Arc Burn in Welding: Causes, Risks, Repair & Prevention

A fabricator once told me the story of a pressure vessel nozzle weld that passed every visual check, only to be rejected days later because an inspector found a single fingernail-sized rough spot an inch away from the seam — a place where a tired welder had tapped his electrode to get the arc started. That tiny ‘arc burn’ looked harmless, but it had hardened the steel, created a stress raiser, and under cyclic load it was exactly where a crack would have begun. This is why a defect the size of a thumbnail can matter more than porosity or undercut, and why understanding arc burn is non-negotiable for anyone who strikes an arc.

What Is an Arc Burn? Defining the Defect Everyone Underestimates

An arc burn is an unintended arc started on base metal outside the intended weld zone. That’s the plain-English definition, and once you can picture it, the rest of the topic makes sense: a welder needs the arc running, taps or scratches the electrode on the plate for a second, drags a live torch while repositioning, or a tired ground clamp arces under current — and the arc touches metal that was never supposed to be welded. The arc does its damage in a fraction of a second, then moves on. What’s left is a small scar on parent metal that no weld ever called for.

Welders call it an arc burn because the arc literally burns a tiny patch into the surface. The formal name in codes and standards is arc strike, and the classification most often cited is ISO 6520-1 code 601, which defines it as a scar caused by the welding arc touching the parent metal outside the weld zone. ISO 6520-1 catalogs weld imperfections, and 601 sits in the family of unwanted metal discontinuities — not cosmetic blemishes, but changes to the material itself.

What It Looks Like and Where It Hides

An arc strike rarely announces itself. Typical appearances include:

• A fingernail-sized rough scar with a craters-and-ripples texture, sitting in otherwise smooth plate
• A small pit with spatter around it, often mistaken for ordinary spatter until you feel the depression
• A discolored, heat-tinted patch — blue, straw, or gray — where the arc grazed the surface
• A rough spot hidden directly under or beside the ground clamp, where nobody thinks to look

Locations follow a pattern: beside the seam rather than in it, along layout and chalk lines, near tack welds and fit-up areas, under earth clamps, on clad or coated surfaces where the arc has burned through to bare metal, and on finished faces that were supposed to stay untouched. If you want a sense of how forgiving the eye can be on coated work, prepping and inspecting surfaces like those covered in Aluminum Cleaner for Welding: How to Prep Aluminum Right shows how quickly a small surface defect disappears once a part is cleaned and dressed.

Why a Thumbnail-Sized Defect Gets Serious Attention

The size is deceptive. An arc strike isn’t just a mark; it’s a miniature, uncontrolled heat-affected zone — hardened metal, a geometric notch, and locked-in residual stress all in one spot. That combination is a ready-made crack initiator, which is why AWS D1.1, Structural Welding Code—Steel treats arc strikes as a distinct prohibited condition in Clause 5.29, with its own acceptance provision at C5.29, rather than lumping them in with general surface blemishes. Codes don’t usually write separate clauses for things that don’t matter.

It also explains why arc strikes show up so often on welder qualification test coupons. A coupon can have a structurally perfect weld and still fail, because arc strikes are among the easiest imperfections for an inspector to spot on the surrounding plate. The welder’s hand was good; the discipline around the weld wasn’t.

How Common Is It?

Very. Arc strikes occur in shop work and field work, on pressure vessels, structural steel, pipe, repair jobs, and test coupons, and they happen across every arc process: SMAW (see Advantages of Stick Welding: Why SMAW Still Wins), GMAW/FCAW, and GTAW. A loose or dirty earth/ground clamp can arc under welding current, leaving strikes that go unnoticed because they sit directly beneath the clamp itself.

The cost isn’t the grinding — it’s the workflow. A discovered strike stops the job, triggers an examination, possibly penetrant testing, a documented repair, and sometimes a discussion about re-qualification. Multiply that by a shift where several people are striking arcs, and a “harmless” habit becomes a real rework line item.

Over the next nine sections we’ll separate arc burn from the terms people mix up with it, look at what the strike actually does to steel metallurgically, walk through root causes, code acceptance rules, inspection methods, the correct repair sequence, and the personal-injury side of the phrase — UV burns, arc eye, and radiant-energy exposure — so you leave with a complete picture of a defect that is small, common, cheap to prevent, and expensive to ignore.

Arc Burn vs. Arc Strike vs. Arc Flash: Clearing Up the Terminology

In most shops, one phrase can mean two very different things. Say “arc burn” across the floor, and half the crew could be thinking of a scar on a plate, while the other half could be thinking of a welder in pain from their eyes. That mix-up isn’t harmless, as the response to each is entirely different. To avoid confusion, it’s worth breaking down the three terms properly.

Arc Burn and Arc Strike: The Metal Defect

When an arc is struck outside of the intended weld zone, it leaves behind an unintended mark on the base metal. This is most accurately known as an arc strike. While the colloquial “arc burn” is still used, inspectors and codes use “arc strike” almost exclusively.

