Submerged Arc Welding (SAW): The Complete Guide to Process, Parameters, Defects & Productivity

A fabricator in Ohio once told me his shop rejected an entire 40-foot pressure vessel shell seam because of centerline porosity found only on radiography — a defect that traceable, in hindsight, to one bag of flux left open overnight in a humid shop. Submerged arc welding looks deceptively calm: no visible arc, no sparks, just a carriage crawling along a seam under a mound of gray granules. That quiet is exactly why SAW either rewards disciplined process control or punishes sloppy habits harder than almost any other welding process.

What Is Submerged Arc Welding? A Plain-English Introduction

If you have ever watched a SAW station at a heavy fabrication shop, the first thing you notice is what is missing. No crackle, no sparks, no blinding arc light spilling across the floor. A wire feeder hums, a carriage crawls along a seam, and a quiet mound of gray granules trails behind a contact tip. Somewhere under that pile, an arc is burning at thousands of degrees. That contrast between calm surface and violent reality is the whole identity of submerged arc welding.

Strip away the mystery and the definition is straightforward. The AWS Welding Handbook Chapter WHC2.06 defines the process as producing coalescence by heating with an arc or arcs between one or more bare metal electrodes and the workpieces, with the arc and molten metal submerged in a blanket of granular fusible flux. No pressure is applied, and the filler metal comes from the electrode, sometimes supplemented by powdered metals or strip.

Three phrases do most of the work there: bare metal electrode, granular fusible flux, and no pressure. Unlike stick welding, the wire has no coating to shield the puddle; the shielding comes entirely from the flux bed. Unlike forge or resistance processes, nothing squeezes the joint. It is fusion welding in its most literal form.

Why You Cannot See the Arc

The arc is invisible because it is buried. Before the weld starts, a hopper lays a layer of granular flux ahead of the electrode. When the arc strikes, it melts the flux in front of and around it, forming a pool of electrically conductive molten slag that carries the current, plus a shielding gas that keeps oxygen and nitrogen out of the weld metal. The operator sees only flux. The arc, the droplet transfer, and the molten puddle all happen underneath.

That conductive slag is why SAW is often described as a self-shielding process with a built-in current path, and it is why flux chemistry is never just a shielding decision. Flux influences arc stability, penetration depth, bead width, deoxidation, alloying, and how much current the process can carry before the arc turns unstable. We will get into that chemistry later in this guide, but the key idea is simple: in SAW, the flux is a consumable as important as the wire.

Once the bead solidifies, the unused granules behind the weld are still unfused. A vacuum recovery unit sucks them up, screens out spatter and slag chips, and returns them to the hopper. Reused flux loses some fine powder and picks up contamination, so shops routinely top off with fresh material and screen aggressively. That recovery loop is one reason SAW consumable costs per pound of deposited metal are so low.

How Big Is This Process, Really?

Bigger than most beginners assume. Research published in the Journal of Materials Engineering and Performance estimates SAW alone accounts for roughly 10% of all deposited weld metal worldwide. The productivity numbers explain why: deposition rates approaching 100 lb/h (45 kg/h) have been reported for SAW, against a maximum of about 10 lb/h (5 kg/h) for shielded metal arc welding. That is a tenfold difference in raw metal placement, which is why you find SAW on pressure vessel shells, wind tower sections, ship plates, storage tanks, and heavy structural work.

The tradeoff is position. SAW is normally limited to flat (1G/1F) and horizontal (2F) work, though 2G has been performed with special flux support arrangements. Gravity is quietly holding that blanket over the puddle, and tilting the joint fights the process. If your work is vertical or overhead, stick, MIG, or flux-core is the honest answer. Curious how those processes stack up position by position? Our overhead stick welding guide covers what SAW simply cannot do.

Why SAW Is Almost Always Mechanized

There is no way to hand-hold this process. You cannot see the arc, judge puddle color, or manually manipulate an electrode you cannot watch. So SAW equipment is built around a carriage, gantry, or manipulator that holds the torch, a wire feeder pushing bare wire at a steady rate, and a flux hopper that maintains bed depth. Typical operating currents run from 200 to 1500 amps, with up to 5000 amps used in multiple-electrode configurations. Those are not hand-tool numbers.

That automation cuts both ways. Setup takes planning, and parameter discipline matters more than manual dexterity. But once dialed in, the process runs repeatable beads for hours without anyone touching a control, which is exactly what long-seam production demands.

Is SAW Right for Your Thick-Plate, Long-Seam Work?

Ask three questions. Is the joint mostly flat or horizontal? Is the seam long enough to justify fixture and carriage setup? Does the code or drawing permit a mechanized process with qualified procedure paperwork? Three yes answers mean SAW is probably worth the learning curve. Any no means you are better served by a manual or semiautomatic process, starting with fundamentals like how to set up a MIG welder or understanding what electrodes in welding actually are.

For everything in between, the rest of this guide walks through the process step by step: how flux, arc, and slag interact, what hardware you need, which parameters really control quality, how to read consumable classifications under AWS A5.17, where SAW beats SMAW and GMAW, the defects that show up on radiography, what OSHA and IARC expect from you, and how to write a procedure that passes qualification on the first attempt.

How the SAW Process Works: Flux, Arc, Slag, and the Recovered Flux Cycle

To understand submerged arc welding, you have to picture one continuous, self-feeding loop. Nothing about it is mysterious, but every stage depends on the one before it. Get any link in the chain wrong and the defect shows up 40 feet later on a radiograph. Let’s walk through a live SAW pass the way it actually happens on the machine.

