How Does Arc Blow Affect Welding? Causes, Defects, Fixes

A pipefitter sets up a root pass on a long-seam pipeline, strikes an arc, and watches the puddle race ahead of him like something has grabbed the electrode. The bead comes out narrow, undercut and half-melted down one side, and no amount of technique fixes it, because nothing about his technique is wrong. That invisible tug is magnetic arc blow — and it is one of the most common reasons a perfectly good procedure suddenly produces garbage.

What Arc Blow Is — and How It Affects Your Weld Right Now

Straight answer first: arc blow is the deflection of the welding arc away from the shortest path between the electrode and the workpiece, caused by an unbalanced magnetic field around the arc.

You expect a stable plasma column running straight from the electrode tip to the puddle, but with arc blow that column bends. Lincoln Electric describes it as the arc stream failing to follow the shortest path between electrode and workpiece, so it is deflected forward or backward from the direction of travel, and less often sideways. The cause gets its own section below; for now, accept that a magnetic field can overpower the arc’s self-righting tendency and drag it sideways, forward or backward.

The Four Names You Will Hear on the Shop Floor

TWI defines magnetic arc blow, also known as arc wander, as arc deflection caused by distortion of the magnetic field produced by the arc current. You will hear several names:

• Magnetic arc blow — the umbrella term: deflection caused by magnetic forces rather than your hand, the travel mechanism or the shielding gas. Also called arc wander or arc deflection.
• Forward blow — the arc is pulled ahead of the direction of travel.
• Back blow — the arc is pushed behind the direction of travel, trailing the work.
• Side blow — the arc walks off to one side, near edges, corners and asymmetric joint geometry.

The direction of the blow tells you which defect to expect.

What Arc Blow Does to the Weld You Are Making Right Now

Arc blow is not cosmetic. When the arc wanders, three things stop landing where your WPS says they should:

1. Filler metal — drops fall in the wrong place, so the bead runs off-centre and the reinforcement ends up lopsided.
2. Heat — deposited at a moving point, so penetration and bead profile change along the seam.
3. Arc force — the pressure that shapes the puddle and drives penetration into the root no longer acts along the joint axis.

The result: a wandering puddle, a bead that no longer matches your procedure, and loss of fusion at the sidewall or root. SSAB’s technical support note on magnetic arc blow is blunt: arc blow can produce several kinds of discontinuity, including lack of root penetration, lack of fusion, undercuts and pores, each one a defect a CWI stops the job to discuss and each means grinding out a pass and re-welding it.

The Signs, in the Order They Show Up

Read the bead, because that is where diagnosis starts.

Type of blow	What you see in the bead
Back blow	Spatter; continuous or intermittent undercut; a narrow, high bead; increased penetration; surface porosity at the finish end of sheet metal welds.
Forward blow	A wide bead of irregular width; a wavy bead; usually intermittent undercut; decreased penetration.

Narrow, high and undercut with spatter means back blow. Wide, wavy and shallow means forward blow. SSAB’s torch angle guidance matters here, since the angle you hold is one of the cheapest ways to fight the deflection.

Not All Arc Blow Is a Problem

Before you chase a phantom: a small amount of arc blow is not always harmful. Lincoln Electric notes that a small amount can sometimes be used beneficially to help form bead shape, control molten slag and control penetration. A controlled deflection is a tool; an uncontrolled one is a defect.

What Arc Blow Is Not

Clear away four common misdiagnoses:

• It is not high arc voltage. A long arc gives a wide, cold, poorly fused bead, but it will not bend sideways off the joint. Different failure mode, different fix.
• It is not a faulty electrode. Changing rods does not remove a magnetic field from the plate. Blow on three different electrodes and the joint is still the common factor.
• It is not a bad grinding wheel. A contaminated wheel causes inclusions, not deflection.
• It is not simply operator skill. This one does the most damage in a shop, because it turns a physics problem into a performance review. A welder on a magnetised joint cannot make the arc behave.

What This Article Will Do for You

The rest of this piece follows four steps: the physics of magnetic deflection, symptom-to-defect mapping, prevention technique at the gun, and demagnetisation for when technique alone is not enough.

One note before you read on: if you work in stainless or aluminium regularly, the joint preparation and filler selection in our guides to aluminum MIG welding and aluminum flux core welding are worth knowing, but they will not rescue a magnetised steel root pass. That is a separate argument entirely.

Let us start with the mechanism, because once you understand why the arc bends, the fixes stop being folklore.

The Electromagnetic Physics Behind Arc Bending

A welding arc is a conductor carrying current, and every current-carrying conductor is wrapped in its own magnetic field. When that field turns lopsided, the arc gets pushed off line. The same principle that swings a compass needle bends a weld arc: iron decides where the magnetic lines of force travel, and the arc follows.

Every Arc Builds Its Own Magnetic Field

Current flowing down the electrode and across the gap creates a self-induced field, with concentric lines of force at right angles to the conductor. Hold a compass near a live DC lead and the needle swings.

The arc is never alone. Current spreads through the workpiece and returns through the work lead, generating a second field. The two distort each other, and the arc is pulled toward whichever side has the stronger magnetic pull. Lincoln Electric puts it neatly: where steel replaces air, flux lines squeeze together as they cross from one material into another, and the arc moves sideways to relieve that imbalance.

Why Edges, Corners and Gaps Pull the Arc

Steel offers a far easier path for flux than air. Wherever the field crosses that boundary the lines crowd together and push on the arc, which is why the trouble spots are so predictable.