These marks are formally classified by ISO 6520-1 as surface damage to the base metal or weld, given code 601. This codification matters, as once identified, acceptance falls under codes such as AWS D1.1 and any project-specific specifications. Depending on location and severity, an arc strike can be unacceptable and require removal.

Arc Flash: Two Meanings For One Phrase

The term “arc flash” is the biggest source of confusion, as it has two completely unrelated definitions:

• Electrical arc flash: A hazardous electrical discharge caused by a fault. This produces extreme heat, intense light, and pressure. It’s an electrical safety issue, most common around welding power sources or plant switchgear, and is handled by qualified electricians under electrical safety standards.
• Shop slang: Many welders use “arc flash” to mean the UV eye injury from an unprotected arc. This is used interchangeably with “arc eye” or “welder’s flash”, and is a personal injury rather than an electrical fault.

Arc Eye, Welder’s Flash, and Photokeratitis

The correct medical term for this eye injury is photokeratitis, which is essentially a UV burn to the cornea. It’s caused by exposure to the arc’s ultraviolet radiation, not a defect in the metal at all.

To prevent it, OSHA 29 CFR 1910.252(b)(2) requires arc welding and cutting operations to use helmets or hand shields with filter shades matched to the process. For example, shade 10 is recommended for 1/16- to 5/32-inch SMAW electrodes, while shades 12 to 14 are used for larger electrodes and high-amperage processes. If this injury occurs, the response is first aid and PPE correction, not grinding or weld repair.

Knowing the Difference Saves Time and Safety

The distinction comes down to two categories: a defect on the workpiece, or an injury to the person. Because of this, they fall under entirely separate rulebooks, as shown below:

Term	What it is	Primary standards	Correct response
Arc strike (arc burn)	Unintended arc scar on base metal (ISO 6520-1 code 601)	ISO 6520-1, AWS D1.1	Mark, examine, grind out, and repair per WPS
Arc flash (electrical)	Hazardous electrical discharge from a fault	NFPA 70E, plant electrical rules	De-energize and only allow a qualified electrician
Arc eye / welder’s flash (photokeratitis)	UV burn of the cornea	OSHA 1910.252, ANSI Z49.1	First aid, medical review, and correct shade/PPE

Both issues can be traced back to the same root cause: arc energy going somewhere it shouldn’t. That’s why prevention, from proper arc control to adequate protection, tends to address them both at once. For those looking to learn how to spot these defects in a structured setting, our Aluminium Welding Course: What You Learn, Costs & Careers covers arc strikes alongside other weld discontinuities.

Going forward, this article will use “arc strike” for the metal defect, and “arc eye” or “photokeratitis” for the injury to keep everything clear.

The Metallurgy: What an Arc Burn Actually Does to Steel

To understand why an inspector gets serious about a mark the size of a thumbnail, you have to look at what happens in the seconds after a stray arc touches the plate. The damage is not really the scar you can see. It is the invisible metallurgical change underneath it.

The thermal cycle of a stray arc

An intentional weld follows a qualified procedure: specified preheat, controlled heat input, monitored interpass temperature, and sometimes a defined post-weld cooling rate or immediate temper bead treatment. A stray arc has none of that discipline. It delivers an intense burst of heat to a very small spot, and everything around it is cold, solid steel acting as a massive heat sink. Heat is pulled away almost instantly, so cooling rates are far faster than anything the qualified weld would ever experience.

That rapid cycle creates what is essentially a miniature heat-affected zone in a place that was never supposed to see one. The base metal near the arc heats above its transformation temperature and then quenches. In carbon and low-alloy steels, that combination of fast cooling and low heat input favors hard, brittle, untempered martensite, along with elevated local hardness and a real hydrogen cracking risk if the electrode or environment introduced moisture. In other words, the spot stops behaving like the plate it sits on.

Residual stress: the spot pulls against itself

As the molten spot solidifies and cools, it contracts. The surrounding metal refuses to move with it, so tensile residual stress gets locked into and around the strike. Now you have two problems stacked on top of each other: a hard, brittle zone with poor toughness, and tensile stress pulling at it before any external load is even applied.

The notch effect: why geometry makes it worse

Most arc strikes leave a rough, crater-like surface or a small pit. Mechanically, that is a geometric notch, and a notch concentrates stress. So the practical recipe for failure is complete: hard brittle material, residual tensile stress, and a stress raiser sitting exactly where the brittle material is. Under cyclic tensile loading, that is a classic fatigue crack initiation site. A weld may pass visual inspection and static proof loading and still fail later at precisely the strike nobody thought mattered. If you have ever wondered where mysterious fatigue cracks in boiler and pressure vessel work originate, unremoved arc strikes are a recurring suspect.

Factor	Qualified weld	Arc strike
Preheat	Specified and applied	None
Cooling rate	Controlled by procedure	Extremely fast (cold heat sink)
Microstructure	Balanced, toughness verified	Hard untempered martensite possible
Residual stress	Managed, often post-weld treated	High localized tensile stress
Surface geometry	Smooth, blended profile	Notch-like pit or rough crater

Which steels are most vulnerable

Not every material reacts the same way. High-strength and alloy steels, quenched-and-tempered grades, and anything with high hardenability are the worst cases, because they transform to martensite readily even at moderate cooling rates. That is why selection of alloy steel welding electrodes and matching procedures matters so much in these grades: the same plate that welds beautifully under a proper procedure can become notch-sensitive with one careless arc start.

Stainless and clad materials pose a different problem. A strike on stainless may look cosmetically harmless, but the localized heat can cause sensitization, where chromium carbides precipitate at grain boundaries and rob the surrounding metal of corrosion resistance. On clad or coated surfaces, the strike can also destroy the protective layer, creating a corrosion cell in service. Visually smooth does not mean metallurgically clean.

The bottom line: an arc burn is small on the surface and large underneath. It introduces a hard zone, residual tension, and a notch into material that was never qualified for any of them, which is precisely why prevention and proper repair matter far more than the size of the mark suggests.

Why Arc Burns Happen: The Real Root Causes

It is tempting to blame arc strikes on a careless welder, and plenty of shop conversations end exactly there. But walk a fabrication floor during a demanding fit-up and you will see that stray arcs are rarely one person’s mistake. They are the predictable output of a system: equipment that was not checked, a joint design that gave nowhere safe to start, a schedule that pushed a tired welder into an awkward corner, a ground clamp bolted to painted steel. Industry guidance on arc strikes consistently groups the causes into a handful of repeatable mechanisms, and almost all of them are fixable without asking anyone to simply “be more careful.”

The seven mechanisms behind a stray arc

1. Arc starts outside the joint. The classic cause. A welder taps or scratches the electrode on the plate to get the arc going, then walks it into the groove. Every one of those taps on parent metal is a potential strike. The correct habits are striking inside the joint, on the bevel, or on a run-on or run-off tab that gets removed afterward.
2. Dragging a live electrode or torch. Repositioning between welds with the arc still live, or letting a hot electrode tip trail across the surface, leaves a line of small craters behind it. This happens most often on long seams where the welder moves along the part instead of stopping the arc first.
3. Accidental contact during movement. In cramped, out-of-position, or overhead work, the nozzle or electrode simply bumps the workpiece on the way to the next start. Tight access multiplies this risk sharply.
4. Defective earth or ground connections. A loose, painted, or corroded clamp cannot carry welding current cleanly, so the current arcs across the gap at the clamp itself. The result is strikes right under or immediately around the clamp, a spot inspectors specifically check.
5. Damaged holders and cables. Cracked insulation, a worn electrode holder, or a live part touching the workpiece creates unintended paths to the arc. Cable routing across sharp edges and walkways accelerates this damage.
6. Magnetic arc blow. On ferrous workpieces, magnetic forces in the plate deflect the arc off its intended path, especially near joint ends and corners. The arc lands where nobody aimed it, and a strike follows.
7. Environmental and human factors. Poor lighting hides small strikes before they matter, fatigue slows reaction time, wind or disturbance wrecks shielding gas coverage and pushes the welder into rushed corrections, and time pressure encourages shortcuts. Research summarized by TWI on heat-affected zone behavior underlines why even a small uncontrolled thermal event matters, but the trigger for that event is usually situational, not skill-related.

Cause	Typical scenario
Striking on the plate	Welder taps the electrode beside the groove to start the arc on a bracket assembly
Live electrode drag	Repositioning along a long fillet seam without breaking the arc
Accidental contact	Nozzle bumps the inner wall while welding an out-of-position pipe joint
Poor ground connection	Clamp attached over primer; current arcs at the clamp, leaving strikes beneath it
Damaged holder or cable	Frayed lead resting on the edge of the plate arcs during travel
Magnetic arc blow	Arc wanders at the end of a butt joint on heavy mild steel
Human and environmental factors	Fatigued welder on a late shift works under dim light and rushes arc starts

Notice how many of these causes are equipment and setup problems rather than skill problems. A welder running alloy steel electrodes on a hardenable plate with no preheat is working with almost no margin: a single stray start on that material creates exactly the hard, untempered zone described earlier. The same welder with a clean clamp, a tab to start on, and adequate light may never produce a strike in a hundred hours.

That is the core argument for systemic prevention. Even highly experienced welders produce arc strikes under fatigue, awkward access, or time pressure, because the causes sit in the work conditions rather than the hands. A shop that only counsels individuals will keep finding the same scars; a shop that audits clamps, redesigns starts around tabs, protects finished surfaces, schedules realistically, and trains deliberately removes the conditions that create strikes in the first place. As practical industry guidance on arc strikes puts it, the defect is common, cheap to prevent, and expensive to discover late, so the real work of prevention happens before the arc is struck at all.