Step 1: The Wire Feeder Pushes Bare Wire Through the Contact Tube

It starts with a spool of bare, deoxidized wire — typically 3/32 or 1/8 inch for carbon steel work — driven at a constant speed by a motorized feed system. The wire passes through a copper contact tube that transfers welding current into it. Unlike stick welding, where you consume a coated electrode a few inches at a time, SAW feeds wire continuously, so arc-off time drops to nearly zero. If you have ever dialed in a MIG machine, the wire-drive mechanics will feel familiar, as covered in our MIG welder setup guide — but the electrical behavior under the flux is very different. The wire has no flux coating of its own; everything shielding and alloying comes from the granular bed above it.

Step 2: The Flux Hopper Lays a Granular Bed Ahead of the Arc

Directly ahead of the moving wire, a hopper gravity-feeds granular fusible flux onto the joint, burying the arc completely. That blanket is why you never see the arc, never get UV glare, and never see sparks. The recommended flux depth over the arc is generally 25 to 50 mm (about 1 to 2 inches) — deep enough to fully suppress the arc flash and contain the shielding atmosphere, but not so deep that un-melted granules choke the bead. Electrode stick-out, or contact-tip-to-work distance, typically lives in that same 25–50 mm band for stable operation.

Step 3: The Arc Melts the Wire, the Plate, and the Flux

Here is the part many beginners miss: when flux melts, it becomes electrically conductive. The molten slag carries actual welding current between the electrode and the workpiece, which is what stabilizes an arc you cannot see. A layer of shielding gas forms above the molten pool, excluding oxygen and nitrogen from the atmosphere, while the molten flux pool floats impurities up as slag. Together these give SAW its exceptionally smooth, quiet, spatter-free character.

Step 4: The Slag Lifts Off a Glassy Bead

As the puddle solidifies, the heavier weld metal sinks and the lighter slag floats and freezes on top as a glassy crust. Because the crust is continuous and brittle, it usually pops off in long ribbons as the bead cools. A clean, uniform slag peel is one of the clearest field indicators that your parameters were right. Ragged, sticky, or blistering slag usually points to flux contamination, wrong voltage, or inadequate current for the wire size.

Step 5: The Vacuum Recovery Cycle

Not all the flux melts. Unused granules lying beside the bead are picked up by a vacuum recovery hose, screened to remove fines and spatter, and returned to the hopper for reuse. This closed loop is one of SAW’s biggest economic advantages — consumable cost per foot drops dramatically compared to coated stick electrodes, such as the E7018 stick electrodes you would otherwise burn by the hundredweight.

But recovered flux has one non-negotiable requirement: it must be sieved. Over time the granules degrade in particle size and pick up metallic fines from the process. Industry troubleshooting guides consistently link degraded, un-sieved flux directly to porosity — the same centerline porosity that condemns pressure vessel seams. If you remember one rule from this section, make it this: screen your recovered flux, every time.

Active Flux vs. Neutral Flux

Flux chemistry drives metallurgy, and the first split to understand is active versus neutral:

• Active fluxes deliberately contribute manganese and silicon to the weld pool as deoxidizers and alloy additions. They are economical and effective for single- and two-run welds on clean plate, where each pass is fresh metal.
• Neutral fluxes add essentially no alloy content, making them the right choice for multi-pass welds, where repeated passes through an active flux could over-alloy the metal and embrittle it.

The second concept is the basicity index, a ratio describing the flux’s chemical basicity. Miller’s SAW productivity guidance recommends selecting the lowest basicity index that still meets your required mechanical properties. Low-basicity fluxes run more forgivingly at high current and fast travel speeds, while high-basicity fluxes generally deliver better toughness. It is a genuine engineering tradeoff, not a “buy the expensive one” decision.

Flux also performs five simultaneous jobs: atmospheric shielding, arc stabilization, deoxidation of the pool, alloying where required, and bead shape control — that last one is why flux manufacturers tune slag viscosity so tightly. For the full classification picture behind these choices, see our breakdown of welding electrode types and numbering.

Get the flux cycle right — dry storage, correct depth, sieved recovery — and most of your porosity and contamination problems never start. As AWS technical resources on the process emphasize, flux condition is not a consumable detail; it is a primary process variable, on par with current and voltage (AWS Publishing, Canadian Metalworking).

SAW Equipment: Power Sources, Wire Feeders, Torches, and Automation Systems

A submerged arc welding system only works when its parts pull in the same direction. The power source, wire drive, flux handling gear, travel mechanism, and control system all rely on one another. If one of them falls out of sync, you can still lay down a bead, but the results can quickly fall short of what the job demands.

Power Sources: DC, AC, and Variable-Balance Square Wave

The power source sets the ceiling for the entire setup. For most SAW applications, polarity is the first decision to lock in:

DC arc blow is the main reason many shops switch to AC. On long vessel seams or heavily magnetized workpieces, the arc can wander toward the edges or crater. AC helps cancel that pull. As noted in AWS Welding Digest’s practical guide to submerged arc welding, the biggest leap forward is the variable-balance square-wave AC machine. It lets operators dial a split, such as 80% DCEP and 20% DCEN, to pull penetration from the positive half and deposition from the negative half without swapping leads.