Where flux concentrates	What happens
Plate edges and open corners	Flux crowds into the remaining steel instead of air
Thickness steps and mismatched joints	Flux concentrates at the step and pulls the arc sideways
Root gaps and open-root preparations	The gap gathers flux from both plates at the puddle
Seam starts, ends and run-on areas	The magnetic circuit changes abruptly where the joint ends

The arc settles on the path of least magnetic reluctance, the easiest magnetic route through the component. That is also why stick welding on DC suffers more than a TIG root pass on the same magnetised part: current density is higher and the arc has less room to recover.

The Ground Effect

Every DC arc carries a ground effect, and it drives much of the blow welders fight. The work clamp completes the circuit, so its position sets the direction of the returning current, and the arc is deflected generally away from the ground connection. The effect can add to other blow forces or cancel them, which is why moving the clamp sometimes reverses the problem rather than removing it.

Residual Magnetism: Why Some Joints Fight Back

Some steel arrives at the weld bay already magnetised, and the list of culprits is long:

• Magnetic lifting gear, magnets and lifting clamps
• Storage near transformers, motors and generators
• In-line magnetic flux leakage (MFL) inspection pigs, which leave permanent magnetisation in the pipe wall
• Plasma cutting, which can leave a magnetised edge in the heat-affected zone
• Mechanical stress and plastic deformation, leaving remanent magnetisation

Composition matters too. Low-alloy and high-strength grades hold magnetism more readily than mild steel, which matters if you run alloy steel welding electrodes on a magnetised joint.

Measurement is where people get caught out. A field checked in free air at the end of a pipe may read a few tens of gauss, yet once the joint is fitted up the root gap concentrates it and readings near 1000 gauss are possible. The Fabricator describes the same trap on pipeline work.

How Much Field Does It Take?

Published thresholds differ because the limit moves with geometry, current and joint design.

Source	Reported threshold
Wikipedia, arc blow entry	Begins at field densities as low as 10 gauss; severe at 40 gauss or higher
SSAB (Hardox technical support)	Disturbances at about 40 gauss or above; lower levels do not affect stability
TWI	Above roughly 50 gauss, arc blow may be experienced
Avdić, Pašić and Hasanagić (2014)	0–20 gauss: very low effect. 20–40 gauss: evident but removable by changing the return connection, current or electrode inclination. Above 40 gauss: demagnetise the component

Reconciling them is simple. Treat 10 gauss as the floor where deflection becomes visible, 40 gauss as the point where it turns into a serious production problem, and 50 gauss as a conservative shop rule of thumb. The 2014 study is the most useful because it splits the range into an action guide: below 20 gauss ignore it, between 20 and 40 fix it with technique, above 40 reach for the degausser.

DC vs AC, Polarity, and Multi-Electrode Welding

Arc blow is overwhelmingly a direct current problem. Nearly every field report of a wandering arc, a wandering puddle and a bead that will not hold shape traces back to a DC power source. Switch the same job to AC and the problem usually disappears, because AC never lets a one-way field build up.

Why AC Fixes It

On AC the arc extinguishes and re-establishes every half cycle: 100 times per second on 50 Hz mains, 120 on 60 Hz. Each time the current reverses, the self-induced magnetic field reverses with it and the sideways forces flip sign, so the one-way field that bends a DC arc never gets the chance to build. Repeated re-striking also re-centres the arc force on the work.

TWI’s guidance on magnetic arc blow lists running the root on AC rather than DC as a standard countermeasure. The catch is procedural, not physical.

Lincoln Electric makes the same point more emphatically for SMAW: with DC, especially above 250 amps, a change to AC may eliminate the problem entirely. That matters because the self-induced field grows with current, so arc blow gets worse at high amperage, right where pipeline root passes live.

DCEP vs DCEN: Why Polarity Matters

On DC you choose electrode positive (DCEP, reverse polarity) or electrode negative (DCEN, direct polarity). That flips the direction of the current, and therefore of the magnetic field it generates, so polarity decides which way a magnetised joint throws the arc. DCEP is often said to penetrate deeper, but the direction of the arc force and of the field-driven deflection both depend on polarity. Flip polarity and a forward blow can become a back blow and vice versa, which helps, because a forward blow is easier to counter with electrode angling.

The root pass is usually the worst moment. On a magnetised joint, the residual field in the plate and the self-induced field from the welding current combine in the narrow root gap, where flux concentrates and the arc has the least material to shield it. Any deflection throws heat and filler metal at an already thin edge, while a fillet pass has a fat, forgiving toe to hide behind.

Multi-Electrode Welding: Arcs That Interact

When two or more arcs share a workpiece they stop being independent. SSAB’s technical support material sets out the rule: with multiple electrodes on DC, equal polarity between the electrodes supports inward arc deflection, while unequal polarity supports outward deflection. Matched polarity on parallel runs pulls the arcs together; opposite polarity pushes them apart. Here magnetic interaction is a tool, not a nuisance.

The same logic explains the Scott connection used on double-arc AC rigs. A Scott connection phase-shifts the current of one arc by 80 to 90 degrees from the other, so one current and its magnetic field peaks while the other reaches its minimum. The fields partially cancel instead of reinforcing each other’s deflection, which greatly reduces arc blow on double-arc equipment. Twin-wire systems rely on that phasing for stability.

The Code Catch: AC Is a Procedure Change

TWI notes that some codes treat switching from DC to AC as a change of an essential variable, so the procedure must be re-qualified rather than adjusted on the fly. Under AWS D1.1, current type and polarity sit in the essential-variable list. If your qualified procedure says DCEP on E7018 and you switch mid-job to AC, you are no longer welding to that procedure.