Code and Standard Rules: When an Arc Burn Is Acceptable (Almost Never)

Ask a fabricator whether an arc strike matters and you’ll usually get a shrug. Ask an inspector and you’ll get a completely different answer. The reason is that codes don’t evaluate arc burns by size or appearance; they evaluate them by location, by loading condition, and by what happened to the metal underneath. What looks like a freckle on the plate is, in the eyes of the standards, an uncontrolled thermal event in material that was never qualified to receive one.

What AWS D1.1 Actually Requires

The structural welding code most shops work under is AWS D1.1, Structural Welding Code – Steel. Its position is blunt: arc strikes outside the area of permanent welds should be avoided on any base metal, and any crack or blemish caused by an arc strike must be ground out and the affected area restored smooth. There is no “it’s only cosmetic” allowance. Once a strike has hardened or cracked the surface, grinding to sound metal is the entry ticket, not the finish line.

D1.1’s repair provisions, governed by section 5.29 (and the commentary at C5.29), then split the response into two very different paths depending on how the part is loaded:

Condition of the base metal	Permitted restoration path	Verification
Not subject to cyclic tensile stress	May be restored by welding under a qualified repair WPS, with preheat and controls matching the base metal	Soundness verified by NDT where the code or specification requires it
Subject to cyclic tensile stress	Restorable only with engineer approval, under a repair WPS, following the code’s stricter pathway	Specified NDT; grinding alone is rarely the end of the story

That single distinction is why the same thumbnail-sized strike can be a five-minute grind-out on a static bracket and a documented engineering decision on a bridge girder or a crane chord. Under cyclic tension, the hard spot and its notch are crack initiators, so the code refuses to let a welder make the call alone.

Projects running under AWS D1.8, the seismic supplement, impose tighter controls still, and individual project specifications frequently outrank both. It’s also worth remembering that D1.1 sits inside a broader safety framework: OSHA 1910.252 requires employers to control welder exposure and hazardous conditions in welding operations, and shops working in sanitary or process piping will find ASME-level acceptance criteria that treat stray arcs with similar severity.

ISO 5817: Three Quality Levels, One Verdict

International work usually runs on ISO 5817, which sets acceptance levels for fusion-welded joints in metallic materials. Its treatment of arc strikes is a clean ladder:

Quality level	ISO 5817 category	Arc strikes
B	Higher requirements	Not permitted
C	Middle requirements	Not permitted
D	Lower requirements	Conditionally tolerated only after grinding and inspection

Read that carefully: only the lowest quality tier even considers an arc strike, and only after it has been dressed and checked. On anything held to level B or C, there is no acceptance criterion to argue with. The industry guidance on arc strikes makes the same point from the shop floor side: codes are written to eliminate the defect, not to rate it.

The Edition Question and the Contract Question

Two practical traps catch experienced shops. The first is the contract. Even where a code allows a repair route, contractual documents can impose zero tolerance, and the publishing trade coverage of arc strikes repeatedly notes that owners and inspectors often treat any strike outside the weld zone as a nonconformance regardless of what a repair clause technically permits. Always read the specification before you grind.

The second trap is the edition. AWS has moved D1.1 to a five-year revision cycle, and the 2025 edition updates prequalified WPS requirements and visual inspection acceptance criteria, so a shop still working from a 2020 or 2015 copy may be applying superseded language. Confirm which edition the contract invokes before pricing a repair.

The takeaway is simple: static versus cyclic loading changes the repair path dramatically, quality level B or C means no arc strikes at all, and the contract has the final word. If you’re still building habits around arc starts and technique, our alloy steel welding electrodes guide covers the consumable side of running clean, controlled arcs on hardenable material.

Finding Arc Burns: Inspection and Detection Methods

Finding an arc burn is less about exotic equipment and more about knowing where to look. Unlike porosity or undercut, a strike hides in plain sight: it sits in parent metal that everyone assumes is smooth, untouched and uninteresting. That assumption is exactly what lets a thumbnail-sized defect travel all the way to final inspection.

Visual Inspection Still Does the Heavy Lifting

Visual inspection remains the primary detection method for arc strikes, and it works because the defect is a surface phenomenon. A trained inspector does not scan only the weld bead and heat-tinted zone; they walk the entire component surface, looking for small rough scars, discolored patches, and the fine spatter-associated pitting that often betrays where an arc landed unintentionally. Under good lighting, per standard visual inspection practice in AWS D1.1, these marks are obvious — they interrupt an otherwise uniform, mill-scale or grind-finished surface.

That contrast is why visual catches most strikes. Welders and inspectors are trained to expect roughness inside the joint, spatter near the seam, and discoloration from heat. Outside the joint, they expect cleanliness. A strike breaks that expectation, and a good inspector knows that the areas around the weld often reveal more than the weld itself.

Where to Look First

Before any formal procedure, work this short checklist. It covers the spots where strikes cluster on real jobs:

• Near and under the ground clamp — a loose or dirty clamp arcing under welding current leaves strikes right at the contact point, sometimes hidden beneath the clamp itself.
• Arc starts and stops — the first scratch of the electrode and the tail of the crater are prime territory.
• Layout lines and tack areas — strikes appear along scribed lines where the welder repositioned, and around fit-up tacks that were struck off-face.
• Repair and rework zones — second and third visits to a joint are where fatigue and hurry do their damage.
• Passageways and cable routes — anywhere a live electrode or torch was dragged while moving between welds.

Escalating to Non-Destructive Testing

When a strike is found in critical service, or where grinding has already been performed to remove it, the visual check is only the starting point. Liquid penetrant inspection (PT) to ISO 3452-1 — equivalent to ASME Section V, Article 6 — is the standard surface method for confirming that a ground-out strike contains no residual crack. It is inexpensive, portable, and sensitive to the tight, tight discontinuities a strike can leave behind after the visible scar has been ground away.

Magnetic particle testing (MT) applies to ferrous components and has the advantage of detecting slightly subsurface indications, which makes it useful where a strike may have driven a crack just below the surface. Eddy current methods (ISO 17643 / ASME Section V, Article 8) are less common for this defect but find use on tubular geometries where a probe can be swept consistently. Eddy current and UV-assisted techniques remain secondary options; most shops still resolve arc strikes with PT or MT plus a good eye.

Hardness Spot-Checking and Documentation

On hardenable carbon and low-alloy steels, a suspect spot may warrant a portable hardness check. Hardness testing can confirm whether the spot’s microstructure stayed within the limits specified for the joint — a quick way to learn whether grinding removed the hardened zone or merely polished it. This matters most on quenched-and-tempered and high-strength materials, where an untempered martensitic patch can survive a clean-looking grind-out. The underlying metallurgy is covered in TWI’s guidance on hardness, microstructure and toughness in steel HAZs, and selecting the right consumable for alloy work is a related concern addressed in our guide to Alloy Steel Welding Electrodes.

Documentation closes the loop. Strikes and their repairs are normally recorded on the welder’s or repair record, and repeat occurrences tend to trigger a conversation about re-qualification rather than another grind-out. That feedback loop is what turns a one-off repair into a systemic fix.

Repairing an Arc Burn: The Correct Step-by-Step Response

Finding an arc strike is only half the job. What you do in the next ten minutes decides whether the mark disappears from the record or follows the part into service. The good news is that repair is rarely complicated. The bad news is that sequence matters, and skipping steps is how a thumbnail-sized defect turns into a re-repair, a rejected inspection, or worse, a crack that starts in service.

The Six-Step Repair Workflow

1. Stop and mark the strike immediately. Do not let it get buried under paint, primer, or coating. Circle it with a paint pen, note the location on your fit-up sketch, and report it before moving on. A strike discovered under a coat of primer costs several times more to fix than one caught while the metal is still bare, because now someone has to strip, expose, re-inspect, and touch up the coating as well.
2. Examine it properly. Start with visual inspection under good lighting: is it a rough scar, a discolored patch, or is there a visible crack line? If the part is critical, or the strike sits in a stress-sensitive area, escalate to liquid penetrant testing (PT, per ASME Section V Article 6 / ISO 3452-1 practice), or magnetic particle testing (MT) for ferrous components. This tells you whether you are polishing out a blemish or removing a crack.
3. Grind out the affected metal completely. The goal is not to make the spot pretty; it is to remove enough material to eliminate the hardened heat-affected zone and any crack indication underneath. Blend the edges smoothly with no sharp transition, then stop and look again.
4. Verify with re-inspection. Penetrant applied after grinding is the common acceptance gate — it confirms the hardened and possibly cracked material has been fully removed. If PT still shows an indication, you grind deeper. Do not proceed to the next step until the indication is gone.
5. Decide whether grinding alone satisfies the code and specification. For many static, non-critical locations, a properly blended grind-out is the accepted finish, especially where loss of thickness is negligible. Check the contract documents before assuming, because project specifications can impose tighter requirements than the base code.
6. If base-metal thickness must be restored, weld only under a qualified repair WPS. AWS D1.1’s repair provisions draw a sharp line here: base metal not subject to cyclic tensile stress may be restored by welding under a prepared repair procedure, with soundness verified by the appropriate NDT when the contract documents specify it. Base metal subject to cyclic tensile stress requires engineer approval before repair welding, reflecting how fatigue-sensitive that location is. Either way, preheat and interpass temperature must suit the base metal, and the filler should match the filler selection for that alloy.

Finally, re-inspect the completed repair, document what was done, and log it on the welder’s or repair record. Strikes and their repairs are usually tracked, and a paper trail protects both the shop and the welder if a pattern shows up later.

The Classic Mistake: “Just Touch It With a Little Weld”

This is the single most common bad fix in the shop. A welder spots a strike, grabs a rod, dabs a quick blob over it, and grinds it flush. But filling a strike without a repair WPS repeats the exact same uncontrolled thermal cycle that created the problem in the first place: arc heat dumped into cold base metal with no preheat, no interpass control, and no cooling discipline. The result is often an even larger unqualified HAZ than the original strike, now hidden under fresh weld metal. If the strike needs filling, it needs a procedure like any other weld.