Wire Feeders, Electrodes, and Flux Delivery

A reliable wire feeder keeps electrode feed speed steady under heavy load. For submerged arc welding, three filler types cover most work: solid wire (commonly EM12K, EM14K, or EM15K), metal-cored wire (faster deposition with better wet-out and fit-up tolerance), and composite electrodes (high deposition with controlled penetration). Sizes run from 5/64 in. up to 7/32 in. (about 2 to 5.6 mm). For a familiar middle ground, 3/32 in. wire sees frequent use. For a breakdown of filler designations, see our guide to what electrodes are in welding.

Flux is fed from a hopper ahead of the arc to form a blanket that fully covers it. Behind the weld, a vacuum recovery nozzle collects unconsumed granules, a screening station removes slag and spatter, and clean flux returns to the hopper. Dry storage and routine screening go a long way toward preventing porosity. As industry tech tips point out, those habits also stretch your flux spend over many passes.

Torches, Travel, and Connectivity

The torch and travel mechanism position the electrode and maintain consistent stick-out and offset. Options range from self-propelled tractor carriages and track systems to column-and-boom manipulators, gantries, and turning rolls. Handheld semi-automatic SAW guns exist, but flux coverage is hard to guarantee without a rigid setup. On any platform running sustained high current, cable connectors and gun leads must stay tight and properly sized to avoid overheating. The constant-voltage logic that drives these feeders shares much with other wire processes, which is covered in our MIG welder setup guide.

Automation, Seam Tracking, and Programmable Controls

With long seams, repeatability matters more than speed alone. Most cells use one of these seam-tracking approaches:

• Mechanical guide: A shoe or roller rides the groove. Rugged and simple, but blind to fit-up changes.
• Touch or photoelectric probe: Follows a prepared edge or marked line with minimal complexity.
• Optical or laser vision: Images the joint ahead of the torch and corrects laterally in real time for shifting fit-up.

These systems feed into a programmable controller that stores voltage, wire speed, travel speed, and torch offset for each pass in a multi-pass sequence. Miller’s productivity research consistently shows that arc-on time and parameter consistency, not raw amperage, drive cost per foot.

Twin-Wire vs. Tandem: Two Approaches to Higher Deposition

Twin-wire is best viewed as a torch upgrade that uses your existing power source to push two wires into one pool. Tandem adds a second circuit, but lets you tune the lead and trail arcs independently for maximum output. Either way, those gains only hold if your hopper, recovery loop, carriage, and controls can keep pace with the extra metal going down.

SAW Welding Parameters: The Variables That Decide Your Weld Quality

Watch a submerged arc carriage run for ten minutes and it looks like autopilot: no visible bead, a steady hum, and a glassy-smooth weld emerging under a lifting crust of slag. Then a radiograph finds lack of fusion three feet in. Quality in SAW is decided before the arc strikes, by whoever set the dials, because a mechanized head will repeat a mistake perfectly for four hundred feet.

Wire Feed Speed: The Dial That Really Sets Your Amperage

On a constant-voltage SAW power source you set voltage and travel speed, and the machine answers: wire feed speed is the primary factor controlling welding current. Feed faster and the arc shortens and pulls more amps; slow the feed and current drops. It is the same relationship you learn when you set up a MIG welder, scaled up to wire diameters manual GMAW never sees. Practically, current is not dialed in, it is arrived at. If amps drift mid-seam, suspect wire feed friction, spooling, or contact tip condition before blaming the power source.

Arc Voltage: Bead Width, Profile, and Flux Melting

Voltage sets arc length under the flux and therefore how wide the pool and slag spread. Raise it and the bead gets wider and flatter, with sidewall fusion improving up to a point. Drop it too far and you get a narrow, high-crowned bead with heavy sidewall bite that can tear at the toes. Voltage also interacts with electrode position: a contact tube angled forward or offset behind the arc shifts the effective reading without you touching the dial, which is why seam-tracker drift appears as mysterious bead-width changes on long seams.

Travel Speed: Heat Input Per Inch and the Gas-Trapping Trap

Travel speed is where productivity and metallurgy collide. Slow down and penetration rises, but so does heat input per inch, widening the heat-affected zone and costing toughness. Speed up and you cut heat input, but leave less time for molten slag to release trapped shielding gas and unreacted flux, risking lack of fusion.

Travel speed and heat input are not locked together, though. A 2024 study in The International Journal of Advanced Manufacturing Technology raised travel speed from 320 to 768 mm/min while holding heat input constant and cut active welding time by roughly 61% with no change to HAZ microstructure or hardness. The lever was more current and voltage at higher speed, which is the whole secret to improving productivity in SAW applications: run hotter and faster, not hotter and slower.

Electrode Stick-Out (ESO or CTTD)

Electrode stick-out, more precisely the contact-tip-to-work distance, is the free wire between the contact tube and the arc. Longer stick-out adds resistive I²R heating before the wire reaches the arc, which raises deposition but slightly reduces penetration at the same wire feed speed, and makes the arc fussier about joint geometry. Most setups target roughly the wire diameter plus a small constant, and the value belongs on your WPS as an essential variable, not on whoever loads the flux hopper that morning.

Polarity, Current Type, and Wire Diameter

Polarity is a design decision, not a preference. DC electrode positive (DCEP/DCRP) gives deeper penetration at lower deposition efficiency, while DCEN gives shallower penetration and much higher deposition, ideal for buildup and surfacing. AC balances the two and cuts arc blow on heavy magnetic assemblies, which is why AC and variable-balance square-wave units dominate thick multipass work, as our guide to electrodes, polarity, and selection explains. Wire diameter matters too: 3/32 in. (2.4 mm) is the plate-fabrication workhorse, 1/8 in. and larger for twin-wire heads and heavy buildup, finer wire for root passes and thin sections.

Flux Depth, Flux Type, and Joint Design

The flux blanket must cover the arc completely, typically one to 1.5 inches deep ahead of and around the electrode. Too shallow gives arc flare, spatter, and porosity; too deep hinders gas escape and encourages inclusions. Chemistry matters too: active fluxes containing metal oxides raise current-carrying capacity and penetration but narrow the bead, so identical dial settings behave differently than on a neutral flux. Joint design closes the loop: single-V, double-V, U-groove, and narrow-gap joints each demand different electrode angles, heat distribution, and root faces, and a right joint often removes an entire pass, which is cheaper than any later tuning.

A Practical Starting Point: ½-Inch A36 Plate

Portland Community College’s WLD 290 training program gives students a representative flat-butt setup worth borrowing as a sanity check:

• Electrode: EM14K (L61) carbon steel wire, 3/32 in. diameter
• Flux: F7A2 classification, buried to full coverage
• Current/polarity: ~500 A, DCRP (DCEP)
• Voltage: ~35 V
• Travel speed: 24 ipm
• Base metal: ½-inch A36 plate, flat position, butt joint

Do not treat those numbers as universal; use them to check whatever your procedure calls for. Note that F7A2 flux paired with a matching electrode defines a specific tensile and impact envelope, and swapping either component silently invalidates the combination. Consumable classification codes are explained in the next section.

Heat Input, Carbon Equivalent, and Preheat

Every variable above feeds one governing number: heat input = (Voltage × Amperage × 60 × efficiency) ÷ travel speed in mm/min, with SAW assigned an efficiency near 0.95 because the buried arc loses very little heat. Using the example above at 500 A, 35 V and 24 ipm (about 610 mm/min), you land near 1.7 kJ/mm (roughly 43 kJ/in). Combined with the base metal’s carbon equivalent, that drives preheat under AWS D1.1: the IIW formula CE = C + Mn/6 + (Cr + Mo + V)/5 + (Ni + Cu)/15 means higher CE and thicker sections push you toward preheat to avoid hydrogen-induced cold cracking.

D1.1 requires minimum preheat at a distance from the joint at least equal to the thickness of the thickest member, but never less than 3 inches in all directions from the point of welding. And when notch toughness is a design requirement, the AWS position statement issued after the Northridge Earthquake recommends that interpass temperature not exceed 550°F (about 300°C), so the thermal cycles of successive passes do not coarsen the microstructure and destroy the toughness you selected the consumables for. This same interplay runs through welding hardened and hardenable steels, applying to SAW exactly as to manual processes, only with bigger numbers.

SAW Consumables: Electrodes and Fluxes per AWS A5.17 and A5.23

In submerged arc welding, the consumables aren’t really consumables in the ordinary sense — they are the process. The arc is buried, so you can’t watch a puddle and adjust on the fly. Whatever the wire and flux do together is what the radiograph will show six weeks later. That is why choosing the electrode and flux is the first real engineering decision on a SAW job, not a purchasing afterthought.

What AWS A5.17/A5.17M Actually Classifies

For carbon steel work in the United States, the governing document is AWS A5.17/A5.17M, Specification for Carbon Steel Electrodes and Fluxes for Submerged Arc Welding, available through the AWS publications catalog and issued in its current A5.17/A5.17M:2025 revision. The spec covers both solid and composite (metal-cored) carbon steel electrodes, plus the granular fluxes they are qualified with.

The classification system runs on two independent tracks, and confusing them is the most common spec-reading mistake in a fabrication shop:

• Electrodes are classified by chemical composition — the manganese, silicon, carbon and deoxidizer package actually in the wire.
• Fluxes are classified by mechanical properties — the tension, impact and soundness of the weld metal that flux produces when used with a classified electrode.

So a flux class number is a promise about a finished weld, never about a bag of gray granules in isolation. If you change the wire, the flux’s published properties no longer legally apply. The same logic applies to stick electrodes, where the AWS number tells you chemistry and positioning, as covered in What Are Electrodes in Welding? Types, Numbers, Polarity & How to Choose.

Reading Common Electrode Designations

The letters and digits you will see on drawings and certification documents follow a predictable pattern. The leading E means electrode, the following letters and numbers describe the wire chemistry, and a suffix such as K indicates a potassium-bearing deoxidizer that promotes a finer-grain, tougher deposit. Widely used carbon steel designations include:

• EM12K — a general-purpose, easily formed bead with good out-of-position tolerance for a SAW wire; the default root and fill choice on many plate jobs.
• EM14K and EM15K — higher manganese/silicon tiers used where more strength, faster travel, or heavier deposition is needed.
• EA1 and EA2 — alternate designations you will still find called out on drawings and in older procedure records; always verify the exact chemistry against the current spec table rather than assuming equivalence.

Composite (metal-cored) electrodes follow the same framework but trade pure chemistry control for dramatically higher deposition rates, which we look at again when we get to Flux Core Wire Welding Tips.

Diffusible Hydrogen: H4, H8 and H16

A5.17 requires more than a chemistry check. Mandatory testing includes tension tests, radiography of the deposited weld metal, Charpy V-notch impact testing, and diffusible hydrogen determination performed by the method in AWS A4.3. Hydrogen results are reported as H4, H8 or H16, meaning the maximum millilitres of diffusible hydrogen per 100 grams of deposited weld metal — 4 mL, 8 mL or 16 mL respectively. On high-restraint, high-carbon-equivalent plate, the H4 designation is cheap insurance against delayed cracking.