Your options are to qualify the same electrode for AC, or qualify a revised WPS covering AC and the related parameter ranges. Neither happens in the time it takes to spot a wandering arc.

Practical Summary

• Arc blow is a DC problem first: AC cycles the field 100 to 120 times per second.
• Polarity steers the blow, and flipping it can turn a back blow into a workable forward blow.
• The root pass is the danger point, and above 250 amps on SMAW the problem gets worse.
• Multiple arcs interact: equal polarity deflects inward, unequal polarity outward, and the Scott connection offsets two AC arcs.
• Do not switch to AC on the spot: confirm the WPS allows it and that the electrode is AC rated.

On aluminium the current type question looks different: the polarity and filler choices in aluminium MIG welding and the machine settings in aluminium flux core welding carry their own DC-only requirements. The arc’s magnetic behaviour is the same, but consumables and gas are not interchangeable with steel.

The Defects Arc Blow Creates in the Weld Metal

Arc blow isn’t a problem with the base metal, it’s a problem with where the heat ends up. Once the arc starts to wander, the puddle, metal transfer, and fusion all fall out of control. The Australian Welding Institute lists magnetic arc blow as one of the primary causes of lack of fusion, right alongside joint preparation, welding parameters, and welder technique.

Undercut: The Most Visible Signature

Undercut is a groove melted into the base metal at the toe of the bead. It’s one of the first defects to appear when the arc pulls away from where it should be fusing.

With forward blow, the puddle is dragged ahead of the arc, starving the leading edge of heat and creating intermittent undercut along one side of the bead. With backward blow, the puddle piles up behind the arc, forming a narrow, high bead that still loses sidewall heat and leaves undercut in its wake. Regardless of direction, these grooves act as stress raisers and remove cross-sectional area. For many code work, the acceptable limits for undercut are strict, as noted by AWS CWI notes on undercut inspection. A small groove on a walkway can mean a reject on a pressure vessel or crane girder.

Lack of Fusion and Inter-Run Lack of Fusion

Lack of fusion occurs when the weld metal bridges over unmelted parent metal. If the arc is deflected to one side, an entire edge of the joint can go without heat and form a cold lap. When this happens on a root or filler pass and is then covered by later passes, it becomes inter-run lack of fusion.

This is the most deceptive form of the defect. From the outside, the bead can look perfectly sound until sectioned. For this reason, the Australian Welding Institute states that ultrasonic inspection is normally more effective than radiography for detecting inter-run lack of fusion, as it can locate and size the flaw better than a radiograph can for a thin, planar unfused boundary. When these flaws need to be removed, air arc welding and carbon arc gouging is the usual method for taking out a bad pass.

Changes in Penetration

Amperage and travel speed don’t guarantee consistent penetration when the arc is deflected. Forward blow pulls the heat ahead of the joint, often reducing root penetration. Backward blow can drive the arc force deeper into the puddle and increase penetration locally, however this comes at the cost of sidewall fusion and undercut. The operator can’t simply shift their hand to correct either of these outcomes, as the arc is no longer going where the electrode is pointed.

Porosity, Spatter, and Lost Deposition

A wandering arc agitates and freezes the puddle unpredictably. This erratic movement can draw in and trap shielding gas, leading to porosity. At the same time, control of droplet transfer is lost, causing a sharp rise in spatter. As the International Design Center’s welding arc zone reference points out, severe arc blow increases spatter and decreases deposition efficiency. That means more filler metal is lost, and less of it ends up in the joint where it’s needed.

Spotting and Repairing the Damage

The bead shape is often the first warning. Wavy width, irregular ripples, a narrow high bead with undercut, or a bead much wider than the procedure allows all point to a magnetic field taking over. If these appear, the field needs to be addressed before more passes are laid down.

Defect	What Arc Blow Does	Why It Matters
Undercut	Deflected arc starves the toe or sidewall of heat	Loss of section, stress raiser, and code rejection risk
Lack of fusion	Arc never reaches part of the joint surface	Unfused boundary with a poor load path
Inter-run lack of fusion	Blown pass is covered by subsequent passes	Hidden structural defect, expensive to find later
Reduced root penetration	Forward blow drags the puddle ahead	May fail to meet required penetration
Porosity	Erratic pool traps shielding gas	Voids, reduced strength, and crack initiation sites
Spatter	Loss of droplet transfer control	Wasted consumables, added rework, and increased exposure

Surface-breaking defects can be found with penetrant or magnetic particle inspection. Sub-surface flaws require radiography or ultrasonic inspection, with UT being the preferred choice for inter-run lack of fusion. Magnetic particle inspection is limited strictly to ferromagnetic materials and surface-breaking defects, so it cannot detect buried, inter-run flaws.

For repair, gouge or grind back to sound, clean metal, inspect the excavated surface, and re-weld to the WPS. However, if the magnetic condition causing the blow is still present, the repair pass will blow as well. Address the cause first, document the removal and re-weld range, and hold the repair to the same inspection standard as the original joint.

Where Arc Blow Lives: Joints, Jobsites and Materials That Trigger It

Arc blow is not random. It lands in predictable places, on predictable jobs, in predictable materials. Learn the map and you can spot trouble before the first arc is struck.

Pipeline repair after in-line inspection

The classic trap is a long-seam pipeline inspected with in-line magnetic flux leakage tools. Their permanent magnets saturate the pipe wall as they travel the line, and that magnetisation is a direct, common cause of arc blow on pipeline repair welds. It bites hardest where the torch is going: the root pass.