On aluminum and coated or clad surfaces, add one extra habit: clean the ground area thoroughly before and after grinding so embedded grinding residue or oxide does not contaminate any subsequent weld. The pre-weld chemistry matters as much as the grind geometry.

Treated this way, an arc burn repair is a five-minute task with a paper trail. Skipped or improvised, it becomes an inspection finding, a corrective-action report, and a conversation about welder re-qualification nobody enjoys.

Prevention Playbook: How to Stop Arc Burns Before They Start

The best fabrication shops know that arc strikes are not a product of carelessness, but of process gaps. When time is tight and space is limited, telling a welder to simply “be more careful” rarely works. Instead, the most reliable approach is to build a system where the arc has nowhere to land except where it belongs. This playbook focuses on four layers of prevention: technique, equipment, job setup, and shop systems. Each one plays a role, and neglecting any of them will allow strikes to return.

Technique: Control Every Arc Start

Where an arc begins is the most impactful habit to lock in. Every strike should start in the groove, on the bevel face, or on a run-on tab, never on finished parent metal. As noted in industry guidance on arc strikes, beginning the arc in the joint or on a run-on tab allows it to stabilize before moving into the weld path. Those extra seconds can prevent hours of grinding later.

For stick welding, deliberate start methods like scratch or lift starts are effective, but they need to be intentional to the procedure. When working with SMAW, revisiting the fundamentals of arc control helps reinforce consistency. This becomes especially vital when using techniques for all-position welding rods in overhead or out of position work, where fumbled starts are far more common. Most importantly, reposition with the arc off, and when an electrode becomes short, swap it for a fresh one rather than dragging a live rod across the workpiece.

Equipment Discipline: Grounds, Cables, and Holders

Even the best technique can be undone by faulty electrical connections. Loose, corroded, or undersized ground clamps create poor contact points that can arc just as easily as an electrode tip. These strikes often cluster near the clamp or beneath it, making them easy to miss. With this in mind, ground connections should be inspected, tightened, and cleaned back to bright metal before every job. This practice falls directly under the maintenance requirements outlined in OSHA 1910.252 for welding equipment in safe operating condition.

On the hot side, the same scrutiny applies to electrode holders, cable lugs, and insulation. Route and secure cables away from sharp edges, walkways, and clamp zones to prevent them from shifting mid-weld. For crews looking to standardize these checks, implementing a practical welding equipment inspection routine is one of the most cost-effective defenses against stray current.

Job Setup: Design the Work So Strikes Cannot Happen

Proper setup removes opportunities for error before the first spark. Pre-position and tighten all clamps, jigs, and fixtures prior to striking an arc to prevent repositioning with a hot conductor. Where possible, add sacrificial run-on and run-off tabs, especially on visible surfaces, with removal planned as part of finishing. Keep tacks inside the joint or on areas destined to be welded or ground, and protect finished faces with scrap plate or heat-resistant covering. For stainless, clad, or hardenable alloys, this protection is critical. Understanding the proper use of alloy steel welding electrodes reinforces just how unforgiving these materials are to stray heat and arc burns.

Shop Systems: Training, Lighting, Scheduling, and Feedback

The environment around the welder directly affects strike frequency. Training and qualification programs should explicitly test arc starting skills, not just final weld quality. Adequate workface lighting is vital, as a strike is far easier to prevent than to locate under primer. Realistic scheduling also matters, as fatigue and rushed fit-ups drastically increase the risk of error. Finally, close the inspection loop by utilizing proven monitoring and inspection practices that flag strikes early, allowing recurring patterns to be traced to a station, clamp, or procedure, rather than placing blame on an individual.

Ten Rules to Post on the Wall

1. Start every arc in the joint, on the bevel, or on a run-on tab.
2. Never strike on finished parent metal, ever.
3. Reposition with the arc off; swap short electrodes instead of dragging them.
4. Check, tighten, and clean the ground clamp before every job.
5. Inspect holders, lugs, and cable insulation at the start of each shift.
6. Pre-position clamps and fixtures before welding begins.
7. Use sacrificial tabs on visible and inspectable surfaces.
8. Tack inside the joint, not on finished parent metal.
9. Protect adjacent surfaces, especially stainless, clad, and alloy steels.
10. Mark and report any strike immediately, before coatings go on.

None of these actions require new technology or a larger budget. They only require treating the arc start as a controlled operation and granting ground clamps and cables the same level of respect as the electrode itself. With consistency, arc burns shift from a routine repair item to a preventable occurrence.

Arc Burns on People: Flash, Arc Eye, and Radiant-Energy Burns

So far this article has treated “arc burn” as a metal defect. But type that phrase into a search bar, or say it in a shop, and you will just as often get a story about a person. A welder finishes a long afternoon on structural plate, drives home, and by evening his eyes feel full of sand. This is the other meaning of arc burn, and it deserves the same seriousness we give arc strikes in steel.

Photokeratitis: The “Sunburn” You Get Inside a Helmet

An open arc produces an enormous amount of ultraviolet radiation. That UV is what makes the process work, and it is also what burns living tissue. When it reaches the eye, the result is photokeratitis, an actual burn of the cornea’s outer surface. Welders call it arc eye, flash burn, or welder’s flash, and the mechanism is exactly what it sounds like: the same injury as a severe sunburn, only on the clear front window of the eye instead of your shoulders.

The cruel part is the timing. Symptoms typically arrive several hours after the exposure, which is why most welders don’t realize they were flashed until the end of the shift, or the next morning. Classic signs include:

• Sharp or gritty, sandy pain in one or both eyes
• Extreme light sensitivity, sometimes enough that a dim room still hurts
• Excessive tearing and a feeling of grit under the lids
• Temporary blurred or hazy vision

Mild cases resolve in a day or two with rest and time. What matters for long-term health is repetition. Cumulative UV exposure is associated with chronic eye damage, including an increased risk of cataracts later in a welding career. Each individual flash feels temporary; the damage tally is not.

Infrared, Visible Light, and Skin: The Rest of the Radiation Story

UV is not the only hazard coming off the arc. Infrared radiation penetrates deeper into the eye than UV does, and its damage to the retina is both cumulative and painless, which makes it the quiet hazard. Intense visible radiation causes temporary flash blindness, and that is its own workplace danger: a welder who has just been flashed can misjudge a step, a crane load, or a hot surface.

Skin is vulnerable too. Exposed arms and necks receive UV burns that behave like a bad sunburn, which is the real reason the dress code is what it is. Flame-resistant jackets,gauntlets-style gloves, and neck protection or a collar are not decoration. Process matters here as well. Processes that run hotter and brighter, or that involve auxiliary operations like air arc welding and carbon arc gouging, push radiation output up sharply compared with low-amperage work.

The Protection Framework: What the Rules Actually Require

The governing U.S. requirement is OSHA 29 CFR 1910.252(b)(2), which requires helmets or hand shields with appropriate filter lenses for all arc welding and cutting operations. OSHA’s accompanying table recommends specific shades by process, with roughly shade 10 for common SMAW electrodes and around shade 12 for gas-shielded ferrous welding. Alongside that, ANSI Z49.1 from the American Welding Society provides the current, more detailed filter-selection guidance that most shops actually follow, and it is referenced alongside OSHA’s eye and face protection requirements.

Work	Typical filter shade
Low-amperage GTAW	Shade 8 to 10
Common SMAW electrodes	Shade 10
Gas-shielded ferrous welding (GMAW/FCAW)	Shade 12
High-amperage SMAW/FCAW, heavy section work	Shade 13 or higher

Auto-darkening helmets cover this range well, generally from about shade 8 at low amperage up to shade 13 and beyond for heavy stick and flux-core work. The convenience is real, but the shade still has to match the job, and the helmet must be paired with clear protective spectacles with side shields underneath. If you are running demanding processes such as all-position flux-core on structural steel, our alloy steel welding electrodes guide covers the amperage and technique settings that affect how much radiation you are dealing with.

The same logic extends to every radiant-energy process in the shop. Anything that concentrates energy on a surface, from laser systems to high-output plasma, demands its own eyewear calculation, which is why laser operations are governed by separate standards and equipment-specific protection, as covered in our alpha laser welding guide. The principle never changes: match the protection to the radiation, not to habit.

The takeaway: on the metal side, an arc burn is a defect to grind out and document. On the human side, it is a preventable injury with delayed symptoms that welders tend to shrug off. Both come from the same source, and both are managed the same way: respect the arc before you strike it, not after.

Common Mistakes, Troubleshooting, and a Practical Daily Checklist

Most shops that keep finding arc strikes are not dealing with careless welders. They are dealing with habits and systems that were never designed to catch the problem early. Here is what keeps the cycle going, how to read the pattern when strikes repeat, and a checklist your crew can actually use.

Mistakes That Keep Arc Strikes Coming Back

• Grinding for looks, not for metallurgy. A quick pass with a flap disc makes the scar shiny and the inspector’s eye slide past it, but a cosmetic grind rarely removes the full hardened heat-affected zone underneath. Acceptance is not “it looks smooth” — the ground area must be smooth and free of residual cracks, verified by inspection where the code or specification calls for it. If a strike sits in a stress path, that verification is penetrant or magnetic particle testing after grinding, not a glance.
• Touching it with a little filler. Dabbing weld metal over a strike without a qualified repair WPS simply recreates the same uncontrolled thermal cycle that caused the problem: no preheat, no interpass control, no cooling discipline. You have traded one defect for an identical one with extra grinding to come.
• Burying it under coating. Priming or painting a strike before inspection turns a five-minute grind-out into a stripping, re-inspection, and rework job. Nothing should cover parent metal until the surface has been accepted.
• Treating it as cosmetic. Codes treat arc strikes severely relative to their size precisely because the consequence — fatigue crack initiation — is disproportionately large compared with the thumbnail of damaged metal involved.
• Ignoring strikes on cyclically loaded or tubular connections. These are the highest-consequence locations, and the repair pathway changes entirely. Cyclically stressed base metal generally needs engineer approval and specified NDT before it goes back into service.
• Skip the paperwork. Repairs and their inspection results should be documented. Under code frameworks, repeat welder issues can trigger re-qualification discussions, and you cannot have that conversation without records.
• Blaming the individual instead of the system. If the same shop keeps producing strikes, the fix is layout, fixturing, equipment maintenance, and procedure design — not another lecture. For background on why welders chase clean starts under pressure, our Aluminium Welding Course: What You Learn, Costs & Careers breakdown of how arc-start discipline is actually taught is worth a read.

Troubleshooting: Read the Pattern, Not the Blame

Pattern you see	Likely root cause	First action
Strikes keep appearing at the same workstation	Dirty or loose ground clamp, damaged holder, compromised cable insulation arcing under welding current	Inspect and clean clamp contact surfaces, tighten connections, check holder and insulation before touching technique
Strikes cluster at arc starts	Wrong amperage, arc started on parent metal instead of in the joint or on a run-on tab	Correct settings, retrain start method, add sacrificial tabs on visible surfaces
Strikes appear during fit-up and tacking	Poor fixturing, rushed tacking, no protected layout surface	Review tack placement, pre-position clamps, mask finished parent metal
Random strikes across many parts	Fatigue, time pressure, poor lighting, inadequate training	Adjust scheduling, improve task lighting, revisit qualification

The Daily Checklist

Before the arc:

1. Ground clamp tight, clean, and biting sound metal; cables routed away from edges and walkways.
2. Holder, liner, and insulation free of damage; no live parts able to touch the workpiece.
3. Amperage and process settings confirmed for the joint, so the arc starts the first time.
4. Run-on/run-off tabs, masking, or protective cover in place on finished surfaces.
5. Lighting adequate to see the whole workpiece, not just the puddle.

After the weld:

1. Scan the full component — near the ground clamp, at starts and stops, along layout and tack lines, and across repair and fit-up zones — before any coating goes on.
2. Mark any strike immediately with a paint pen so it cannot disappear under primer.
3. Report it, grind it out per the repair steps, and get the required PT/MT verification.
4. Log the repair and its inspection result on the welder’s record.

Everything in this article circles back to one asymmetry: an arc burn is tiny, cheap to prevent, and expensive to ignore. The difference between a professional fabrication shop and one living in repair mode is not talent or equipment price. It is whether the strike gets caught today, fixed correctly, and recorded — or rediscovered by an inspector with a magnifier and a rejection stamp.

Frequently Asked Questions

What is an arc burn in welding?

An arc burn — standardized as an arc strike (ISO 6520-1 code 601) — is a small scar left when the welding arc is accidentally initiated on base metal outside the intended weld joint. It creates a localized hardened heat-affected zone that can crack or act as a fatigue crack initiation site.

Are arc strikes allowed by welding codes?

Essentially no. AWS D1.1 requires cracks or blemishes from arc strikes to be ground out and the area restored smooth, and ISO 5817 prohibits arc strikes at quality levels B and C, tolerating them at level D only conditionally. Project specifications may be stricter still.

Why are arc strikes dangerous if they look so small?

The rapid heating and cooling of a stray arc can form hard, brittle untempered martensite and lock in tensile residual stress, creating a stress concentration. Under cyclic loading, that small spot becomes a classic fatigue crack initiation point — the reason codes treat them far more severely than their size suggests.

How do you repair an arc strike on structural steel?

Grind out the affected material completely, verify with visual and, where required, penetrant or magnetic particle inspection that no crack remains, then leave it blended smooth or restore the base metal by welding under a qualified repair WPS. Repairs on base metal under cyclic tension require engineer approval per AWS D1.1.

What is the difference between an arc strike and arc flash?

An arc strike is a defect in the metal — an unintended arc on the workpiece. ‘Arc flash’ is informal language that can mean an electrical discharge event or, in most welding shops, the UV eye injury (photokeratitis, also called arc eye or welder’s flash) caused by looking at the arc without proper filter protection.

Can arc strikes cause cracks in the base metal?

Yes. The hardened, brittle local microstructure combined with residual tensile stress and the notch geometry of the strike can initiate cracks, especially in high-strength, alloy, or hardenable steels and in components subject to cyclic loading.

How do I prevent arc burns while welding?

Start the arc in the joint or on a run-on tab — never on finished parent metal; keep the electrode or torch controlled when repositioning; maintain tight, clean ground clamps and sound cables; and use adequate lighting and realistic scheduling. These systemic controls outperform ‘be more careful’ reminders.

What filter shade protects against arc flash (arc eye)?

OSHA’s guide recommends shade 10 for common SMAW electrodes and shade 11-12 for gas-shielded welding, rising to 14 for large electrodes and carbon arc work. Current practice follows ANSI Z49.1 shade tables, and auto-darkening helmets typically run from shade 8 for low-amperage TIG up to shade 13+ for high-amperage processes.