Root Face Welding: Why That Small Flat Edge Decides the Weld

The fitter swears the bevel is perfect, the welder strikes an arc, and thirty seconds later a glowing hole opens in the pipe root. Nothing about the machine changed. What changed was a 2 mm flat strip of metal nobody measured that morning: the root face. It is the smallest dimension on the drawing and the one most likely to decide whether the joint passes inspection or gets cut out.

Root Face Welding: The Tiny Flat That Decides Pass or Fail

The fitter swears the bevel is perfect. The welder strikes an arc, holds the keyhole for maybe thirty seconds, and then a glowing hole opens in the pipe root. Nobody touched the machine. Nobody changed the amperage. What changed was a narrow strip of flat metal at the bottom of the groove that no one measured that morning: the root face.

Here is the definition, before anything else. The root face is the flat, unbeveled strip of base metal left at the bottom of a groove, running along the joint root. In standard terminology, AWS A3.0 describes it as the portion of the groove face adjacent to the joint root. It is the metal the arc must melt through before the root actually opens up and the weld can penetrate. Every other dimension on the drawing describes the shape of the groove. The root face describes how much resistance the arc meets at the bottom of it.

That single dimension drives the two failures that ruin root passes. A root face that is too thick is a leading cause of lack of root penetration, because the arc never gets through the land and unmelted metal stays at the bottom of the joint. A root face that is too thin promotes burn-through and suck-back, because there is almost nothing left to hold the puddle. And when a root pass is rejected, the fix is rarely cosmetic. Rework usually means gouging or cutting the joint out, re-beveling, re-welding, and re-inspecting, with the schedule slipping while it happens.

So what changes with the land width?

• Too thick: the puddle sits on top of cold metal, penetration stalls, and lack of fusion hides under a weld that looks acceptable from the outside.
• Too thin: the arc punches through, the root caves in, and the welder spends the next pass trying to rebuild metal that should never have been lost.
• Just right: the land gives the first bead a shelf to sit on while still letting the arc melt cleanly into the root, producing a controlled penetration bead with consistent reinforcement.

This guide is written for the people who actually deal with that number: fitters who cut and match the bevel, welders who have to burn through it, supervisors and inspectors who sign it off, and hobbyists making the jump from plate to pipe where fit-up suddenly matters more than settings. There are no claims of shop testing here. Everything below is documented practice drawn from published guidance, manufacturer technical material such as Hobart’s groove weld reference, and recognized standards.

Bevel dimensions are not arbitrary. ASME B16.25 butt-weld ends, for example, commonly specify a 37.5° bevel, which produces a 75° included angle, paired with a root face of roughly 1.6 mm ± 0.8 mm (ASME B16.25). Move that land from 1.6 mm to 3 mm and the arc behaves like a different machine is running.

Across the sections that follow, we will work through the vocabulary that prevents most confusion, the geometry linking root face with root gap, bevel angle and wall thickness, and the consequences for penetration, fusion and code acceptance. From there it moves into process-specific practice for SMAW, GTAW, GMAW, FCAW and SAW, what the standards actually fix versus what the WPS controls, and how to cut and measure a land properly. The final sections cover root-pass technique, defect troubleshooting, and a shop-floor checklist you can pin to the wall. If you also work aluminium, some of the same fit-up discipline carries across, and the principles in our aluminium MIG welding guide translate well to groove preparation on lighter sections.

The principle worth carrying through all of it is simple. Control the root face, and you control the root.

What Is a Root Face? Land, Root Gap, and the Vocabulary of a Groove

Most arguments about root faces in a fabrication shop are not really about welding. They are about words. One fitter calls it the land, the drawing calls it the root face, the inspector asks about the root opening, and the welder talks about the gap. Everyone is looking at the same 2 mm strip of steel and describing it differently, which is exactly how a joint gets prepared, measured, welded, and rejected without anyone actually disagreeing about the metal. Get the vocabulary straight once and most of the confusion in the rest of this topic disappears.

The Parts of a Groove, Described in Words

Picture a single-V butt joint on 10 mm plate, cut and ready to fit. Every dimension below lives inside that picture.

• Joint root: the point where the two members come closest together, at the very bottom of the groove. It is a location, not a measurement.
• Groove face: the angled cut surface of the bevel. This is the slope the arc has to climb as it fills.
• Bevel angle: the angle of one groove face measured from a plane perpendicular to the plate surface. It describes one side only.
• Groove angle (included angle): the total angle between the two groove faces. Two 30° bevels make a 60° included angle. Confusing the two is a classic estimating error.
• Root face: the flat, unbeveled strip of base metal left at the bottom of the groove, next to the joint root.
• Root edge: the boundary where the root face meets the root. If the root face is zero wide, the bevel runs straight to a knife edge, and that is a feather edge.
• Root opening (gap): the distance between the two members at the root, measured after fit-up.
• Land: the shop word for the root face in a groove weld. It also has a second, different meaning in partial-joint-penetration work, covered below.
• Depth of preparation: the distance from the plate surface down to the root face, in other words the bevel depth.
• Root penetration: how far the weld metal extends through and past the root.
• Root reinforcement: weld metal sitting proud of the root surface.
• Root concavity: a hollow or sag in the root surface, the shape welders call suck-back.

Why Land and Root Face Mean the Same Thing

Pipeline hands have said “land” for generations, and older drawings still print it. AWS standardised the language through documents such as AWS A2.4, which governs the symbols used to call out groove welds, root openings, groove angles, and root faces, and “root face” became the preferred term because it describes where the feature sits rather than what it does. The same housekeeping renamed the gap: in AWS terminology, root opening is preferred, and “gap” is a nonstandard shop term for the identical dimension.

Shop-floor term	Preferred term	What it actually is
Land	Root face	The flat strip at the bottom of the groove
Gap	Root opening	Distance between members at the root
Bevel face	Groove face	The angled cut surface
Included angle	Groove angle	Total angle across both faces
Knife edge	Feather edge	A root face of zero width

Root Face vs Land in Partial-Joint-Penetration Welds

This is where the two words genuinely part company. In a full-penetration groove weld, the root face is simply metal the arc must consume to reach through. In a partial-joint-penetration (PJP) joint, the unwelded base metal left below the preparation is also called the land, and it is deliberately left there to limit how deep the weld goes. The designer sized it that way on purpose. Grinding it away, or welding deeper than the WPS allows, changes the joint’s behaviour rather than improving it.

Feather Edges: Zero Root Face, on Purpose

A feather edge leaves no flat at all, just a fine knife edge where the two bevels meet. It gives maximum root access and is used on thin material, on some TIG roots where the welder wants the arc to keyhole immediately, and where a drawing specifies a full-penetration joint with no land. It is also unforgiving: with nothing for the root pass to sit on, gap control and heat control have to be right the first time. Keep the groove faces clean before fit-up too, since contamination at the root shows up on the inside of the pipe. Aluminum cleaner for welding covers that prep step for aluminium work. For a broader look at groove geometry, Hobart Brothers publishes a useful groove weld reference.

Measured Along the Bevel, Not Along the Plate

Here is the convention that trips up nearly every trainee. The root face is measured parallel to the bevel face, along the direction of the groove face, not vertically down through the thickness and not along the plate surface. On a 37.5° bevel the difference is noticeable; on a steep J-groove it is large. Inspectors regularly find fitters measuring vertically with a rule and reporting a land that is 10 to 20 percent off, which then feeds straight into the welding variables. Measure the way the symbol is drawn, not the way that feels natural with a scale in your hand.

With these terms fixed, the next question becomes practical: how do the root face, root opening, groove angle, and wall thickness actually influence each other on a real joint?

How Root Face, Root Gap, Bevel Angle, and Thickness Work Together

A groove is not a set of independent numbers. Change the land by half a millimeter and the gap that worked yesterday suddenly burns through, or the bevel angle that fed the puddle nicely starts leaving cold lap at the root. Root face, root gap, bevel angle, and wall thickness behave like four dials wired to the same circuit. Turn one and the others change what they mean.

The Four Variables and What Each One Controls

• Root face (r): the flat land left at the bottom of the bevel. It is the metal the arc must melt before the root opens. More land means more support and less burn-through risk, but it also means less penetration for the same current.
• Root opening or gap (g): the space between the two faces at the root. It gives the arc a path to the back side and lets the keyhole form. A wider gap improves penetration but adds fill volume and invites drop-through.
• Included groove angle (θ): the total angle between the two bevel faces. It sets how much room the electrode or torch has and how much filler the joint swallows. Wider angles mean easier access and much more metal.
• Wall thickness (t): the base dimension everything else scales against. Thick wall increases heat sink, restraint, and fill volume; thin wall gives the puddle nowhere to hide.

Arc access, shrinkage, and penetration all follow from these four. A tight angle with a thick land is a joint that punishes every small mistake in current or travel speed.

Why Fill Volume Scales With the Square of Bevel Depth

For a single-V groove, the cross-sectional area of the beveled portion is approximately (t − r)² × tan(θ/2). That squared term is the part people underestimate. Double the bevel depth and you roughly quadruple the area to be filled, which drives filler consumption, arc time, heat input, and transverse shrinkage together. Increasing the groove angle adds filler and shrinkage as well, which is why narrow-groove designs dominate heavy-wall work where weld volume becomes a cost and distortion problem.

Arc energy follows the same logic per unit of length. Heat input in kJ/in can be estimated as 60 × volts × amps ÷ (1000 × travel speed in in/min), so a pass at 150 A, 24 V and 4 in/min works out to about 54 kJ/in. More fill volume usually means more passes at that heat input, and the part absorbs all of it.

Two Worked Comparisons

Half-inch plate, 60° included, 1/8 in gap. With a 1/8 in land, bevel depth is 0.375 in and the bevel area is about 0.141 in². Push the land to 1/4 in and bevel depth drops to 0.25 in, so area falls to roughly 0.0625 in² — about 2.25 times less filler, but now a quarter inch of flat steel sits between the arc and the far side of the joint. That is a lot of metal to melt with a root pass, and lack of penetration becomes the likely outcome unless the gap and current are adjusted.

Pipe with a 25 mm wall and 60° included angle. A 6 mm land leaves 19 mm of bevel depth; a 10 mm land leaves only 15 mm. The bevel areas come out near 208 mm² and 130 mm², so the thick-land joint needs about 60% more metal in the groove while offering much more resistance to the arc breaking through. Neither number is wrong on its own. They belong to different techniques.

Typical Starting Ranges by Thickness

Thickness band	Starting root face	Starting gap	Practical notes
Thin sheet, under 3 mm	0–0.5 mm (feather edge)	0–1 mm	Square groove common; fit-up and current matter more than land
3–6 mm	1–1.5 mm	1.5–3 mm	Small V or square groove; watch for burn-through on the first pass
6–12 mm	1.5–2 mm	2–3 mm	The everyday pipe and plate range; most WPS values live here
Heavy wall, over 12 mm	1.5–2.5 mm; 3–4 mm on U and J grooves	2–3 mm	Narrow-groove and compound bevels cut fill volume and shrinkage

Standard butt-welding ends under ASME B16.25 narrow the field further: a 37.5° ± 2.5° bevel with a root face of 1.6 mm ± 0.8 mm (1/16 in ± 1/32 in). That is a useful sanity check when a drawing looks unusual.

The Trade-Off Logic in Plain Terms

A bigger root face gives the root pass a shelf to sit on and resists burn-through, but it steals penetration. A bigger gap buys penetration and arc access, but it adds fill volume and makes drop-through more likely. Wider angles make the joint friendlier to weld and more expensive to fill. Thickness magnifies every one of those effects, because a heavy section pulls heat away from the root faster than thin sheet ever will.

The practical limit also shifts with position. A land and gap that keyhole cleanly in the flat position can collapse in 6G, where gravity works against the puddle and travel speed changes. That is why welders working overhead or on the vertical often step down slightly in current rather than changing the fit-up, and why rod selection matters as much as geometry — the all-position welding rod guide is worth reading alongside any fit-up chart.

The useful habit is to treat the groove as one system. Decide the land from the thickness and process, set the gap to match the penetration you need, then confirm the angle gives the arc room to reach the root. For a deeper look at how groove geometry is described in manufacturer guidance, Hobart Brothers’ groove weld reference is a solid starting point.

Why the Root Face Matters: Penetration, Fusion, Strength, and Code Acceptance

On the drawing, the root face is just a small dimension. In service, it becomes the difference between a joint that carries load as one piece of metal and one that merely looks welded from the outside. The root is the first metal to form and the last line of defense against leakage, cracking, and rejection.

The Root Is the Load Path

Root penetration and root fusion are not cosmetic details. They determine whether the weld connects fully through the joint thickness. When the land is too thick, the arc may melt the bevel faces and leave a shelf of unmelted base metal at the bottom. That is incomplete penetration: the weld sits on top of the land instead of reaching through it. The result is a reduced effective throat and a sharp notch at the root, which acts as a stress concentrator under pressure cycles, vibration, or bending.

A lack of root fusion is even more deceptive. The bead can look full and the cap can pass a visual check while the root never actually bonded to one groove face. That unfused plane behaves like a crack waiting for a reason to grow. In fatigue service, small root defects matter because stress concentrates at sharp edges, not at smooth transitions.

How the Root Pass Sets Up Every Pass After It

The root pass is the foundation for the hot pass, fill, and cap. A proud, uneven root creates a ledge where slag can hide and where the hot pass may bridge instead of fusing. That leads to slag traps and lack-of-fusion indications that are expensive to find and worse to repair. A sunken root is not a safe compromise either. It reduces the effective throat and may exceed the allowable root concavity.

ASME B31.3 is direct about the consequences. ASME B31.3 treats lack of fusion as an unacceptable imperfection in its weld acceptance criteria, and Table 341.3.2 caps root surface concavity at about 1.5 mm for wall thickness up to 13 mm, 3 mm up to 51 mm, and 4 mm above 51 mm. Those limits are not suggestions. They reflect the fact that a thin, concave root may not retain pressure or resist cracking as designed.

Symptom	Likely root-face cause	Consequence
Incomplete penetration	Land too thick for the process and gap	Reduced throat and a sharp root notch
Burn-through or ice hole	Land too thin, often with a wide gap	Cut-out, re-prep, and re-weld
Proud, lumpy root	Excess land or too much root reinforcement	Slag traps and lack of fusion on the hot pass
Sunken root	Thin land, high heat, or poor gap control	Concavity beyond code limits and lost throat
Root corrosion in stainless	Poor purge and excessive root oxidation	Sugaring, crevices, and reduced corrosion resistance

Fatigue, Pressure, and Corrosion Consequences

Pressure retention depends on a continuous, sound throat. A root that is unfused or deeply concave gives the internal pressure a thinner section and a geometric discontinuity to work on. Under cyclic loading, the notch at an incomplete-penetration root can initiate a fatigue crack long before the cap is stressed. In stainless steel, the root face also affects purge quality. If the land is thick and the root does not open cleanly, shielding may be poor, producing sugar or heavy oxide scale on the inside surface. That scale is not just ugly; it creates crevices and reduces corrosion resistance.

What Rejection Actually Costs

The shop cost of a bad root is rarely limited to the weld itself. A rejected joint may need to be cut out, re-beveled, re-fitted, re-welded, and re-inspected. Repair often starts with gouging out the defect, followed by grinding and a new root pass. That is where a process like air arc gouging earns its place, but it also adds heat, time, and the risk of damaging the surrounding metal. Schedule slips follow, especially on pipe racks where one joint holds up the next spool.

The bottom line is simple: the root face is a acceptance variable, not a fitter’s detail. Control the land, the gap, and the fit-up, and the root pass has a chance to fuse cleanly. