Worm Weld Defects Explained: Wormhole Porosity, Worm Tracks, Causes, Fixes & Code Limits

A structural shop in Ohio had three 300 mm girth welds come back from radiography stamped ‘reject — herringbone indications.’ Nobody on the floor could read the film, and the welder swore his parameters hadn’t moved all shift. The real culprit was a bucket of flux that had sat open overnight next to a wash bay. That is the story of the worm weld: a defect that grows quietly inside the metal, hides under slag, and only shows its true face on a film or a leak test — and it is almost always traceable to something a crew could have controlled before the arc was ever struck.

The Day the Radiograph Came Back Black: Why ‘Worm Weld’ Problems Keep Welders Up at Night

Three 300 mm girth welds, a fabrication shop in Ohio, and one bad bucket of flux. The welds went out to radiography and came back stamped reject — herringbone indications. Nobody on the floor could read the film, and the welder insisted his parameters hadn’t drifted all shift. The actual culprit turned out to be a sack of submerged-arc flux that had sat open overnight beside a wash bay, soaking up moisture all evening before anyone struck an arc.

That is the reality of what searchers mean by a worm weld: a defect that grows quietly inside the metal, hides under a clean-looking layer of slag, and only reveals itself on film, on a leak test, or worse, in service. It is almost always traceable to something a crew could have controlled before the torch was ever lit.

So What Are People Actually Searching For?

When a welder, inspector, or QA engineer types “worm weld” into a search box, they are usually chasing one of two closely related things:

• Wormholes — elongated, tunnel-shaped gas pores running through the weld metal. Unlike a round, isolated pore, a wormhole stretches lengthwise, often parallel to the weld axis. On a radiograph, they produce a distinctive herringbone appearance that TWI describes as the classic wormhole signature — a dark, feathery line that inspectors learn to spot instantly.
• Worm tracks (shop slang: wagon tracks) — the surface version of the same problem. These are long, ropey ridges or valleys on the weld face, most common in flux-cored arc welding, caused by gas that escaped the solidifying pool but got trapped under the slag instead of escaping cleanly. Our friends at The Shed Journal’s field write-up on worm tracks captures how they look in practice: parallel streaks that resemble wagon wheel ruts cut into the bead.

They share a root cause — gas that could not escape in time — but they are found in different places and judged differently by code. Telling them apart is the first step toward fixing either one.

Why This Defect Matters More Than Ordinary Porosity

Here is an uncomfortable fact for anyone who has shrugged at a few pinholes: porosity is the single most common imperfection family that welding codes allow in limited quantities. AWS D1.1, ISO 5817, API 1104 — they all permit some level of distributed porosity, because a handful of tiny round pores rarely threatens structural integrity.

Wormhole porosity is the exception. Because it is elongated and often interconnected, it behaves like a leak path through the wall rather than a scattering of isolated voids. A pipeline that passes a pressure test with round porosity can still weep through a wormhole tunnel that runs from the inner surface to the outer. In pressure-containing, cryogenic, or cyclically loaded service, that difference is not academic — it is the reason a coupon gets cut out at 2 a.m.

The gas itself rarely arrives alone. TWI’s porosity guidance notes that as little as 1% air entrainment in the shielding gas causes distributed porosity, while levels above 1.5% produce gross surface-breaking pores. In other words, a slightly loose gas fitting or a cracked hose can be the whole story behind a rejected weld.

What This Article Covers

This guide walks the full path from definition to disposition, in the order a real job hits it:

1. Precise definitions — how ISO 6520-1, AWS D1.1, and shop slang each describe wormholes and worm tracks.
2. The physics of why gas gets trapped instead of escaping.
3. A catalog of the real-world causes, from contaminated plate to damp flux.
4. Process-by-process risk in SAW, FCAW, SMAW, GMAW, and GTAW.
5. How to actually find the defect — and which NDT methods will never see it.
6. Code acceptance limits under AWS D1.1, ISO 5817, and API 1104.
7. A prevention routine you can run on the floor tomorrow.
8. Repair strategy, including when the only right answer is to cut it out.

A quick word on safety while we are here: chasing a rejected weld often means grinding, gouging, and re-entry into a hot work zone, and that is exactly where crews get careless with eye, hand, and respiratory protection. If your shop’s PPE routine is thin, our Welding PPE: The Complete Guide to Protective Gear Every Welder Needs (2026) is a solid place to tighten it up before the rework starts.

The Ohio shop’s story ended the way these stories usually do: the flux was replaced, the烘干 procedure tightened, and the next three girth welds passed without a single herringbone on the film. Nothing about the fix was exotic. That is the encouraging part about worm weld defects — they are severe, they are sneaky, but they are overwhelmingly preventable by people who know what they are looking at.

What Exactly Is a Worm Weld? Definitions, ISO/AWS Terms, and How Wormholes Differ from Worm Tracks

Before you can fix a worm weld, you need to know exactly what you’re looking at. In most shops, the term “worm weld” is less of a technical classification and more of an umbrella. It can refer to a tunnel-shaped gas cavity buried in the weld metal, the surface trace it leaves behind, or the radiographic indication that flags it for rejection. To avoid crossed wires between welders, engineers and inspectors, it helps to fall back on the language that the codes actually use.

The ISO Definition: Reference 2016, the Wormhole

The cleanest starting point is the international standard. Under ISO 6520-1, gas pores fall under series 201, with the wormhole singled out as reference 2016. It is defined as an elongated, tunnel-shaped gas cavity formed within the weld metal or at the weld face. The defining trait is elongation: a wormhole is a tube, not a dot, and its length is meaningfully greater than its width.

This fits into a broader family. Any gas cavity that is longer than it is wide is classified as elongated porosity, and the wormhole is its most pronounced, tunnel-like form. In practice, a cavity roughly three times as long as it is wide is likely to earn that label. These distinctions matter on paper, as they keep reports consistent between parties and prevent vague calls from slowing down a job.

AWS Language: Piping Porosity, Wormhole Porosity, and B1.11

The American Welding Society uses slightly different terminology, but arrives at the same geometry. In AWS D1.1 Annex J, the glossary refers to elongated porosity whose major dimension lies approximately perpendicular to the weld surface as piping porosity. It also notes that this is commonly called wormhole porosity. That perpendicular orientation is built directly into the definition.

For surface indications, AWS B1.11 takes over. When a welder spots “worm tracks” on a fresh bead, B1.11 classifies them, at their root, as elongated porosity. The ripples on the face are a symptom of a cavity underneath, not the defect itself. In pipeyards and structural shops, these axial traces are often called “wagon tracks”, a term that has no formal listing, but one that nearly every inspector will recognize on the floor.

Orientation Is the Whole Game

The most useful distinction to carry into any inspection is orientation. It separates two defects that can look similar at a glance, but point to very different causes:

• Piping porosity (wormholes): Run roughly perpendicular to the weld face. These are vertical tunnels, with gas attempting to escape straight up through the bead.
• Worm tracks (wagon tracks): Run roughly parallel to the weld axis. These are horizontal tunnels along the length of the bead, where gas gets dragged by the motion of the molten pool.

Both are elongated gas cavities, and both can be rejectable under the right code. However, perpendicular tunnels typically point to sustained gas generation, while axial tracks tend to point to travel speed, current, or shielding issues stretching the bubbles along the pool. Reading the orientation will narrow your suspect list almost immediately.

Reading the Herringbone

The radiographic signature is often the tie-breaker. Wormhole indications on RT film read as tapering traces arranged in a repeating herringbone pattern, angling up and out from the weld centerline like the bones of a fish. This separates them from round pores, which appear as isolated dark discs, and from elongated slag lines, which are irregular, blunt-ended and lack that symmetry. If your technician can describe that pattern, they can usually name the defect before the film is dry.

Finally, these geometries aren’t process-specific. From submerged arc to GTAW, the same principles apply across the processes covered in this article, including Flux Core Wire Welding Tips. Keep these definitions in your back pocket, and you’ll use the same accurate vocabulary no matter which station you’re at.

Why Wormholes Form: The Physics of Gas That Can’t Escape

Every wormhole starts the same way: gas dissolves into liquid metal, then the metal turns solid faster than the gas can float away. Understand that one sentence and you can prevent most worm weld problems before they reach a radiograph. Here is the mechanism in plain language, so you can read a defect and work backward to its cause.

The Solubility Cliff: Why Gas Boils Out of Solidifying Steel

As TWI’s technical knowledge base explains, porosity forms when nitrogen, oxygen and hydrogen are absorbed by the molten weld pool and then released as the metal solidifies. The physics is elegantly simple: gases are far more soluble in liquid steel than in solid steel. While the pool is molten it behaves like a sponge, holding a surprising amount of dissolved gas. The moment the solidification front passes through, solubility falls off a cliff.

Hydrogen is the most dramatic case. Its solubility drops by roughly an order of magnitude between liquid and solid states. That is why steel appears to “boil” hydrogen out as it cools, and why hydrogen shows up in delayed cracking discussions as often as in porosity discussions. The gas has to go somewhere. If the pool surface is open and calm, bubbles rise, pop, and leave no trace. If the metal freezes first, the gas is stranded.

Turbulence: How a Bubble Becomes a Tunnel

A single trapped bubble is a round pore. A wormhole needs more: sustained gas generation plus a directional escape route that keeps getting sealed behind it. Research presented at the 2016 World Conference on Nondestructive Testing found that tunnelling porosity is caused primarily by turbulence in the weld pool, usually from excessively high welding currents. That radiography paper is worth reading if you have ever stared at a herringbone pattern wondering what the film was really telling you.

Here is what happens step by step:

1. Above a certain current, arc force stirs the pool violently instead of letting it sit calmly.
2. Bubbles get swept around and stretched lengthwise along the direction of travel instead of rising straight up.
3. The solidification front advances as a rough, tree-like (dendritic) surface of tiny interlocking fingers.
4. A bubble caught between those dendritic arms cannot rise; buoyancy is no match for the freeze rate.
5. Each new bubble feeds the same cavity, and the advancing front seals a continuous tunnel of porosity parallel to the weld axis.

High travel speed makes it worse, because the top of the pool skins over before dissolved gas finishes escaping. That is why wormholes often cluster along one edge of a bead or near the crown: the last places to lose their liquid.