Residual magnetism is primarily an issue in tack welds and root runs. After those sequences the magnetism in the joint is normally low enough to reduce arc blow, according to SSAB’s technical support guide on magnetic arc blow. Get through the root and the tacks without chasing the arc around the circumference and the rest of the weld usually behaves.

The earth’s field adds more. A buried north-south line can pick up magnetisation over years underground. Weaker than the pig effect, but enough to deflect the arc, especially on a full-bore root pass where the welder is also fighting gravity.

Structural steel and the awkward corners

In structural work, arc blow concentrates where the magnetic path changes abruptly. Picture a column base plate: a fillet weld around a base, column tacked, crane still holding the piece. Plate edges are where flux bunches up, so the arc hunts toward an edge.

Box sections welded from the inside are recurring offenders. The arc sits in a steel box surrounded by steel in nearly every direction, and the return current takes whatever exit is convenient. The path changes constantly along the seam.

Stiffener corners and seam ends follow the same logic. Start and end are the worst moments, because the path changes abruptly there: at the start you establish one, at the end the arc breaks out of a settled pattern. Air arc welding and carbon arc gouging often cleans up those ends, but gouging a blown seam exposes more magnetised metal, so the root still has to be solved.

Mismatched thickness

When welding plates of different thicknesses, the arc may deflect toward the thicker plate in the joint, again per SSAB. The heavy section offers the path of least magnetic reluctance, so the arc leans into it: a root pass running hot and wide on the heavy side and starving the thin side, with lack of fusion on the thin plate and undercut on the heavy one.

Material families that fight back

Not all steels behave alike. TWI cites 9% Ni steel as particularly prone to high levels of residual magnetism, where de-magnetisation may be a final manufacturing operation. Those plates need careful handling, transport and storage so the magnetism does not come back before welding, which is why cryogenic LNG tank and pipework in 9% Ni often carries degaussing as a routine manufacturing step rather than a repair.

SSAB reports that sensitivity to magnetic arc blow caused by residual magnetism is generally more emphasised with higher-hardness and higher-strength grades. That matters as wear plate replaces mild steel in handling equipment and abrasion liners. Grades like Hardox, Strenx, Armox and Toolox are used more and more, and the harder and stronger the grade, the less forgiving it is of a wandering arc. That trend is precisely why arc blow complaints are rising across the abrasion-resistant plate industry.

Fixtures and workholding

Most shops overlook the biggest steel object in the room: the fixture. Lincoln Electric recommends a minimum of one inch of clearance between the supporting beam and the work for fixtures used on longitudinal cylinder seams, calls for nonmagnetic clamping fingers or bars, and warns against inserting a copper strip in a steel bar for backing, because the steel part increases arc blow.

The principle is simple: avoid large masses of steel on one side of a seam. Every kilogram nearby is another flux path competing with your joint.

One qualification from SSAB: fixturing effects are usually not a problem with stick-electrode welding when the current does not exceed 250 amps, but fixtures for higher currents and mechanised welding should be studied specifically. A 400 amp mechanised seam inside a magnetised jig makes the fixture a serious variable, not a guess.

Trigger reference table

Trigger	Why it matters	First thing to change
Post-MFL pig inspection	Magnets saturate the pipe wall; root pass sees the worst field	Measure the gap, plan to degauss
North-south buried line	Earth’s field leaves residual magnetisation over years	Check field before striking
Base plates and box sections	Edges and enclosed steel bunch flux, changing the path	Move ground clamp to the weld point
Start and end of a seam	Magnetic path changes abruptly	Use run-off tabs
Mismatched thickness	Arc leans to the thick plate, starves the thin side	Angle electrode, control heat
9% Ni steel	High residual magnetism; may need degaussing in manufacture	Control storage; degauss before welding
Hardox, Strenx, Armox, Toolox	Sensitivity rises with hardness and strength	Short arc, lowest practical current
Magnetised fixtures	Steel masses compete with the joint for flux	1 in. clearance, nonmagnetic fingers, no steel in backing bars

The pattern is the same everywhere. The magnetised steel is not in the joint, but it is in the way. That is good news, because it means the problem usually has a physical, checkable answer rather than a mysterious one.

Diagnosing Arc Blow: Reading the Bead and Measuring the Field

Arc blow is one of the few welding defects you can diagnose with minimal equipment. The bead tells you what the arc was doing; a gaussmeter tells you how much magnetism you have to fight. Run this sequence before anyone changes an electrode, raises the voltage or blames the welder.

Step 1: Read the Bead

Start with what came out of the last pass. Bead shape, width, undercut and penetration all track the direction the arc was deflected.

Signature	Forward Blow	Back Blow
Bead width	Wide, irregular width	Narrow, high bead, usually with undercut
Bead shape	Wavy, wandering	Stacked or humped, narrow crown
Undercut	Intermittent	Continuous or intermittent
Penetration	Decreased	Increased
Other	Arc force pulls the puddle ahead of the puddle	Spatter; on sheet metal, surface porosity at the finish end

Lincoln Electric makes a key distinction here: in an open-arc process the deflection is directly visible, because you can see the arc wander. Under submerged arc it is much harder to diagnose, and direction has to be inferred from the defect you end up with. When the coating peels away or the bead edge goes dark and irregular, you cannot watch the arc move, so you read the defect backwards to find the cause.

