of fusion — the code treats both as sharp, crack-like features rather than rounded pores. Acceptance levels simply move the goalposts; they never reclassify a planar flaw as a rounded one. That is a deliberate choice: it reflects how seriously a planar, unbonded plane affects fatigue and fracture behavior.

If you are untangling defect names, it helps to compare notes with Overlapping Welding: Technique vs. Defect Explained, where intent again decides whether a feature passes or fails.

How it looks. On film, LOP reads as a continuous dark line running along the weld root. Its density is consistent, its edges are straight, and it sits centered on the joint. Experienced interpreters call it one of the most recognizable indications in radiography. In a macro section, an etched cross-section shows the same truth: a sharp, straight, unbonded land of base metal at the root, with no fusion line bridging it. The TWI job knowledge page covers these incomplete root conditions in more depth.

Get the terms right early and the rest of the repair and inspection conversation gets a lot shorter.

The Root Causes of Incomplete Penetration

Incomplete penetration rarely has one cause. It’s usually two or three small problems stacking up: a joint that was never fit to fuse, welded with parameters that couldn’t reach the root, tracked by an arc that drifted off the centerline. Once you understand the mechanism behind each cause, you know which one to fix.

Fit-Up Errors: The Cause You Can’t Out-Weld

A closed root gap or an oversized root face gives the arc no path to the bottom of the joint. The arc melts the top edge of the land while the root stays cold and unfused. This is the classic geometry problem: no amount of current can reliably fix a fit-up error. Turning up the amps to force burn-through just widens the top of the groove, undermines the sidewalls, and risks blowing a hole.

Welders often try to compensate by slowing down and holding the arc on the joint, but the arc energy is spent melting the land and the top edges — there’s no molten pool at the bottom to fuse with. It’s a geometry limit, not a power limit. In practice, a common shop figure is a root face of about 1–2 mm, with root gaps of roughly 2–4 mm for SMAW and TIG and 1–3 mm for MAG. An included angle that’s too narrow is the same disease in another form: the electrode or arc can’t physically reach the root, so the groove has to be opened up before welding starts.

Linear misalignment, or high-low, leaves one root face recessed so the arc skips it entirely — the raised side fuses, the low side doesn’t. That’s why AWS D1.1 and AWS B1.10 treat alignment as a controlled variable; for pipeline work, API 1104 limits root alignment to 1/8 in. (3 mm). Exceed it and you’ve guaranteed a cold root on one side.

Heat Input, Travel Speed, and Electrode Choice

Too little current gives a narrow, convex bead and a wedge-shaped penetration profile that tapers out before the root. Too much current pushes the opposite way: the welder can travel fast enough to bridge the root without ever penetrating it, laying metal across the top of the gap. Travel speed too high does the same thing — the molten pool leads the arc, solidifies over the root, and traps a void underneath.

An electrode or wire that’s oversized for the groove won’t fit down into the root, so the arc sits shallow. Wrong polarity shifts the heat balance away from the workpiece. In consumable-electrode processes the filler is deposited automatically as the electrode melts, so the welder has limited independent control of penetration — the parameters do the work for you, for better or worse. For the fundamentals, our guide to 3/16 E7018 amperage covers current settings, polarity, and technique together.

Arc Behavior and Contamination

Even with perfect fit-up and correct current, the arc has to point where you want the fusion. An off-centerline track or a bad electrode angle sends arc force into one wall, leaving the opposite root cold. Magnetic arc blow deflects the arc away from the joint, especially with DC, so it wanders and skips the deepest point.

Contamination — rust, oil, paint, mill scale, moisture — boils off under the arc, disturbs it, and leaves porosity that blocks sound fusion. On TIG in particular, the wrong heat balance between cleaning and penetration can starve the root; that’s covered in our breakdown of TIG welding polarity.

Quick Reference: The Main Causes

Here’s the full cause matrix at a glance:

Root gap too small or closed — no path for the arc to reach the bottom.
Root face/land too thick — too much metal for the arc to melt through.
Included angle too narrow — electrode can’t reach the root.
Current too low — wedge-shaped profile that dies before the root.
Current too high plus fast travel — metal bridges the gap before it melts.
Oversized electrode — arc sits shallow in the groove.
Wrong angle or off-centerline tracking — fusion on one wall only.
High-low misalignment — recessed root face skipped by the arc.
Magnetic arc blow — arc deflected away from the root.
Contamination — rust, oil, moisture, mill scale disturb the arc.

Quick Note: Fit-up sets the ceiling on what your parameters can achieve. If the root gap is closed, more amps buy you a bigger bead on top, not a deeper root.

Why Lack of Penetration Matters: Structural and Safety Consequences

A lack of penetration defect rarely announces itself. The cap looks fine, the toes blend, and the part passes a visual check. Under the surface, though, the root never fused into the joint, and that single change flips the weld from a load-bearing member into a stress concentrator waiting for the right load cycle.

The Effective Throat Shrinks

A full-penetration groove weld is engineered on the assumption that the throat carries the load across its entire depth. When the root doesn’t fuse, the effective throat is reduced, sometimes dramatically. A joint designed for a given tensile or shear capacity now has less material doing the work, so the actual stress on the remaining metal climbs above what the designer intended. On a fillet weld or a partial-pen groove, that shortfall can be the difference between passing and failing a calculated margin.

Planar Defects Act Like Cracks

Here is the part that separates LOP from harmless porosity. Because LOP is planar — a flat, sharp-edged discontinuity rather than a rounded void — it concentrates stress at the sharp tip of the unfused root much like a crack does. Porosity is rounded and blunts stress; a planar flaw does the opposite. It focuses the load into a tiny region, and that focused stress is exactly what drives crack growth.

The American Welding Society’s AWS D1.1 Structural Welding Code and related filler metal and qualification standards treat this defect family with real gravity. AWS A3.0 lists incomplete joint penetration alongside incomplete fusion as a serious planar discontinuity, and the TWI notes that an incomplete root fusion or penetration reduces the cross-section that actually carries stress. Planar flaws are the ones codes are written to reject.

Fatigue Is Where It Bites

Under static loading, a modest LOP may simply reduce a safety margin. Under cyclic loading, the story changes. The sharp root tip is a perfect fatigue initiation site, and cracks start earlier and grow faster than they would from a sound root. Bridges, machine frames, trailers, pressure vessels, and anything that flexes over thousands of cycles are especially unforgiving. A weld that survives one proof load can still crack after a few thousand service cycles.

Leak Paths and Pressure Service

In pressure and pipeline service, an unfused root provides a potential leak path straight through the joint wall. It is also an initiation site for fatigue cracking that can open that path under pressure cycling. For piping, tube fittings, and vessels, this is doubly serious: you get a structural weakness and a through-wall flaw in the same location. That is why pressure codes demand full fusion and why these joints are hydro-tested and radiographed rather than trusted to a visual pass. The relevant OSHA welding standards frame the broader environment these tasks sit inside.

Quick Note: LOP is not a cosmetic defect. Because it is planar, sharp, and often buried, it behaves far more like a crack than like porosity — which is why structural and pressure codes reject it rather than accept it on a size allowance.

The Business Cost of Late Detection

The economics are brutal. Repair of a completed weld is far more expensive than prevention at fit-up because gouging and re-welding consume time and can introduce distortion, which then demands rework of its own. If the defect is caught in the shop, you pay for a weld repair. If it is caught by a third-party radiograph, you may pay for re-inspection of the whole lot. If it reaches service, you can pay for a shutdown, a failure investigation, and reputational damage.

Shop detection: cost of gouge, re-weld, and re-inspect.
Late detection: scrap, schedule slippage, and re-testing entire batches.
Field failure: downtime, liability, and potential injury.

A shop laid out for cleaner fit-up, good lighting, and easy access for inspection catches more of these before they become expensive. That is where a well-planned layout for welding facilities quietly pays for itself in defect prevention. Weld it right the first time and the cost stays at fit-up, not at teardown.

How to Detect Lack of Penetration: Inspection Methods Compared

Lack of penetration hides in plain sight, sitting at the bottom of the joint beneath a sound-looking cap. Catching it demands the right method at the right moment. Some techniques confirm the root only when you can reach it; others see straight through the section. That distinction decides whether a defect is caught in the shop or found in service.

Visual Testing: Useful, But Limited

Visual testing to ISO 17637 is the first line of defence, and on open-root work it is revealing. Where the back of a pipe or plate is accessible, an inspector can examine the root bead directly for a clean, evenly fused connection through the full wall. A borescope extends that reach into confined spaces.