Matching Flux to Electrode for Tensile and Impact Requirements

Procurement happens in matched sets, written in the form F7A2-EM14K: the flux class first, the electrode it was tested with second. The number and letter combination behind the F encodes the impact tier and test temperature, so the notation is meaningless without its wire. When your procedure calls for 20 ft-lb at −20 °F, you buy the exact combination that was tested to it, not a flux from one supplier and a wire from another because both were on the truck. Impact requirements tighten further at low-temperature service, so confirm the qualified combination against the actual service temperature before cutting a PO.

A5.23 for Low-Alloy Steel and Two-Run Work

Carbon steel is only half the story. For chromium-molybdenum, nickel, manganese-molybdenum and other alloyed base metals, the correct document is AWS A5.23/A5.23M, Specification for Fluxes and Electrodes for Submerged Arc Welding of Low-Alloy Steel. Two-run flux-electrode classifications — engineered so a single pass on each side of a joint produces a sound, properly fused weld — are addressed in A5.23 rather than A5.17. Ordering a two-run flux under A5.17 will get you a material review flag at best.

Where the Specification Stops

A5.17 is deliberately quiet about safety. The specification explicitly directs users to ANSI Z49.1, Safety in Welding, Cutting and Allied Processes, and to AWS Safety and Health Fact Sheets, noting that safety issues fall outside the scope of a filler metal specification. In practice that means the flux SDS, ventilation plan and protective equipment decisions belong to your shop, starting with the baseline requirements in What Does a Welder Wear? The Complete Welding PPE Guide, alongside OSHA’s welding, cutting and brazing fact sheet and 29 CFR 1910.252.

Get the consumable pairing right, and everything downstream — parameters, procedure qualification, defect rates — gets easier. Get it wrong, and no amount of seam tracking will save the weld.

SAW vs. Other Welding Processes: When SAW Wins and When It Doesn’t

Ask any shop foreman which process they’d pick for a 60-foot seam on half-inch plate, and the honest answer depends on where that seam sits. Submerged arc welding is not a general-purpose process. It dominates a narrow lane and gets beaten badly outside it, and knowing where that lane starts is the difference between a productive investment and an expensive paperweight.

Deposition Rate and Operating Factor: The Productivity Math

SAW’s headline strength is raw metal deposition. A bare continuous electrode carries current along its full length while the arc runs buried in flux, which gives SAW high deposition rates, deep penetration, and readily sound welds with minimal fume or arc light. In mechanized setups that also means a high operating factor: the arc stays lit, the carriage keeps moving, and nobody stops to change a stick or reposition a stinger.

Manual processes lose arc time in other ways. SMAW burns one rod at a time and every swap costs arc time, and dialing 3/32 E7018 amperage correctly helps, but it cannot fix the arc-off penalty baked into stick welding. GMAW and FCAW improve the operating factor by feeding wire continuously, yet a human hand still guides the torch. SAW hands travel to a carriage. On thin sheet it can also reach genuinely startling speeds, with published figures above 100 in/min (2.5 m/min) where the joint and fixturing allow it.

Weld Quality and Consistency

Once a SAW procedure is dialed in, it repeats nearly identical welds pass after pass, which is why pipe welding procedures and pressure vessel work lean on it. Manual processes depend on the operator’s steadiness, so consistency drifts over a long shift. The tradeoff is that SAW has almost no ability to correct itself. It welds exactly what the parameters and joint preparation tell it to, which is why SAW quality problems tend to show up in groups rather than one-offs.

The Flat-Position Tradeoff

This is the biggest limitation, and it is non-negotiable. SAW is essentially restricted to flat and horizontal work because granular flux must sit on top of the joint and stay there; it has nowhere to go in vertical or overhead positions. Thatmeans the workpiece usually has to be positioned, rotated, or fixtured so the weld sits flat, which demands positioning equipment and extensive setup time. The arc is also invisible under the flux, so the operator cannot watch the puddle and react the way a MIG or stick welder can.

Setup Time and Cost Structure

SAW’s economics reward volume. Fit-up and parameter tuning can eat hours before the first arc strikes, so a 12-inch weld rarely justifies the effort. Run that same setup across hundreds of feet and the cost per inch collapses. Manual processes invert this: cheap to start, expensive per foot. A shop comparing a MIG setup against a SAW station has to weigh total seam length, not hourly rate.

Material Limitations

SAW lives in the ferrous world. It handles carbon steels, stainless steels, and nickel alloys well, but it is not the go-to for aluminum the way GMAW is. Flux chemistry is tuned for specific steel families, so exotic alloys usually push you back toward GTAW or pulsed GMAW.

Electroslag welding and SAW get confused often, but ESW welds in the vertical position and relies on resistance heating through molten slag rather than a true arc. Narrow-gap SAW shrinks the joint groove to cut fill volume on very thick steel, trading some accessibility for major consumable savings.

Where SAW Wins

SAW dominates high-volume applications where flat positioning and long seams align: shipbuilding, oil and gas fabrication, structural steel, tube and pipeline joints, pressure vessels, nuclear containers, wind tower shells and flanges, and oil, water, or LNG tank construction. On a wind tower shell or a storage tank course, the process turns a joint that would take days into hours.