Worth memorising: welding past a current return connection bolted to the centre of the plate drags the bead sideways every time. That pattern, illustrated in the guidance from TWI and the Australian Welding Institute, is unmistakable once you have seen it. If the bead curves consistently at one point along the seam, look for a clamp, bolt, work lead attachment or heavy lump of steel under that spot.

Step 2: Measure the Field

Visual inspection tells you what happened. Measurement tells you how bad it is and where to weld.

A compass is the free first check. Hold it near the joint and watch the needle. If it swings when held away from the joint axis rather than along it, the workpiece is magnetised. Crude, but it finds problems in seconds.

For real numbers use a handheld gaussmeter with a blade-type Hall probe. The routine recommended by TWI: put the probe into the weld gap, move it around the circumference, note the maximum reading, then mark that position on the pipe with a paint pen or tape. The marked spot becomes your most hostile welding position and the one you plan around.

Pipeline repair follows the same logic. The procedure in Oil & Gas Journal requires measuring both the magnitude and the polarity of the residual magnetic field in the groove, then verifying the compensation with the probe still in the groove before striking an arc. Polarity matters because the return path has to oppose the field you measured, not reinforce it.

Step 3: Rule Out the Usual Suspects

Not every ugly bead is magnetic arc blow. Before moving clamps around, rule out the ordinary causes:

• Voltage too high. A long, wandering, oxidised arc is a voltage problem, not a magnetism one. Check arc length against amperage.
• Travel speed too fast. It leaves the puddle behind and drags the arc forward, mimicking forward blow.
• Current too high for the electrode diameter. Excess amperage lengthens the arc, raises spatter and can mimic back blow.
• Wrong electrode work angle. An electrode pushed too far forward drags the puddle and undercuts the leading edge.
• Poor joint preparation. A mis-machined bevel, an oversized root face or a gap that changes along the seam ruins the bead for ordinary reasons.
• Dirty or wet surfaces. Rust, paint, moisture, oil or water in the root gap cause porosity and undercut unrelated to flux.
• Incorrect polarity. Wrong polarity changes arc force, spatter and penetration. Confirm it against the WPS, not memory.

Once those are clear and the distortion still points in one consistent direction, you are looking at magnetic arc blow.

Step 4: Confirm and Document

If the metal is suspect, confirm the defect rather than argue about it. Magnetic particle inspection finds surface and near-surface cracking and lack of fusion in ferromagnetic material, and ultrasonic inspection is the standard tool for deeper inter-run lack of fusion in structural work, per AWS CWI guidance. Enclosed or underwater work on the same joint uses the same principles with the right probe setup.

Record what you found while it is fresh: joint ID, pass number, measured gauss value and location, direction the arc blew, corrective action taken. Two minutes now saves hours of gouging and re-welding later.

And if lack of fusion keeps returning in the same joint after you have degaussed the pipe, repositioned the work lead and confirmed the field is low, stop blaming magnetism. The procedure or the welder needs attention. Root pass work on all position welding rod has its own skill curve, and a welder fighting a normal joint in a normal position is a different problem.

Diagnose, measure, rule out, confirm.

Welder-Level Fixes: Techniques That Work on the Spot

Most arc blow can be corrected without buying anything. The key is to work at the torch in order of speed and cost. Start with the connection, then adjust your arc, then your travel, and only as a last resort reach for one of the trickier field fixes.

1. Move the Work Lead Before Anything Else

The position of your ground clamp is often the culprit. According to Lincoln Electric, it should always be the first thing checked. Moving the clamp closer to the puddle shortens the return path, limiting how far the arc can be pushed away from it.

If a single, close clamp isn’t possible for a long member, balance multiple earth connections along the joint or split the lead with a proper welding cable splitter. As TWI notes, multiple earth connections may solve the problem altogether. AWS CWI guidance echoes this, placing the work lead as close to the active weld as possible.

2. Short Arc, Lower Current, and an Angled Electrode

Once your ground is solid, adjust your arc itself. Start by holding as short an arc as you can. The arc force helps resist magnetic deflection, a trait that is one of the reasons SMAW still wins on site work where conditions are unpredictable.

Next, run the lowest practical current. The self-induced magnetic field scales with arc current, so dropping your amps weakens the force doing the pushing. You may need a smaller diameter electrode to hold the arc at this setting, so be sure to cross check it against your rod selection for the position and diameter without sacrificing the heat input required by your procedure.

Finally, angle your electrode into or against the direction the arc is blowing. SSAB recommends angling opposite to the blow to mechanically straighten the arc, whether it’s starting to wander or has already been pulled.

3. Direction, Sequencing, and Joint Preparation

Where you travel is just as impactful as how you set your arc. Welding toward the work connection tends to reduce forward blow, whilst welding away from it can reduce back blow. For MMA, TWI recommends welding toward the return connection, as a small amount of controlled back blow can even be desirable for certain passes.

To keep the field consistent throughout the seam, try the following:

• Back-step welding: Deposit short increments, re-establishing the arc in a more stable field each time.
• Weld to existing metal: Tacking to a heavy tack or an existing weld can help draw the arc back into line.
• Heavy tacks: Tack at both ends of the seam and frequently along loose fit-ups to give the arc a firm return path.
• Run-off tabs: Let the arc terminate in a magnetically neutral zone, rather than dying out in a corner where the field tends to pile up.

4. Wrap the Work Cable Around the Workpiece

For round sections and pipe, one free trick stands above the rest. Wrap the work cable around the component one or more times before striking an arc. The returning current creates its own opposing magnetic field that can cancel out a large portion of the offending one. The trade-off is mobility, as this compensation only holds where the wrap is, so it will need to be reset as you travel.