But visual inspection has a hard ceiling. It confirms the root only where the back is reachable, and it cannot show how deep the weld metal travelled inside the section. On a double-sided weld, a backing bar joint, or any closed configuration, the root is invisible. Sharpening your eye still matters — see Good and Bad MIG Welds: How to Tell the Difference for the visual cues.

Quick Note: A good-looking cap tells you nothing about penetration. The most convincing bead can hide a continuous unfused root underneath it.

Surface Methods: PT and MT

Liquid penetrant testing (PT) and magnetic particle testing (MT) excel at surface-breaking discontinuities, but lack of penetration is a volumetric, planar defect buried at the root. Neither reaches it unless the flaw breaks the surface. Treat them as tools for surface cracks and porosity, not internal root fusion.

Radiographic Testing: The Classic Root View

Radiographic testing to ISO 17636 is the traditional workhorse. Because X-rays pass through the section and expose a film or detector, an unfused root shows up as a continuous dark line along the joint’s bottom — a clear, documentable signature. RT also gives a permanent record, dominating code work on pressure piping and structural joints.

The catch is orientation. RT is strongest when a flaw has volume or an opening aligned with the beam. A tight, crack-like planar root flaw can be nearly invisible on film, since the beam passes through the thin gap with little density change.

Ultrasonic Testing: Better for Tight Planar Flaws

This is where ultrasonic testing pulls ahead. UT sends a sound beam into the weld and listens for echoes. A flat, planar reflector like lack of penetration returns a strong signal when the beam strikes near perpendicular. That sensitivity is why UT is preferred over RT for these flaws — it does not depend on the defect having volume or a surface opening.

UT also works from one side and needs no radiation safety zone. It demands a skilled operator, and it produces a signal rather than a picture, so interpretation carries more judgement.

In-Process Monitoring: A Screening Signal, Not a Verdict

Infrared thermography and high-speed weld-pool imaging watch the root pass as it happens. A cold, narrow pool or odd thermal signature can flag that penetration is falling short. These tools catch a drifting process early, stopping a run of bad welds. But they are screening signals only, never the acceptance result. No camera replaces code-required inspection.

Destructive Testing: The Proof of Root Soundness

For procedure qualification and test coupons, destructive methods give ground truth radiography alone cannot:

Macro section: cut, polish, and etch across the weld to reveal how deep fusion reached.
Nick-break: a notch is cut and the specimen broken open to expose the root.
Root bend: bends the root into tension, opening any unfused plane wide.

A macro, nick-break, or root bend on a coupon is definitive proof of root soundness. If RT passes a joint but the macro shows a cold lap at the root, RT was fooled; the destructive test was not.

Choosing the Right Combination

No single method covers everything. A sensible route pairs visual inspection with UT or RT for internal confirmation, then validates the procedure with destructive testing. Compare:

Method	Finds tight root LOP?	Best Use
Visual (ISO 17637)	Only if back is accessible	Open-root surface check
PT / MT	No (surface only)	Surface cracks, porosity
RT (ISO 17636)	Often, as dark line	Documented volumetric records
UT	Yes, strong on planar flaws	One-sided, tight root detection
Thermal / pool imaging	Screening only	Real-time process control
Macro / nick-break / bend	Yes, definitive	Procedure qualification

Match the method to the geometry, the code, and the flaw you fear most. When tight root penetration is the concern, that usually means UT backed by destructive proof on coupons.

Prevention: Fit-Up, Parameters, and Welder Technique That Work

Lack of penetration is rarely a mystery. In most cases the root never had a fair chance: the gap was wrong, the current was too low, or the arc never reached the bottom of the joint. Fix those three things and you remove the defect before it happens. Prevention is cheaper than repair, and it starts long before you strike an arc.

Fit-Up: Control the Gap, Then Re-Check It

Everything begins with the joint geometry in your work procedure. Match the root face and root gap to the WPS before you tack, and verify them with a gauge rather than by eye. A root face that is too thick leaves metal the arc cannot melt through; a gap that is too tight does the same thing.

Tacks are the hidden culprit. As each tack cools it shrinks and pulls the plates together, quietly closing the gap you so carefully set. Re-measure the root gap after every tack, or set it slightly wider from the start to compensate for the pull. On long seams, tack in a sequence that works back toward the start so shrinkage does not drag the gap shut ahead of you.

Use spacers, wedges, and strongbacks to hold the gap against shrinkage. They bridge the joint, hold dimension, and keep the root opening stable until the weld is complete. For pipe and tube work, the same discipline applies as with weld tube fittings, where a few thousandths of an inch decide whether you get fusion or a dark line on the film.

Parameters: Current and Travel Speed for the Real Thickness

Set amperage for the actual material thickness and welding position, not for the last job you ran. Too little current cannot melt both root faces; too much burns through and forces you to rush. Check that your machine settings match the WPS and the electrode or wire diameter in hand.

Travel speed matters just as much as current. Move too fast and the arc rides on top of the joint, freezing before the root faces fuse. The practical rule: slow down until the weld pool is roughly twice as wide as the root gap. That width tells you the arc is melting down into the root, not just laying metal across the surface.

Keep the arc length short and steady. A long arc loses focus, wanders off the centerline, and spreads heat where you do not need it. For a comparison of how polarity and arc control interact, see TIG welding polarity, but the principle holds for every process: a tight arc puts energy into the root.

Quick Note: A properly penetrating open root shows a small keyhole at the front of the pool, with both root faces melting into it. If that keyhole closes, penetration is being lost. Watch the keyhole, not just the bead surface.

Technique: Reach the Root and Hold a Keyhole

On open-root joints, the keyhole is your best instrument. Hold it open and let both root faces feed into the pool, moving forward just fast enough to keep it from closing. On MMA root runs, a 2.5 mm or 3.25 mm electrode is common because it can physically reach the root and still let you control the weld pool. Match electrode size to the joint; one that is too large cannot get into the gap.

Grind the ends of your tacks to a taper. A tapered tack lets the root pass tie in smoothly instead of stopping dead against a cold, raised lump. That kills the unfused stop/start that so often hides a lack of penetration.

Preheat, Cleanliness, and Back-Gouging

Apply the preheat your WPS calls for, especially on thick or high-carbon steels, so the root wets and fuses rather than chilling out. Clean the joint faces of mill scale, rust, oil, and moisture before you weld; contamination steals heat and blocks fusion. When a sealing run follows a root pass, back-gouge to sound metal first. Grinding or gouging back until you see bright, solid metal removes the imperfect root and gives the sealing pass something real to fuse to. If you want to build these habits into a routine, our guide to welding practice walks through the drills that make them automatic.

Do these things and lack of penetration stops being luck. It becomes a variable you control.

Process-by-Process: SMAW, GMAW/FCAW, GTAW, and SAW

Each arc process reaches the root in its own way, and each has its own way of failing to. Here’s how they behave at the root, and where they tend to go wrong. The consumable-electrode processes — SMAW, GMAW, FCAW, and SAW — deposit metal as the electrode melts, so penetration and deposition are tied together, making them more prone to lack of root penetration. GTAW works differently, and that difference is the reason it produces cleaner roots.

SMAW (Stick): Electrode Size and Current Control the Root

Stick welding puts the root decision in your rod choice and your amps. For an open root, a small-diameter electrode — 2.5 mm (3/32 in) or 3.2 mm (1/8 in) — run at fairly high current gives the concentrated heat needed to melt the land through. Drop the amperage to be gentle and you get the opposite: a rod that sits on top of the joint instead of keyholing into it.

E6010 and E6011 are the classic open-root rods because their fast-freeze slag lets you control the keyhole. A slight back-and-forth or whipping motion keeps the root edges melting just ahead of the puddle. Oversize rods at low current, or a travel speed that outruns the arc, are two of the fastest routes to LOP.

On fill passes, a low-hydrogen rod still needs the right current window — the wrong setting on something like 3/16 E7018 amperage hurts fusion just as badly.

GMAW/FCAW (MIG/MAG and Flux-Cored): Arc Length and Transfer Mode

MIG/MAG and flux-cored push molten wire across the arc, so the filler can bridge a gap before the root ever melts. Arc length (set by voltage) and transfer mode therefore decide penetration. A short arc length, matched to the material thickness, drives heat into the root; a long arc spreads it across the surface and leaves the bottom untouched.

Short-circuit transfer runs cool and is the most LOP-prone mode on thick plate; spray or pulsed-spray transfer is often required on heavier sections. Contact tip condition matters more than most welders admit — a worn tip lets the wire wander and destabilises the arc, and that shows up as inconsistent penetration.

Flux-cored adds polarity to the equation. Gas-shielded FCAW usually runs DCEP for deeper penetration, while many self-shielded wires run DCEN, which is less penetrating. Getting polarity wrong flips the heat balance and can quietly kill root fusion. Self-shielded wire trades penetration for portability, which is exactly why it needs careful technique on open roots.

GTAW (TIG): Independent Filler Gives Real Root Control

TIG is the least LOP-prone arc process for one simple reason: the welder adds filler independently of the arc. With no wire feeding itself into the puddle, you can hold the arc on the root and let it melt through before any metal is added. On an open root that means running a keyhole — a small hole that forms at the leading edge and closes behind the puddle as you travel. Current is critical: too little and the arc stalls on the surface, too much and you blow through. The right TIG polarity sends the bulk of the heat into the workpiece.

SAW: Current, Flux, and Alignment for Deep Penetration

Submerged arc buries the arc under a blanket of granular flux, and deep penetration depends on high current plus correct flux and electrode alignment. High amperage drives the weld deep into the joint, the flux shapes the arc, and the electrode must sit on the joint centreline — even a small offset pushes penetration sideways. Robotic SAW earns its keep here: automation removes operator variability, but it also repeats a wrong parameter set on every pass. If the current or alignment is off, you don’t get one bad weld — you get hundreds.

Quick Note: If you switch a self-shielded FCAW wire from DCEN to DCEP hoping for more penetration, retune voltage and travel speed too. Changing polarity alone often just overheats the wire.

Read together, the pattern is clear: anything that feeds metal into the arc demands tighter control of current, arc length, and alignment to reach the root. GTAW buys that control back, which is why it remains the reference process for critical open roots.

Lack of Penetration vs. Lack of Fusion, Overlap, and High-Low

On the shop floor these four names get swapped constantly, and that confusion costs real money. Each defect has a different cause, a different inspection response, and a different repair. Get the label right and the fix usually follows.

Lack of Penetration vs. Lack of Fusion

The cleanest split: lack of penetration means the arc never reached the intended depth of the joint. The weld stops short and leaves unfused land at the root. Lack of fusion means the arc did reach that interface, but the molten metal never bonded to it.

In short, penetration is about where you got to; fusion is about whether you stuck once you got there. The TWI guidance groups incomplete root fusion and penetration together because both are planar defects, which is why they behave so similarly under load.

Defect	What it is	Where	On a radiograph
Incomplete penetration	Failure to reach intended depth	Root of groove	Straight dark line along root
Root lack of fusion	No bond at the root face	Root interface	Straight dark line along root
Sidewall lack of fusion	No bond at a bevel the arc reached	Along sidewalls	Often too thin to register
Overlap / cold lap	Metal rolls over base without fusing	Toe, surface	Dark edge at the toe
High-low	Misaligned joint faces	Root, across joint	Offset image, density shift

Here is the trap. On a radiograph, both incomplete penetration and root lack of fusion appear as a straight dark line along the root, so the film alone may not tell you which one you have. Repair quality depends on root-cause analysis, not just the image.

Sidewall lack of fusion is worse. When the planar gap is too thin to register on film, radiography misses it, while ultrasonic testing detects it reliably. Codes that demand volumetric inspection of planar defects lean on UT for exactly this reason. That is also why overlap defect detection and fusion-defect detection are treated as separate problems.

Macro examination settles the argument. Cut, polish, and etch a cross-section and the difference is obvious: incomplete penetration shows a gap at the bottom of the joint, while lack of fusion shows a crack-like line between passes or along a bevel where metal sits beside metal without blending. The AWS D1.1 structural code treats these as distinct discontinuity types for acceptance, so mislabeling one as the other can send a repair in the wrong direction.

Overlap, Cold Lap, and High-Low

Overlap is a surface imperfection: weld metal rolls over the base metal at the toe without fusing. Cold lap is its GMAW/FCAW cousin, where the puddle freezes ahead of the arc and lays metal down with no bond. Unlike penetration, both live at the surface and are usually visible to a trained eye. For the technique side of that defect family, see overlapping welding: technique vs. defect.

High-low is different again. It is a misalignment of the joint faces, not a bonding failure at all. It thins the effective root section and makes penetration harder to reach, which is why fit-up belongs in the same conversation. Solid welding practice starts with matching the two faces before the arc ever strikes.

Fixing one as if it were the other is how rework multiplies. Naming the defect correctly is half the repair.

Quick Note: If a radiograph shows a dark line along the root, do not assume lack of penetration. Confirm whether the arc reached the interface (lack of fusion) or never got there (penetration) before you choose a repair.

Repair Methods and Common Mistakes to Avoid

When lack of penetration (LOP) turns up on a radiograph or a penetrant check, it will not fix itself, and no cosmetic cap will hide it. The defect has to be cut back to sound metal and rebuilt. A repair done well is permanent; a repair done badly is a second failure waiting for the same load. Which method you use comes down to one question: do you have access to the back side of the joint?

Repair With Back Access

The standard fix is unglamorous. Gouge or grind the defect out until you hit clean, fully fused metal, then re-weld it, normally following the same qualified procedure that produced the original weld. The removal step deserves as much care as the welding. Arc-air gouging leaves a thin carburized layer on the groove faces, and that hard skin will drive porosity and poor fusion if you weld straight over it. Grind the carburized layer off before you strike an arc. Once the groove is clean, a sealing run from the back closes the root and restores through-thickness penetration; many shops follow the root with a second pass to be certain of the seal. The re-weld should follow the same qualified procedure, including preheat, interpass temperature, and consumable choice, because a repair is no place to improvise. On deep defects, remove the metal in stages and inspect between them rather than gouging blindly.

Quick Note: Grind, don’t just gouge. The carbon layer left by arc-air is only a few thousandths thick, but it is enough to ruin the repair if it stays in the joint.

Repair Without Back Access

If the joint is closed, boxed in, or welded from one side only, you cannot seal it from behind. Turning up the heat and running a pass from the face rarely helps, because a hotter arc simply cannot reach a root that sits buried under the fill. In this case the joint may have to be cut out, re-bevelled, and welded again on a procedure that guarantees penetration from the side you can actually reach. It is slow and costly, but it is the only repair that truly gets to the defect.

The Mistakes That Make It Worse

Cranking up the amperage. More current often widens the pool and causes undercut long before it reaches the root. If the root is closed, heat is the wrong lever; technique and fit-up are.
Welding over an unfused root. Capping a defect just traps it under good-looking metal. The unfused root remains a crack initiation site long after the face looks clean.
Running an oversized root electrode. A fat rod in a narrow gap bridges the opening instead of fusing it. Match the rod, and the wire in a wire welder, to the root gap.
Ignoring high-low. A mismatched pipe offset means one side of the groove is short and one is deep; the shallow side gets no penetration while you chase the deep side. Correct the fit-up, not the travel speed.
Skipping fit-up verification. Root gap, land, and alignment decide penetration before the first arc. A root gap that drifts by a millimetre changes penetration more than any dial on the machine. Check it, every joint.
Re-examining only part of the repair. Repaired welds should be checked visually and then radiographed in full under the repair procedure, not just at the spot you fixed.

Repairing LOP is a discipline problem as much as a welding one. Clean the groove completely, control the variables that caused it, and inspect the whole repair. Fine-tune your heat control with our guide to AC or DC for TIG welding steel, and treat every root as if the radiograph is already coming. Log what you removed, how you re-welded it, and the final inspection result so the repair stays traceable.

Your Action Plan: Weld Roots Right the First Time

Every technique in this guide points to one conclusion: the cheapest weld is the one you never have to repair. Prevention costs minutes of measuring and setup. Detection and repair cost cut-out time, rework, inspection fees, and sometimes a failure in service. Workmanship standards such as AWS D1.1 treat incomplete joint penetration as a rejectable discontinuity, not a cosmetic quibble. The welders who rarely see this defect on a radiograph are not lucky. They are methodical.