Hydrogen’s Favorite Back Doors

Hydrogen enters through ordinary doors: damp flux, humid shop air, oil or grease on the joint, paint and primer, wire drawing compounds, and electrode coatings. The arc’s heat dissociates water vapor, and the resulting atomic hydrogen diffuses readily into molten metal. On rapid cooling the metal becomes supersaturated, and the excess comes out as porosity or, in susceptible steels, as cracking. This is the physics behind the low-hydrogen discipline you already know from stick welding: an E7018 that has absorbed moisture is a hydrogen delivery system with a flux coating on it. Amperage matters too, since running a rod too hot drives more moisture and coating gases into the pool; the 5/32 E7018 Amperage guide covers those settings in detail.

FCAW wire tells the same story from another angle. Wire left in a humid shop picks up moisture through its flux coating, and the deoxidizers and hydrogen scavengers built into that coating only work if the wire is in good condition. Our Flux Core Wire Welding Tips guide walks through storage and handling. Nitrogen comes mostly from air entrainment: a shielding gas leak, a shop draft, or flow set high enough to pull air into the nozzle. For the full setup picture, see How to Set Up a MIG Welder.

Deoxidizers: The Metals That Tie Oxygen Up Before It Bites

Manganese and silicon in the parent metal, electrode or wire act as deoxidizers: they scavenge dissolved oxygen and carry it out of the pool as harmless slag (MnO and SiO₂) that floats to the surface. That is why mild steel fillers are “killed” steels with deliberate manganese and silicon content, and why MIG wires for mild steel carry higher deoxidizer levels than their stainless cousins.

Rimming steels are a different animal. They are not fully deoxidized, so they hold a high oxygen content. As the pool cools, that oxygen reacts with carbon to make carbon monoxide, and because the reaction continues while the metal freezes, it delivers exactly the sustained gas generation worms require. The traditional fix is adding aluminum to the pool, since aluminum is a far stronger deoxidizer than manganese or silicon. The Fabricator’s list of 22 porosity causes flags incompatible filler and base metal chemistry for the same underlying reason.

One Pulse or a Slow Leak?

This mental test separates diagnosis from guessing. A round pore usually comes from a single nucleation event: one bubble trapped with nowhere to go. A wormhole requires a slow leak: gas generated continuously over several seconds while the front advances and freezes the trail behind it. Think of one drop of water falling on soil versus a steady trickle carving a groove.

Keep that distinction in mind as you move through the cause catalog next. Prevention is almost always one of three moves: keep gas out of the pool, give it a clear path to the surface, or slow the freeze long enough to let it leave.

The 12 Real Causes of Worm Weld Porosity (From Contaminated Plate to Damp Flux)

Wormholes rarely have a single culprit. What makes them maddening on the shop floor is that the cause and the symptom are separated by time: the contamination enters the arc in a fraction of a second, and the tunnel only reveals itself days later on a film. The good news is that the failure modes are well catalogued. The Fabricator’s widely cited 22-cause list for porosity covers nearly every version of the problem you will meet in production, and almost every one of them is controllable before the arc is struck.

Here is the working catalog, grouped the way a supervisor would actually troubleshoot it.

Category 1: A Dirty Joint (Causes 1–3)

1. Oil, grease and cutting fluid. Hydrocarbons decompose in the arc and feed carbon and hydrogen straight into the pool. Along the weld centreline, this is the classic signature TWI describes in its submerged-arc porosity guidance: elongated wormhole porosity sitting on the weld centreline, produced by decomposition of oil, paint, grease and hydrated oxides.
2. Rust, mill scale and hydrated oxides. Scale is not inert. It carries chemically bound water, and a damp plate on a cold morning can outgas long after it “looks” dry.
3. Paint, primer and NDT residue. Leftover penetrant, magnetic particle ink and developer are organic films that burn badly. Even legitimate weldable primer can go too far: TWI notes that low-zinc weldable primer typically should not exceed about 20 µm on the edge of a T-joint vertical member before porosity becomes likely.

Category 2: Wet Consumables and Bad Gas (Causes 4–7)

4. Damp or improperly stored SAW flux. Flux is a sponge. A bag left open beside a wash bay, or a hopper that has sat idle in humid weather, will pick up moisture and hand it to the pool as hydrogen.
5. Moisture pickup in FCAW wire. Flux-cored wire is porous by design; it absorbs humidity from the air if the spool sits out. This is one reason the shop routines in our Flux Core Wire Welding Tips guide hammer on sealed containers and short open-air exposure.
6. Shielding gas faults. Leaking regulator or cylinder connections, plugged or spatter-filled nozzles, excessively high flow that sucks in surrounding air, and drafts faster than roughly 4–5 mph all destroy the shield. The Fabricator points out that even an overhead fan 25 feet away can wreck SMAW and FCAW shielding.
7. Contaminated hoses and liners. A moisture-laden air line, an oil-filled shop compressor, or a wire liner packed with drawing compound will contaminate the arc while everything on the machine display looks perfect. Correct gas and liner setup is part of any solid MIG welder setup.