One unusual regulatory perk reflects how tightly the arc is buried. OSHA 29 CFR 1910.252(a)(3) exempts submerged arc welding from the helmet and hand-shield requirement that applies to other arc welding, because the arc is fully covered by flux. That does not remove other PPE needs, and a proper welding PPE setup still applies for burns, slag, and radiant heat.

Note: The right question is not “is SAW better?” It is “does this job have long, flat, thick seams repeated many times?” If yes, SAW usually wins on cost and quality. If the answer is no, a flux core setup or manual process will beat it every time.

Common SAW Defects and How to Fix Them: Porosity, Lack of Fusion, Slag Inclusions, Cracking

One of the quiet risks of submerged arc welding is that the process hides its own evidence. There is no spatter to read, no puddle color to judge, and no visible arc to tell you the parameters have drifted. Defects stay buried under a gray mound of flux until someone chips slag or a radiograph comes back with a row of white dots. That is why SAW troubleshooting is a skill worth building before the first production run, not after the repair bill arrives.

Porosity: The Defect With the Longest Suspect List

Porosity tops every SAW defect list because it has the most causes. An analysis published through the AWS Welding Journal identified eleven distinct causes of porosity in submerged arc welding, and most of them are process-control issues rather than equipment failures:

• Joint contaminants such as oil, mill scale, paint, moisture, or cutting residue
• Rusty or dirty electrode wire feeding through the contact tube
• Insufficient flux coverage, so the arc breaks out of the blanket
• Contaminated recycled flux, especially if the recovery system picked up slag or grinding dust
• Entrapped slag from a previous pass or from a tack weld
• Wrong parameters, particularly arc voltage set too high
• Excessive travel speed, which traps shielding gas before it can escape
• Slag residue left by tack welds made with covered electrodes
• Flux moisture picked up from an open bag or a humid shop
• An overly heavy flux burden that lets gas build up under the cover
• Arc blow, worst on DC setups and near the ends of plates or tack welds

Flux condition deserves special attention. Granular flux is hygroscopic, and a bag left open overnight in a humid shop can carry enough moisture to gasify during the weld. Keep flux sealed, store it in a heated oven when required by the procedure, and recondition recovery-hopper material before it goes back into circulation. Understanding how flux and wire interact in related processes also builds good intuition here.

Centerline Porosity on Two-Sided Joints

Second-side welds on double-sided joints have a frustrating signature: a straight line of pores running down the center of the bead. The industry guidance is consistent — that centerline porosity usually sits exactly where the first side failed to fuse. Gas generated by incomplete fusion on side one vents upward through the fresh weld metal. Fix the fusion and the porosity almost always disappears, so stop adjusting flux and voltage and go look at first-side penetration, head alignment, and joint prep instead.

Lack of Fusion: Usually Too Much Metal, Not Too Little Current

Technicians reaching for more amps is the classic misdiagnosis. In SAW, lack of fusion is frequently caused by the opposite problem: too much weld metal for the travel speed. When travel is too slow, molten metal rolls forward ahead of the arc and cushions the arc force, so the arc digs into its own puddle instead of the base plate. The fix is often to speed up, not to turn up. Other contributors are inconsistent joint prep, wrong contact-tip-to-work distance, and on pipe work, an incorrectly set weld head offset that walks the arc up the bevel wall. Review the technical fundamentals behind these

1. Slag inclusions. Deep, narrow grooves plus low travel speed or excessive current trap flux and slag where the bead cannot wash it out. Slow down the deposition rate, widen the groove or change the pass sequence, and chip and brush every layer before the next pass goes in.
2. Undercut. A groove melted into the base metal at the bead toe, usually from voltage that is too high, travel that is too fast, or a work lead placed so current concentrates on one side. Undercut is a measurable, coded defect, so know the limits before you grind: the acceptance criteria are laid out in the guide to undercut causes and acceptance limits.
3. Cracking. Hydrogen from damp flux, high joint restraint, and a high carbon equivalent are the usual trio. Control hydrogen with low-H4-rated consumables, apply preheat per the AWS D1.1 carbon-equivalent tables, and reduce restraint through sequencing. Materials that harden readily deserve extra caution — see how to weld hardened steel without cracking.

A Defect Checklist Before Production

Run through this short list before the first production pass:

• Flux conditioned, dry, and screened; hopper bed deep enough to fully cover the arc
• Wire clean, contact tube free of scale, correct stick-out set
• Tacks chipped and brushed, or welded with E6010, E6011, E7015, E7016, or E7018
• Joint prep, head offset, and travel speed verified on a run-off plate
• Interpass cleaning method defined for multi-pass beads

Documented parameter windows and a qualified procedure are what turn these fixes into repeatable results rather than operator memory. When the arc stays submerged and the settings stay honest, SAW repays the discipline with beads that need almost no post-weld repair.

SAW Safety, Fumes, and Health: What OSHA, NIOSH, and IARC Say

Submerged arc welding has a reputation as one of the “clean” processes, and on the shop floor it’s easy to see why: no visible arc, no shower of sparks, no plume of smoke rolling off the weld zone. But calm doesn’t mean harmless. The hazards that remain with SAW are the quiet, cumulative kind, which is exactly why they get underestimated. If you’re responsible for a SAW cell, here’s what OSHA, NIOSH, and IARC actually say, and how to translate the rules into practice.

Why SAW Produces Less Fume, but Not None

Because the arc is buried under granular flux, the fume and ultraviolet output that escape into the breathing zone are dramatically suppressed compared with open-arc processes. That is a genuine advantage. But flux is a chemical mixture, not sand. As it melts, flux constituents release gases, and coating or base-metal contaminants still vaporize into the weld zone. Fluxes containing bonded sodium silicate can add to the fume load, and fluoride-bearing fluxes deserve their own handling discipline. The bottom line: lower fume is not no fume.

That’s the practical context behind two landmark health decisions:

• In 2018, the International Agency for Research on Cancer (evaluated welding fumes and classified them as carcinogenic to humans (Group 1)). The classification covers overall welding fume and does not single out one process, so SAW exposures fall under it too.
• The ACGIH withdrew its previous welding fume TLV-TWA of 5 mg/m³ back in 2004, leaving no single-number threshold to hide behind. NIOSH instead recommends keeping welding exposures as low as feasible, and notes that welders can be harmed even when individual metal components sit below their OSHA PELs.

One specific PEL worth memorizing: hexavalent chromium is limited to 5 µg/m³ as an 8-hour TWA under 29 CFR 1910.1026 and 1926.1126. Stainless and hardfacing work through a SAW cell can push you there faster than most people expect.

What OSHA 1910.252 and 1910.254 Require

Two OSHA standards anchor SAW compliance. 1910.252 covers welding, cutting, and brazing in general industry; 1910.254 covers arc welding specifically. The requirements that most often trip up SAW shops:

• Hazard communication. 1910.252(c)(1)(iv) requires hazard communication covering hazardous materials in fluxes, coatings, coverings, and filler metals, including labels and access to Safety Data Sheets. Every bag on your pallet should be readable and current.
• The ventilation notice. 1910.252(c)(1)(v) requires containers of filler metal and fusible granular material to carry the exact notice: “Do not use in areas without adequate ventilation.” If your flux bags don’t have it, that’s a finding waiting to happen.
• Eye protection for others. 1910.252(a)(3) requires helmets or hand shields during arc welding and cutting operations excluding submerged arc welding — a rare regulatory wink at SAW’s buried arc. But helpers, fitters, and recovery-system attendants still working nearby need proper eye protection from radiant energy and flying slag.
• Confined spaces and ventilation. Tank interiors, box sections, and vessel shells demand written atmospheric monitoring, forced ventilation, and permit programs under 1910.146.
• Respirators. Required whenever engineering controls can’t bring exposures to acceptable levels, per 1910.134.

For personal protective equipment beyond the eyes — leathers, sleeves, metatarsal guards, and hearing protection around vacuum recovery and screening equipment — see our welding PPE guide.

The Hierarchy of Controls, Applied to a SAW Cell

The most useful takeaway from the research is practical: NIOSH-reviewed field studies found that local exhaust ventilation can reduce total fume, manganese, and Cr(VI) exposure by 40–50% or more compared with natural ventilation, though no LEV system is 100% effective. So build your controls in order:

1. Eliminate or substitute. Pick lower-hydrogen, lower-throwoff flux-electrode combinations where the procedure allows; control coating and grease contamination before welding.
2. Engineering controls. Source-capture hoods at the flux hopper and recovery point, enclosed cells with downdraft ventilation, and proper general dilution airflow sized for the shop volume.
3. Administrative controls. Rotate operators, schedule high-fume welds in ventilated bays, monitor exposure by task, and keep the arc running by reducing stick-out mistakes and rework — every rejected weld doubles the exposure.
4. PPE. Last line, not first: respirators fit-tested, hearing protection, and burn protection from hot slag and workpieces. For the burn side of the equation, see our guide on weld burns on arms.

Note: Flux deserves storage discipline too. Keep bags on pallets off the floor, sealed against shop humidity, and track lot numbers so a bad batch can be traced back — exactly the failure mode behind the pressure vessel seam in our opening story.

Treat SAW’s calm as a design feature that buys you control, not permission to relax. The shops that pass audits and pass radiography are the ones that document hazard communication, ventilate deliberately, and handle flux like the chemical it is.

SAW Productivity, Costs, WPS/PQR, and Practical Tips for Real Shops

A perfect bead means little if the paperwork cannot prove it, and less still if the job lost money making it. This closing section covers two things: a defensible quality system and a realistic cost per foot of weld.

Qualifying a SAW WPS: Essential Variables, PQR, and Prequalified Joints

Every production SAW run needs a written procedure specification. For structural work, AWS D1.1 lets you build on the prequalified joint details in Chapter 3, but only if your actual process stays inside the essential variable windows. For pressure vessels and pipe you qualify under ASME Section IX, and the variables you watch shift accordingly: current type and polarity, voltage, travel speed, electrode diameter and classification, flux type, heat input range, preheat and interpass temperature, and the joint geometry you actually prepared. Our separate walkthrough of pipe welding procedures, positions, passes, and inspection covers the same qualification logic in more depth.

Behind the WPS sits the PQR, the test record proving the joint meets its specified tensile, bend, and impact requirements. If your product is a coded pressure vessel, the flux-electrode combination also has to be an approved SFA-5.17 listing under ASME Section II Part C, which is why shops keep the manufacturer’s certification batch numbers with the weld records rather than in a filing cabinet across the plant.

Cutting the Cost Per Foot of Weld

Labor is the largest single cost in any welding operation, so Miller’s SAW productivity guidance starts with a blunt truth: small reductions in arc-off time, pass count, and rework move the bottom line more than haggling over wire price. Work the levers in this order:

• Joint design first. Reducing the cross-sectional area of the groove reduces the number of passes, and fewer passes means less arc-on time. A narrow-gap or single-V preparation with a smaller included angle can remove whole passes before you touch a single machine setting.
• Then travel speed. Raising travel speed at constant parameters shortens the seam time, though you must re-check penetration and watch for gas trapping at the trailing edge of the flux pile.
• Then arc-off time. Count every stop: repositioning the manipulator, running out of wire, flux hopper refills, tack-up delays. Mechanizing flux delivery and wire changeovers often buys more than a faster wire feed.
• Then consumables. Metal-cored wire can lift deposition rates and wetting while keeping travel speed up, which is where the real savings appear on long seams.
• Then rework. Every rejected radiograph costs several times the original weld. Undercut, porosity, and lack of fusion all trace back to parameter drift, and undercut in particular is cheaper to prevent than to gouge out and fill.

The numbers can be dramatic. A 2024 experimental study published by Springer on ASTM A516 Gr.70 plate reported weld time savings of about 61% at constant heat input, with comparable hardness profiles across both the weld zone and the HAZ. That last part is the point worth remembering: the savings came from process and joint optimization, not from cheating the thermal cycle.

Flux Storage, Reconditioning, and Reuse

Flux is a consumable you can ruin without ever striking an arc. Store it sealed, off the floor, and dry, then recondition it in ovens at the manufacturer’s specified temperature and dwell. Recovered flux stays within the manufacturer’s permitted reuse ratio, runs through a magnetic separator to catch arc-spatter and tramp iron, then through a properly sized sifter before it goes back in the hopper. Wet or screened-out flux belongs in the scrap bin, not the hopper. For shops coming from gas-shielded processes, the discipline is closer to our flux core wire welding tips than most people expect: contamination control is the whole game.

Training Operators Who Cannot See the Puddle

This is the under-discussed SAW problem. Many SAW operators are not experienced manual welders, and the process hides the one feedback loop manual welders rely on: they cannot see the puddle. An operator who cannot read arc length by eye has to be taught what 500 amps and 35 volts look like in a finished bead, what a flux depth change does to profile, and what a starved joint looks like before radiography tells them. Classroom time on the variables, paired with sectioning and etching sample welds, pays for itself the first time an operator diagnoses a lack-of-fusion root cause without calling the supplier. Knowing how electrodes are classified is a good starting block for that training.

Conclusion: When SAW Is the Right Process

Choose SAW when you have thick plate, long seams, flat or rolled positions, and enough volume to amortize setup. Look elsewhere for out-of-position work, thin sheet, and short one-off joints, where TIG on sheet metal or a wire process wins on setup time. The trajectory is clear: narrow-gap SAW keeps shrinking joint volume, laser and optical seam tracking close the gap on manual fit-up variation, and programmable manipulation turns a shop’s best process knowledge into repeatable recipes. The shops that combine disciplined flux handling, qualified procedures, and operators who understand the variables will keep pulling cost out of every foot of weld, one seam at a time.

Frequently Asked Questions

What is submerged arc welding used for?

SAW is used for long, heavy welds in thick carbon steel: pressure vessel shells and heads, wind tower sections, ship hull plate, storage tanks, pipelines, nuclear containers, and heavy structural plate. It excels whenever joints are long, run in flat or horizontal position, and demand high deposition with consistent quality.

Why can’t you see the arc in SAW?

The arc is completely covered by a blanket of granular fusible flux, which melts into a conductive slag and shielding gas layer over the weld pool. This suppression of the visible arc also cuts spatter, arc light, and much of the fume — which is why OSHA exempts SAW from the arc-welding helmet requirement for the operator.

What are the main disadvantages of SAW?

It is essentially limited to flat and horizontal positions, requires heavy fixturing, positioning equipment, and significant setup time, is mostly restricted to steels, stainless, and nickel alloys, and generates slag that must be removed between passes. The operator also cannot directly watch the weld puddle.

What causes porosity in submerged arc welding?

The most common causes are moisture or contamination in the flux, oil, rust, paint, or mill scale in the joint, insufficient flux coverage, arc blow, and travel speeds so fast that gas cannot escape before solidification. Porosity on the second side of a two-sided joint usually traces back to lack of fusion on the first side.

Can SAW flux be reused?

Yes — unused flux is vacuum-recovered, screened, and recirculated, which is a real cost advantage. It must be sieved to remove fines and metallic or scale contamination, kept dry, and mixed with virgin flux within the manufacturer’s permitted ratio, because degraded particle size and moisture are leading causes of porosity.

What current and voltage does SAW typically run at?

Common ranges are 200–1500 amps with multi-electrode systems reaching far higher, voltages around 24–40 V depending on wire size and bead requirements, and wires from about 5/64 to 7/32 inch. Wire feed speed is the primary current control, and travel speeds can exceed 100 in/min on thin sheet.

Is submerged arc welding safe? Does it produce fumes?

SAW is a comparatively low-fume, low-UV process because the arc is submerged, but it still produces gases and metal fume from flux and base metal. IARC classifies welding fumes overall as Group 1 carcinogenic, so ventilation, hazard communication, SDS access, and respiratory protection when needed remain mandatory under OSHA rules.

How do I qualify a SAW welding procedure?

Develop a WPS listing essential variables — electrode and flux classification, current type and polarity, current, voltage, travel speed, stick-out, preheat and interpass temperature — then qualify it with a PQR per AWS D1.1, AWS D1.5, ASME Section IX, or EN ISO 15614 as the governing code requires. AWS D1.1 Chapter 3 prequalified joint details can reduce qualification testing.