The Prioritized Action Plan

Confirm the WPS and joint design first. Before you touch the machine, read the procedure. Know the required root face, root gap, bevel angle, and process. If the drawing and the WPS disagree, stop until it is resolved.
Measure before the arc starts. Treat fit-up as a non-negotiable first step. Check root face thickness and root gap against the WPS with a gauge, not an eyeball. A tight gap or a heavy face is a lack-of-penetration defect waiting to happen.
Re-check gaps after tacking. Tacks pull a joint. A gap that measured perfect can close up or open wide once heat and shrinkage take hold.
Qualify process and polarity for the actual joint. Never recycle settings from a different job. Polarity shifts heat balance and penetration profile, so confirm it suits this thickness and design.
Control travel speed and arc length. Slow, steady travel with a short arc drives the root in. Racing the puddle or long-arcing leaves the root behind.
Watch the keyhole. A small, consistent keyhole tells you the arc is burning through. Lose it and you have likely lost penetration.
Inspect early and often. Check the root before you bury it under fill passes. Catching a defect early saves an entire weld.

Quick Note: A radiograph that shows a clean, straight dark line along the root is not a mystery. It is the signature of a joint that never fused through, and it is almost always traceable to fit-up or parameters.

The Quick Checklist

WPS and joint design verified
Root face and root gap measured pre-weld
Gap re-checked after tacking
Process and polarity qualified for the joint
Travel speed and arc length controlled
Keyhole watched and held
Root inspected before fill passes

Invest in the Welder, Not Just the Weld

All of these rules only work if your hands can execute them. Consistent training and practice are what turn prevention rules into reliable technique. If you want a structured way to build those habits, see our guide to building real welding skills. A clean, organized shop makes good practice repeatable, so our guide to planning a safe welding shop pairs well with it.

Be honest about your current skill. Shop welds fail when good welders rush good protocols. The welders who avoid lack of penetration for decades are not the fastest; they are the most consistent. Slow down on the root, log what works, and let repetition turn the checklist into muscle memory.

Frequently Asked Questions

What is lack of penetration in welding?

Lack of penetration is a root condition where weld metal does not extend through the full thickness of the joint as the design requires. AWS A3.0 calls it incomplete joint penetration. It is a planar, crack-like discontinuity that concentrates stress at the unfused root.

How do you detect lack of penetration?

Back-side visual inspection works only where the root is accessible. For closed joints, radiographic testing shows a continuous dark line along the root, while ultrasonic testing is often better for tight planar flaws. A macro section, nick-break, or root bend on a coupon proves root soundness.

Is lack of penetration always a defect?

No. In a complete-joint-penetration weld any incomplete penetration is rejectable, but in a partial-joint-penetration weld the limited penetration is intentional. Always read the welding symbol and WPS before judging a root.

What are the main causes of lack of penetration?

The most common causes are a root gap that is too small, a root face that is too thick, insufficient amperage or heat input, travel speed that is too fast, an electrode too large for the groove, a narrow included angle, off-centerline tracking, and high-low misalignment.

Does more amperage always fix lack of penetration?

No. If the root gap is closed or the root face is too thick, extra current mainly widens the pool and can cause undercut without reaching the root. Fit-up is usually the cheaper and more reliable fix.

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

Lack of fusion is a bonding failure at an interface the arc reached, such as a sidewall or interpass. Lack of penetration is a failure to reach the intended depth at the root at all. Both are planar and treated seriously, but the mechanism and repair differ.

What are the acceptance criteria for incomplete penetration?

Acceptance depends on the code. In API 1104, inadequate penetration without high-low is a defect when an individual indication exceeds 1 in. (25 mm) or the aggregate exceeds 1 in. per continuous 12 in. ISO 5817 does not permit it at quality levels B and C, with a limited allowance at level D.

Can you repair lack of penetration without cutting the weld out?

If there is back access, you can gouge or grind the defect back to sound metal and re-weld a sealing run. Without back access, the joint often has to be cut out, re-bevelled, and welded again, because a hotter pass from the face cannot reach the buried root.

Is TIG welding less prone to lack of penetration?

Generally yes. GTAW is a non-consumable process where the welder adds filler independently of the arc, so penetration can be controlled before metal is deposited. Consumable-electrode processes deposit metal as the electrode melts, giving the welder less independent control.

Can you grind out lack of penetration?

Grinding works only if it removes the entire unfused root and exposes sound metal, which usually means grinding through much of the weld. In thicker sections, gouging plus re-welding is more practical than grinding alone.