Category 3: Parameters That Trap Gas (Causes 8–10)

8. Excessive current. Too much heat makes the pool deep and fluid, giving bubbles less time to reach the surface.
9. Travel speed too high. A fast pass freezes the top of the pool while the centre is still boiling, locking tunnels in place.
10. Wrong stickout or CTWD. Excessive contact-tip-to-work distance preheats the wire, destabilizes the arc and changes shielding geometry. Getting stickout right is half of proper amperage and polarity control.

Category 4: Geometry That Holds Gas (Causes 11–12)

11. Crevices under T-joint vertical members. Gas generated in the root has nowhere to vent, so it re-emerges as tunnels higher in the bead.
12. Deep, narrow grooves. A high depth-to-width ratio traps shielding gas and turbulence. Widen the joint, stagger the passes, or use a backing arrangement that lets gas escape.

ESAB’s submerged-arc guidance adds two process-specific traps worth remembering: a flux burden that is too shallow to fully blanket the arc, and running the wire dead in line with the joint, which can drive the arc out from under the flux. Work through these twelve and you will have removed the overwhelming majority of wormhole risk before anyone reaches for a grinder.

Process by Process: Where Worm Welds Hide in SAW, FCAW, SMAW, GMAW and GTAW

The same worm-like void flagged on one job can have a completely different cause on the next. A worm weld is an elongated, caterpillar-shaped pocket of gas that forms when a bubble gets trapped by a moving solidification front instead of floating clear of the pool. What changes from process to process is how the bubble is made and why it never escapes, so the troubleshooting list has to change with it.

SAW: moisture in the flux, arc on the backing bar

Submerged arc hides the pool, so worm tracks usually show up later as a string of rounded voids on a radiograph rather than anything visible on the bead. Moisture is the first suspect: granular flux left in a humid shop picks up water and releases it exactly where the metal is freezing. Keep flux sealed, re-bake it on the manufacturer’s schedule, and specify a hydrogen-controlled flux on critical work in line with American Welding Society practice. Next comes the fit-up, where mill scale, oil and rust along the joint line feed the same reaction. Excessive voltage for the plate thickness can drive the arc out of the flux and onto the backing bar, while current set too low freezes a deep, narrow bead before gas can rise. A flux bed too shallow to cover the arc does the same thing.

FCAW: shielding, voltage and drafts

In flux-cored arc welding, the tubular wire supplies both slag and gas, and gas-shielded wires lean on an external jet of shielding gas as well. Excess voltage lifts the arc up into the wire so droplets freeze before they reach the pool, leaving a trail of trapped gas; too little voltage buries the wire and drives slag forward ahead of the arc. Self-shielded wires are the most weather-sensitive of the group, because the flux alone has to hold the envelope, and a fan, an open bay door or a compressed-air nozzle is enough to break it. Wet plate, cutting fluid and a spatter-clogged nozzle finish the list.

SMAW: damp electrodes and arc blow

The coating that shields an SMAW pool is also a gas generator. Damp or badly stored low-hydrogen rods vent moisture straight into the arc, which is why rod ovens and sealed quiver storage earn their keep; a re-bake per the electrode maker’s instructions clears most cases. Magnetic arc blow is the second classic, worst at the end of a run, around a corner, or on high-current DC. The arc wanders, the pool churns, and the gas freezes where it stands. Paint, zinc primer, oil and rust on mild steel are the third source, along with current set low enough to leave the pool sluggish and slow to release what it has taken in.

GMAW: the shielding envelope

Almost every GMAW worm weld traces back to gas coverage. Flow turned up too high makes the jet turbulent and pulls ambient air into it; flow too low leaves the leading edge of the pool bare. Drafts, a leaking hose, a spent regulator or a mixture that drifted off spec typically show up as porosity within the first inch of the bead. Mechanically, inspect the contact tip-to-work distance, a liner packed with grinding dust, a grimy wire reel and a nozzle mouth caked with spatter. For the underlying mechanism, the TWI porosity explainer is worth a read.

GTAW: almost always housekeeping

GTAW gives the cleanest pool and the best view of it, so worm holes here are rare and nearly always a cleaning failure. Oil, grease, marking crayon and chlorinated cutting fluids volatilize under the arc and get caught as the bead freezes. On aluminum, the oxide film and the moisture it traps are the main gas source, so fresh wire brushing and a solvent wipe belong before every pass. On stainless, skipping the trailing purge on a root run produces exactly the same long void, and it lands in the same place every time you rerun the joint.

Name which of these five starting points you are dealing with and the bead plus the radiograph start telling a coherent story. Our guide to reading weld porosity picks up from there, with visual cues for telling worm voids apart from blowholes, lack of fusion and slag inclusions.

How to Find Worm Welds: Visual, Radiographic, Ultrasonic and In-Process Detection

A “worm weld” is the shop-floor name for elongated wormhole porosity, and it is one of the nastiest internal defects a joint can carry. A single gas pore is roughly spherical and relatively harmless; a wormhole is a channel. It can run the full thickness of the weld, act as a stress raiser, and leak under pressure. Worse, it often hides under a bead that looks perfectly acceptable from the outside. Finding it takes more than one pair of eyes.

Visual Inspection: Fast, Cheap, and Only Half the Story

Every examination starts with good lighting, a mirror, a magnifier, and knowledge of what to look for. Surface-breaking wormholes usually announce themselves as chains of tiny dark pinholes aligned in the travel direction, a “trail” or streak that follows the weld ripple, or a cluster of spatter and arc wander. On stainless and aluminum, the mouth of a channel can be barely a hairline wide.

Where a channel may break the surface, dye penetrant testing picks up openings that plain eyes miss, and magnetic particle testing on ferritic steel can flag the surface expression of something that starts underneath. Both methods stop at the surface, which is exactly where the problem with wormhole porosity begins.

Acceptance criteria are set by the governing code, not by opinion. AWS D1.1 Clause 6 judges porosity by size, length within a stated weld length, and total area; ISO 5817 does the same for steel and ISO 10042 covers aluminum, copper, and nickel alloys. None of these standards let an inspector decide a pore is “shallow” by looking at it, which is why surface methods alone can never certify a critical joint. Our guide to weld porosity causes, types, and prevention goes deeper on the underlying mechanisms.

Radiographic Testing: The Classic Method for Wormholes

Radiography is the reference technique for gas porosity because of how the defect images. A wormhole removes metal along its length, so it reads on film or a digital detector as a dark, tapered, streamlined trail. The visual cues separate it from its neighbours:

• Round porosity appears as discrete dark circles, isolated or scattered.
• Wormholes appear as smooth, cigar-shaped or needle-like dark lines, usually aligned with the weld axis and tapering at the ends.
• Cracks appear as sharp, jagged, branching lines with high contrast edges.

One geometry rule matters more than any other: radiography records a defect’s shadow in the plane of the beam. A channel running parallel to the beam projects a short, dense dot, while the same channel lying perpendicular to the beam projects its full length. Inspectors therefore take two exposures at different angles, or rotate the source 90 degrees, so the true extent is recorded. Computed tomography goes further and reconstructs the whole channel path in three dimensions, which is why CT has become the tool of choice in failure analysis.

Practice is governed by ASME Section V, Article 2 for radiography and ASTM E1742 for the test method itself, with acceptance read from the applicable construction code. Digital radiography has largely replaced film in production shops because it removes chemical processing and supports immediate re-exposure. Practical guidance on interpreting what you see sits in our article on reading weld radiographs.

Ultrasonic Testing: Full-Volume Coverage With Caveats

Conventional pulse-echo and phased-array ultrasonic testing sweep the entire weld volume from a single side. That makes UT fast on thick sections, girth welds, and field joints where both faces are not accessible, and it gives depth and length data rather than a shadow. The relevant standards are ASME Section V, Articles 4 and 5 and ISO 17640.

The UT fingerprint of wormhole porosity is not a single dramatic echo. Wormholes are smooth, rounded, gas-filled channels, so they scatter and reflect weakly compared with a planar crack. Inspectors see a train of small, scattered, low-amplitude returns at a consistent depth, plus diffuse backscatter (often called “grass”) inside the gate.

Two limitations deserve respect. First, signal amplitude is orientation-dependent: a channel aligned with the beam path can return almost nothing. Second, geometric reflectors such as weld crown, root reinforcement, and ladder-like steps left by a weave produce false calls. That is why UT indications on critical welds get confirmed by radiography or by grinding out, and why our overview of ultrasonic weld testing stresses interpretation training alongside the equipment.

In-Process Detection: Catching Wormholes Before They Form

Inspection finds wormholes after the fact. Process monitoring stops them being born, and it works by watching the signals that precede them. Practical controls in the shop:

• Gas-side checks: flow meter verification, hose and torch leak tests, dew point and humidity monitoring, and drafts or shielding loss that strip the puddle of gas.
• Joint and consumable preparation: clean liners, dry storage of filler wire and electrodes, degreasing, and a strict time window between de-oxidizing aluminum and welding over it.
• Instrumented monitoring: welding current and voltage waveforms, since short-circuit frequency and arc signature shift when gas becomes trapped in the keyhole; arc emission spectroscopy and photodiodes sensing metal vapor and plasma changes; acoustic emission from the collapsing keyhole; and high-speed imaging of the weld pool.

Machine-learning classifiers now read these traces in real time and flag cells at risk, pausing or retuning the schedule. The important caveat is that such systems detect the conditions that produce wormhole porosity, not the wormhole itself. Final acceptance still runs back through radiography or UT, supported by solid shielding gas practice.

The working combination on a critical joint is straightforward: 100% visual plus penetrant where surface breaking matters, one volumetric method on every weld the code requires, and in-process monitoring on the production line to keep the defect from being created in the first place. Which volumetric method, and how much of it, remains the engineer of record’s call. For the wider inspection workflow, see weld inspection fundamentals.

Accept or Reject? Code Limits for Worm Weld Porosity Under AWS D1.1, ISO 5817 and API 1104

This is the question that decides whether a weld gets a grinder or a green tag. The answer depends on which code governs the job, whether loading is static or cyclic, and whether you are staring at a round pore or a wormhole tunnel. Two nearly identical radiographs can be acceptable on one project and rejectable on another, with no contradiction at all.

The governing principle is simple. Codes do not judge worm welds by how alarming they look. They judge measurable geometry: diameter, cumulative length, frequency, and location. “That looks rough” is not a disposition.

AWS D1.1: Table 6.1 and the Piping Porosity Rules

Under AWS D1.1, Structural Welding Code — Steel, porosity acceptance lives in Table 6.1 for visual inspection, with radiographic limits tied to the same criteria. For statically loaded nontubular groove and fillet welds, the rule is cumulative: the sum of visible piping porosity 1 mm or larger must not exceed 10 mm in any 25 mm of weld length, and must not exceed 20 mm in any 300 mm of weld. It is total length of piping porosity, not a single-pore limit: a 12 mm wormhole tunnel fails even if it is only 1.2 mm wide.

Cyclically loaded work tightens dramatically. CJP groove welds in butt joints transverse to computed tensile stress must have no visible piping porosity whatsoever under cyclic loading criteria. Zero. One wormhole tunnel is a reject, regardless of length.

For fillet welds in cyclically loaded nontubular connections, the frequency of piping porosity must not exceed one per 100 mm of weld length, with a maximum diameter of 2.5 mm. Both conditions must pass — frequency and size together. A single 3 mm pore spaced 150 mm apart still fails on diameter.

ISO 5817: Quality Levels B, C and D

Under ISO 5817, acceptance is expressed as quality levels, and level B is the stringent one. For a single pore in a butt weld at level B, the diameter is limited to d ≤ 0.2s, where s is the butt weld thickness, with an absolute maximum of 3 mm. For linear porosity at level B, the height is limited to h ≤ 0.2s with a 2 mm maximum, and outer pore spacing l ≤ 25 mm. Levels C and D progressively relax those numbers, so the quality level has to be named before any worm weld argument starts.

API 1104: Pipeline Rules

Pipelines play by different arithmetic. Under API 1104, an individual pore larger than 3 mm, or larger than 25% of the thinner joined wall thickness, is a defect. Cluster porosity in the finish pass exceeding 13 mm in diameter is also a defect. On thin-wall line pipe that 25% rule bites hard: a 0.7 mm pore in a 2.8 mm wall is a repair, not a judgement call.

The Worm Track Verdict Most Shops Get Wrong

A worm track without associated porosity is generally not a defect under AWS D1.1 unless the WPS specifically precludes it. Worm tracks are surface irregularities on the weld face, not internal voids that reduce section or create leak paths. Many shop practices that automatically fail a test coupon for worm tracks are simply stricter than the code — a legitimate commercial choice, but one that should be documented in the WPS or the contract, never applied as though it were code. If the governing document is silent, the coupon passes.

The full workflow is covered in our guide to pipe welding procedures, WPS, positions and inspection.

Side-by-Side: Which Limit Applies Where

How an Inspector Should Read This

1. Identify the governing code and edition from the drawing or specification.
2. Identify the loading condition — static or cyclic — because that changes everything.
3. Determine whether you are judging visible surface porosity or an internal indication on film.
4. Measure and document: length, diameter, count, spacing and location along the weld.
5. Compare against the correct clause, then cite it in the disposition.

Worm welds that fail on internal porosity are genuine structural concerns. Surface worm tracks usually are not — unless the WPS says otherwise. Knowing the difference separates a defensible rejection from a costly argument. The same discipline applies to other imperfection families, as we outline in welding defects: undercut causes and acceptance limits.

Prevention Playbook: The 15-Step Routine That Eliminates Wormholes Before They Start

Worm weld defects are almost entirely preventable. Wormholes don’t show up because the arc is moody or because a welder has bad luck. They form because gas dissolved into the molten pool faster than it could escape, and that traces back to a controllable input: a dirty surface, damp flux, a leaking hose, or a parameter drifting out of range. As The Fabricator’s catalog of 22 porosity causes makes clear, the same handful of culprits repeat across every process.

What follows is a three-phase routine: what happens before the arc strikes, how consumables and gas are managed, and what you control while welding.

Phase 1 — Before Welding: Surfaces, Joints, and Shielding

Steps 1–2: Clean the base metal by material type. Not all steels want the same cleaning. TWI’s per-material table is the simplest standard to memorize:

The scrape on aluminum matters because the oxide film regenerates fast and carries moisture with it. For joint prep on light material, see TWI’s porosity defect summary.

Step 3: Keep NDT residue and primer off the joint. Penetrant dye, developer, marking paint, and shop dust all burn into the pool. Where weldable primer is unavoidable, TWI reports a typical maximum of about 20 µm on T-joint edges; anything thicker gets ground off entirely.

Step 4: Design out gas traps. The vertical member of a T-joint hides a crevice right where the arc needs clean shielding, and deep, narrow grooves or backed-up root gaps do the same. Where possible, add vent holes, open the bevel, or back-gouge so trapped gas has somewhere to go.

Steps 5–7: Prove your gas system is leak-free. Run a daily leak check: open the cylinder handwheel, let the system pressurize for 15–20 seconds, shut the valve, and watch the regulator gauge. If pressure drops within a minute or two, find the leak before production starts. Then confirm the flowmeter reading, inspect hoses for cracks and old fittings, and clean or replace the contact tip and nozzle — a plugged nozzle with dried anti-spatter buildup wrecks shielding geometry. If you’re still learning the setup, How to Set Up a MIG Welder covers the full gas-side routine.

Phase 2 — Consumables: Flux, Wire, and Electrodes

Steps 8–10: Treat consumables as moisture-sensitive. Damp SAW flux is one of the industry’s most reliable wormhole generators. Dry it per the flux manufacturer’s recommendations and store it in a heated hopper, not an open bag on the shop floor. TWI’s FAQ on SAW porosity flags the diagnostic signature: gas flats and surface pocking on SAW beads link directly to moisture, low deoxidizer content in the flux, and an overload of fines that restricts gas escape through the slag.

For FCAW, keep wire drums sealed and re-bake or discard wire that has absorbed moisture. For stick work, follow the drying and storage schedule in AWS D1.1 Table 5.8, which sets rebake times and temperatures by electrode classification. An over-baked low-hydrogen electrode is worse than useless — the coating burns off and the arc becomes uncontrollable, a mistake covered in 3/32 E7018 Amperage: Settings, Polarity & How to Run It.

Phase 3 — During Welding: Parameters, Travel, and Draft

Steps 11–15: Hold the process steady. Wormholes grow from sustained gas generation, so the controls that matter are the ones keeping conditions constant:

• Current discipline. Excessive current boils the pool and drives gas inward. Stick to the WPS range, not “what worked yesterday.”
• Travel speed. Too fast traps gas under a rapidly freezing surface. A slow, even travel speed gives bubbles time to reach the top.
• Stickout / CTWD. Long stickout preheats flux-cored wire and decomposes its gases before the arc does. Keep CTWD short and consistent.
• Draft control. Shielding gas blows away in cross-drafts above roughly 4–5 mph. Screen the bay, not the welder.
• Run-off tabs. Start and stop wormholes and crater pipes on tabs you cut off, not in the pipe or vessel wall.

For electrode-specific settings that keep pools calm, see 5/32 E7018 Amperage: Settings, Polarity & Technique and Flux Core Wire Welding Tips. And since prevention only works if the operator stays alert through a full shift, keep your gear current per Welding PPE: The Complete Guide.

The Shop-Ready Daily Checklist

1. Surfaces ground, brushed, and degreased per material type
2. Primer within 20 µm or removed; no NDT residue near the joint
3. Flux dry, from the heated hopper, with no fine-packed dead bags
4. Electrodes from a sealed oven or rod oven, per D1.1 Table 5.8
5. Gas leak check passed (15–20 second pressurization test)
6. Nozzle clean, contact tip correct size, CTWD within range
7. Draft screens up, WPS on the stand, run-off tabs tacked

Repair, Rework, and When to Cut It Out: Fixing Worm Welds Without Making Things Worse

Finding the defect is the easy part. Deciding what comes next is where shops get into trouble, because a wormhole is not a round pore you can peck out with a pick and forget. It is a tunnel, and if you treat it like a dot, you will re-weld over a hidden void, pass the visual, and fail the next radiograph three feet down the line. Here is how professionals actually remove worm weld defects, confirm the metal is clean, and document the repair so it holds up to both the code and the next inspection.

Standard Remedial Action: Gouge, Grind, Reweld

There are two legitimate outcomes, and the choice depends on how much of the weld is affected. For localized indications, TWI’s remedial guidance is straightforward: remove the affected metal by localized gouging or grinding, then re-weld the area. When porosity is widespread or clustered through a long length of the bead, localized digging is a waste of time. The entire weld must be removed, the joint re-prepared, and the joint re-welded to the agreed procedure.

Wire-feed welding education sources list the practical toolkit for this work: grinding, air gouging (carbon-arc or arc-air), and rewelding. The order matters. Every one of them only works after you have eliminated the gas source and the trapping cavity. Otherwise you are simply cutting a clean hole in a weld that will fail the same way again.

Why a Wormhole Is a Line, Not a Dot

The surface pit you can see is usually just the mouth of the tunnel. Grinding until the visible dark spot disappears routinely leaves most of the void buried behind the weld face, running parallel to the surface. Dig along the tunnel axis, not into the first hole you spot, and keep cutting until the runout is clean and the remaining wall thickness still meets the drawing.

Here the diagnosis loop from Welding Defects Undercut: Causes, Acceptance Limits, and How to Fix It applies almost unchanged: grind, then verify, then stop grinding. After every pass, clean the surface and check it with dye penetrant (PT), or magnetic particle (MT) if the part is ferromagnetic. Penetrant wicks into the residual tail of a tunnel far better than your eye does, and it takes seconds between grinding passes. Shop reporting on worm tracks repeatedly shows repairs that “looked fine” until RT was pulled again.

Air gouging deserves its own caution. It is fast, it reaches into deep grooves, and it leaves a carburized surface that must be ground back to sound metal before re-welding. Wear proper welding PPE, because gouging throws sparks and molten slag far wider than a normal bead, and the acoustic levels are high enough that hearing protection is not optional.

Repair Welds: Preheat, Procedure, and Documentation

Repair is welding, so it lives under the same rules as production welding. The repair must be performed to a qualified procedure, preheat and interpass temperature re-applied as required for the base metal and thickness, and the bead laid to the same WPS, positions, and pass sequence the original joint called for. Never wing a repair with whatever amperage feels right; a repair that runs colder than the qualified range is a hydrogen trap waiting to crack.

Two details get overlooked constantly:

• Craters and pipe overlap. Stop-start points and crater pipes sit right where wormhole tunnels begin. Rewelding over a crater without grinding the end of the bead guarantees the overlap zone carries the old defect into the new one.
• Run-on and run-off tabs. On critical butt joints, start and stop outside the joint so the repaired length never overlaps suspect metal.

Documentation closes the loop: record the repair location, method of removal, procedure used, and the re-examination that followed. Code work generally requires the repaired area to be re-inspected by the same method that found the defect, so if RT flagged it, RT clears it. For the gas and joint setup side of the re-weld, the checklist in How to Set Up a MIG Welder: Complete Step-by-Step Guide is a useful reset point before the arc is struck again.

Common Mistakes to Avoid

• Grinding only the visible pit and assuming the tunnel is gone.
• Re-welding before confirming the gas source is actually fixed.
• Skipping PT/MT because “it looks clean now.”
• Using a repair procedure that was never qualified for the joint.
• Auto-failing test coupons for worm tracks when the governing specification does not reject them.
• Failing to re-examine the repaired area with the original NDT method.

The Final Practical Checklist

1. Identify the defect type and confirm it is rejectable under the governing code before cutting.
2. Trace and eliminate the gas source: flux, wire, gas line, joint contamination, or shielding draft.
3. Remove metal along the tunnel axis by grinding or gouging, not just into the surface pit.
4. Verify removal with PT or MT after every grinding pass.
5. Re-qualify preheat, interpass temperature, and the repair WPS before rewelding.
6. Grind craters, overlaps, and stop-starts before depositing over them.
7. Re-examine with the original NDT method and document everything.

Worm welds punish shortcuts, but the pattern across this article is reassuring: they are traceable. Define the defect precisely, understand why gas gets trapped, work the cause list process by process, inspect with a method that can actually see the tunnel, apply the right code limits, prevent the problem at the source, and remove repairs the right way. Do that consistently and the radiograph stops being something your crew dreads. It becomes confirmation that the work was done right the first time.

Frequently Asked Questions

What is a worm weld or wormhole in welding?

A wormhole is an elongated, tunnel-shaped gas cavity in weld metal, formed when a large volume of gas is generated and trapped by the freezing weld pool. ISO 6520-1 lists it as imperfection 2016, and on a radiograph it appears as a tapering trace in a herringbone pattern. Shop crews also call related surface marks ‘worm tracks’ or ‘wagon tracks.’

Are worm tracks the same as wormholes?

No. Worm tracks are surface marks in FCAW caused by gas that already left the weld metal but got trapped under the solidifying slag; they run parallel to the weld axis. Wormholes are subsurface tunnels inside the weld, usually perpendicular to the weld face. Worm tracks alone are generally not rejectable under AWS D1.1 unless a WPS prohibits them.

What mainly causes wormhole porosity?

Gross surface contamination (oil, grease, rust, thick paint or primer), moisture in flux, electrodes or on the workpiece, and excessive weld pool turbulence from high current or travel speed. Deep, narrow joints and crevice geometries such as the gap under a T-joint vertical member make entrapment worse.

How do you detect wormholes inside a weld?

Radiography is the primary method — wormholes show a distinctive herringbone pattern that separates them from round pores and slag lines. Visual inspection only catches surface-breaking ends, and ultrasonic testing struggles with small pores in thick sections. PT or MT work for surface-breaking indications only.

Does AWS D1.1 allow wormhole porosity?

D1.1’s visual criteria address piping porosity: CJP butt groove welds transverse to computed tensile stress get none, and elsewhere the sum of visible piping porosity 1 mm or larger must stay within 10 mm per 25 mm and 20 mm per 300 mm of weld for statically loaded connections, with stricter or frequency-based limits for cyclic and tubular work. Anything beyond those limits must be removed and rewelded.

Can a wormhole weld be repaired?

Yes — the standard fix is to grind or gouge out the affected metal along the full length of the tunnel, verify removal (PT/MT is useful here), and re-weld to the qualified procedure. If porosity is widespread, the entire weld must be cut out and the joint re-prepared before rewelding.

Which welding process is most prone to wormhole defects?

Flux-based and high-current processes carry the highest risk: submerged arc welding (damp flux, insufficient flux burden) and flux-cored arc welding (moisture pickup, wrong stickout). Gas-shielded processes get them from shielding failures, and aluminum GTAW/GMAW is notoriously sensitive because of oxide and moisture on the surface.

How do I prevent wormholes on my next job?

Clean and degrease joint edges to bright metal, keep primer within the maker’s thickness limit, dry and store flux and low-hydrogen consumables correctly, check the gas system for leaks and drafts, control current, voltage and travel speed within the qualified range, maintain proper stickout, and avoid joint shapes that trap gas. Fix the gas source, not just the symptom.