Fix	Why it works	What it won’t fix
Close work lead	Shortens the return path for the arc	A heavily magnetised joint by itself
Multiple earths or cable splitter	Shares current over a shorter, lower resistance route	Long, isolated members with only one practical clamp point
Short arc	Arc force resists magnetic deflection	Wide beads that require extra arc length to fuse
Lower current	Weakens the self-induced magnetic field	Heat input if travel speed isn’t matched
Angled electrode	Straightens the arc mechanically	A field that shifts rapidly as you move
Direction, back-step, tacks & tabs	Keeps the arc in a controllable, repeatable field	Seams that must only be welded one way
Cable wrap	Adds an opposing field around the workpiece	Anything outside of the wrapped area

Work through this list in order, and most cases will resolve before the electrode runs out. If none of these torch-level changes shift the bead, the magnetic field is stronger than technique can counter. At that point, the answer is to degauss the steel itself. Should a pass already need correction, you’ll want to have solid air arc gouging technique before going back over it.

Demagnetising Steel and Controlling Residual Magnetism

Technique fixes like a short arc, low current, electrode angle, back-stepping, or repositioning the work lead will solve most cases of arc blow. However, once those adjustments fail and the arc continues to wander, the residual magnetism is too strong for technique alone. At that point, demagnetising the steel itself is the shortest and most reliable path to a sound weld.

The Shop-Floor Coil: Welding Cable and a DC Source

The most accessible option for many shops is to use the welding machine they already have. By wrapping welding cable around the component and passing a large direct current through it, an opposing magnetic field can be induced to cancel out the trapped field in the steel.

As TWI explains, this is done by wrapping the cable around each end of a pipe, keeping the edge of the outermost coil roughly 60 mm from the joint line, and connecting the arrangement to a DC power source. Current should be increased gradually while watching a gaussmeter. If the measured field rises instead of falling, reverse the polarity and continue. That feedback is not a recommendation, but a necessity.

Research on gas pipeline commissioning by Avdić, Pašić and Hasanagić used a similar stepwise approach, winding a demagnetising coil around the pipe and raising the current in steps to around 140 A to reach a state where welding could proceed. Taking it slowly matters here, as rushing the current risks damaging any coatings or linings on the pipe.

The Fabricator outlines another field method that can double duty. By wrapping the work lead around the pipe three to six times, roughly half the pipe diameter back from the groove and on the side opposite to the pull of arc blow, then depositing several 3/4- to 1-inch tack welds around the joint, the cable wrap offsets the offending field whilst the tacks pin the joint in place.

Why Pulsed DC Beats AC for Big Components

Not all demagnetisers are equal, and this is where older AC “shotgun” coils often fall short. The limitation comes down to the skin effect. With alternating current, the flow concentrates near the surface of a conductor and drops off exponentially with depth. On a thick pipe wall, this only treats the outer layers, leaving trapped flux in the interior to continue causing arc blow.

Pulsed DC demagnetising equipment avoids this by driving a field through the entire thickness of the section with results that hold for the long term. Western Instruments tested this on a 10-3/4-inch OD pipe, demagnetising it with a pulsed DC unit and monitoring it for ten days with a Hall effect magnetometer. No measurable re-magnetisation was found through the full thickness of the pipe wall, a stark contrast to a purely surface treatment.

Compensating Coils in Pipeline Repair

For live pipeline repairs, demagnetising an entire component isn’t always practical. Instead, a compact compensating coil can be used. Rather than erasing the field everywhere, it is positioned near the weld zone to produce a counter field that opposes the residual magnetism exactly where the arc will run. With careful tuning of its position, turns, and current, the arc can behave as if the joint were never magnetised at all.

Equipment Worth Buying

Two purchases will cover the vast majority of jobs. First, a portable gaussmeter or blade-type Hall probe. It will pay for itself the first time it reveals the field in the gap is far higher than assumed. Second, either a dedicated demagnetising coil for your typical work, or a pulsed DC unit if you regularly handle pipe and heavy plate.

For high-strength or hardenable grades, electrode selection is just as crucial as degaussing, so it’s worth reviewing your alloy steel welding electrode selection and safety guide before starting the job.

Finally, never weld with the demagnetising circuit energised, and re-measure after the job. Grinding, handling, or magnet-crane lifts can reintroduce magnetisation even after a successful treatment. If a blown pass still needs to come out, air arc gouging and carbon arc gouging is the fastest way to strip back to sound metal for re-welding.

Common Mistakes, Code Requirements and What to Do After a Bad Pass

Arc blow is cheap to fix and expensive to ignore. The habits that keep it in your shop rarely have anything to do with magnetism. They are about time pressure, cost cutting and paperwork, the same habits that quietly ruin plenty of other things.

The mistakes that keep arc blow in your shop

• Blaming the electrode or the grinder. Swapping consumables because the bead looks wrong costs money and teaches nothing. A fresh rod will not straighten a magnetic field, and a new grinding wheel does not fix a wandering arc.
• Raising voltage instead of shortening the arc. A long, loose arc gives the magnetic field more room to push the arc off axis, and you lose the arc force that would hold it steady.
• Leaving the ground clamp on the far end of a long member. The returning current travels the whole length of the steel and deflects the arc away from that path. Move the clamp to the joint.
• Welding away from the clamp on a magnetised joint. The ground effect pushes the arc away from the connection, so travelling the other way compounds the problem.
• Switching to AC without checking the WPS. It may be the right fix technically and the wrong move contractually.
• Ignoring arc blow on a root pass that “looks fine”. Fusion with the sidewall is hardest at the root and easiest to lose, and buried inter-run lack of fusion is among the hardest defects in the trade to find later.
• Covering a blown pass without recording it or inspecting it. Once the next pass covers the defect, the evidence is gone. Surface-breaking undercut and porosity can only be found by penetrant or magnetic particle inspection while the surface is still clean and reachable.

Switching to AC is not a welder-level decision

AC largely solves magnetic arc blow because the arc extinguishes and re-establishes every half cycle, so the unidirectional field never has time to build. The temptation at the bay is obvious: the root pass stops wandering, so keep welding.

But AWS D1.1 treats current type and polarity as an essential variable, so the qualified range in your WPS covers the polarity you qualified on. Changing from DCEP to AC mid-job is a procedure change, not a technique tweak, and a welder cannot make that call alone. Your options are to qualify the same electrode on AC, or to revise the WPS with supporting PQRs where the change sits outside the qualified range.

Electrode classification matters too. Not every SMAW electrode is approved for AC, so all position rod selection needs checking against the manufacturer’s data sheet rather than assumption. The alloy steel electrode families common on pressure work are often DCEP-first designs, so the change is not always as simple as flipping a switch.

What a CWI wants to see in the record

When a pass goes bad, the report is not an apology. It is a short, factual trail that lets a reviewer reconstruct the decision. Expect the following:

1. Joint identification and the pass range affected, so the gouge and re-weld cover the right length of seam.
2. Nature and severity of the deflection: direction of blow, the defect type observed, and a field reading if one was taken.
3. Corrective action taken, such as clamp repositioning, reduced current, back-step sequence or degaussing.
4. Confirmation that welding resumed inside WPS parameter ranges, with the actual current, voltage and travel speed recorded.
5. Visual inspection result after repair, plus the disposition of any NDE.

If there is any doubt about porosity or lack of fusion in that region, get the NDE done before the joint is covered. Surface and near-surface indications are well served by visual and magnetic particle checks, while buried inter-run indications need radiography or ultrasonic testing. Plan the repair around the procedure, using air arc gouging to remove the affected metal rather than filling over it.

The safety bill nobody puts on the quote

Arc blow drives up spatter and rework, and both drive up fume exposure. That matters more than comfort, because welding fume health effects are governed largely by fume composition, as recorded in the NIOSH Pocket Guide entry for welding fumes.

OSHA regulates hexavalent chromium with a permissible exposure limit of 5 micrograms per cubic metre as an 8-hour time-weighted average. The same fact sheet notes that welding on stainless steel converts chromium into its hexavalent state during the welding process, and that surfaces should be cleaned of coatings before welding. Every gouge-and-re-weld cycle is another cycle of that exposure.

Ventilation is not optional in the closed cases. 29 CFR 1926.353 requires general mechanical or local exhaust ventilation whenever welding is performed in a confined space, and specifically calls out enclosed-space welding of zinc-bearing, lead-bearing, cadmium-bearing and chromium-bearing metals.

Conclusion: A Practical Arc-Blow Playbook for Your Shop

Arc blow bends the arc off the shortest path between electrode and workpiece, and everything else follows: the puddle moves, the bead changes shape, sidewall fusion is lost, undercut appears along the leading edge, porosity traps in the cap, and a routine root pass becomes a gouge-and-re-weld job on a pressure line.

It is not a mystery and not a verdict on the welder holding the stinger. Arc blow is a measurable field condition with a cause you can confirm in minutes using a compass or a gauss meter, and countermeasures that are mostly free.

Tier 1: Diagnose

Before changing anything, prove what you are dealing with. Watch the arc in the first few inches: if it leans forward, drifts sideways or hunts around the joint, that is magnetic deflection. In the bead, a wide irregular bead, waviness, intermittent undercut and reduced penetration point to forward blow, while spatter, a narrow high bead with undercut, increased penetration and surface porosity at the finish end point to back blow.

Then measure. A gauss meter or Hall probe is the standard instrument for checking magnetic flux density at the next weld pass, per SSAB’s technical guidance on magnetic arc blow. Put the probe in the root gap, walk it around the circumference, note the maximum reading and mark that spot on the pipe with a paint pen or tape. Free-air and in-gap readings are not the same number, so measure where you intend to weld. Without a meter, a compass at least shows the field direction relative to the joint axis.

Tier 2: Technique

This is the cheapest tier and often enough. Hold the shortest arc you can work with, because arc force works against the deflection. Run the lowest practical current, which often means a smaller diameter electrode and reduced travel speed to match. Angle the electrode deliberately, into or against the blow, so you steer the arc rather than chase it. SSAB’s angle-of-the-torch guidance covers the geometry: give the arc a shorter path than the magnetic field.

Sequencing matters too. Weld toward the ground connection to suppress forward blow and away from it to suppress back blow. Back-step in short increments to keep the puddle small and centred, and weld toward a heavy tack or an existing weld so the field points somewhere benign. Make solid tacks at both ends, tack often where fit-up is loose, and use run-off tabs so the arc ends in a magnetically neutral zone rather than dying in a corner, where the field changes most abruptly.

Tier 3: Electrical Setup

Now the shop-controlled variables. Positioning the current return so welding is always toward or away from the clamp, and using AC instead of DC in MMA welding, minimises distortion of the arc current magnetic field, as TWI explains. Move the clamp close to the weld, add a second earth connection, or split the work lead with a cable splitter. Wrapping the work cable around the workpiece so the returning current neutralises the field causing the blow is a recognised fix, documented by Lincoln Electric. Crude and cheap, worth trying before you escalate.

Tier 4: Change the Process

When technique is not enough, the options get bigger. An AC root pass, where your WPS permits it, largely removes the problem: the arc re-establishes every half cycle and the unidirectional field never builds. Pulsed DC demagnetisation handles residual magnetisation in heavy sections an AC degaussing coil cannot reach. When magnetism is high, the only fully effective means is often to demagnetise the steel before welding, as TWI frames it. A shop gauss meter and a demagnetising coil built from welding cable cost less than one gouged root pass and the schedule slip that follows.

Tier 5: Document and Verify

Close the loop. Record the joint ID, pass range, nature and severity of the deflection, the corrective action and evidence that welding resumed inside WPS parameter ranges. Get NDE on suspect regions before they are covered or painted, since lack of fusion and porosity become far harder to detect once buried. If you changed current type or polarity, confirm your qualification and WPS cover it: AWS treats current type and polarity as an essential variable in D1.1.

Three Things to Check Tomorrow Morning

A blown pass means gouging, re-welding, added fume exposure and a slipping schedule. Spatter and rework drive fume generation, and OSHA’s welding safety guidance covers the exposure controls that follow, including local exhaust ventilation for enclosed-space welding on metals of toxic significance.

1. Check the work lead position on every DC job. See where the clamp actually sits relative to the live arc, and move it close to the weld if it is not.
2. Probe the next joint before striking an arc on it. One minute with a gauss meter in the root gap tells you whether you are walking into a magnetised joint, and where.
3. Confirm that arc-length discipline is taught and enforced. A long, lazy arc is the most common amplifier of arc blow and the easiest habit to fix on the spot. For a refresher on electrode handling and positioning, our all-position welding rod guide covers the setup basics for the consumables you are probably using.

Treat arc blow as a solvable field condition, not a mystery and not a welder failing. Diagnose it, tier up through technique, electrical setup and process change until the bead lands where you intended, then write it down so the next shift inherits a fix instead of a scar.

Frequently Asked Questions

What is arc blow in welding and what does it do to the weld?

Arc blow is the deflection of the welding arc away from the shortest path between the electrode and the workpiece, usually caused by an unbalanced magnetic field around the arc. It moves the puddle off the joint, distorts the bead, and produces defects such as undercut, lack of fusion, lack of root penetration, porosity and excessive spatter. In structural and pressure applications these are stress concentrators that can force gouging and re-welding.

Why does arc blow happen mostly with direct current welding?

DC current produces a steady, unidirectional self-induced magnetic field around the electrode that interacts with fields in the workpiece and any residual magnetism. An AC arc extinguishes and re-establishes every half cycle, roughly 100 to 120 times per second on 50/60 Hz power, so the field never builds enough to pull the arc off line. That is why switching the root pass to AC is one of the most reliable remedies, when the procedure permits it.

How do I tell if the arc is blowing or if my technique is wrong?

Forward blow gives a wide, irregular or wavy bead, intermittent undercut and reduced penetration; back blow gives spatter, continuous or intermittent undercut, a narrow high bead and increased penetration. Check the field with a compass or gauss meter in the root gap, and rule out high voltage, excessive travel speed, wrong electrode angle and poor joint preparation. If lack of fusion continues and it is not magnetic arc blow, the procedure or the welder needs reviewing.

What is the fastest fix for arc blow on a long structural member?

Move the work lead close to the active weld point, or clamp at both ends of the member and use a welding cable splitter. The position of the ground connection sets the direction of the returning current and therefore the direction of the blow. Shortening the arc, lowering the current, and angling the electrode against the deflection are good supporting habits.

How much residual magnetism is too much to weld?

Published thresholds vary with how and where you measure, but disturbance becomes noticeable around 20 to 40 gauss and severe at 40 gauss and above, with a common rule of thumb that problems appear above roughly 50 gauss. In practice the value matters most in the root gap, where it can be many times higher than a free-air reading. Many pipeline crews target below 30 gauss in the groove, and some specify 5 to 7 gauss.

Does using AC on the root pass require a new qualification?

It very often does. Current type and polarity is treated as an essential variable in AWS D1.1, so a welder cannot unilaterally switch from DCEP to AC without WPS authorisation. Typically you either qualify the same electrode on AC or revise the procedure with PQR support, and you must confirm the electrode classification actually performs on AC before welding.

Can arc blow be eliminated completely?

No, and it does not need to be. In practice it is usually reduced to a level where the bead is stable and within the procedure. Combining short arc length, the lowest practical current, a sensible electrode angle, correct work lead placement and, where necessary, demagnetisation handles the overwhelming majority of cases.

Does arc blow affect mechanical properties and code compliance?

Yes. Lack of fusion, undercut and porosity reduce the load-carrying area and act as stress concentrators, and inter-run lack of fusion is a serious structural defect that can be missed once the pass is covered. A blown pass that causes the welder to stop or change technique should be recorded in the production weld log, with the corrective action and the visual or NDE result documented.

Does arc blow increase safety risk for the welder?

Indirectly, yes. Arc blow raises spatter and forces gouging and re-welding, both of which increase fume and spatter exposure and rework time. OSHA regulates hexavalent chromium with a permissible exposure limit of 5 micrograms per cubic metre as an 8-hour time-weighted average and requires local exhaust ventilation for many enclosed-space welding operations, so the extra time and rework should be factored into your fume control.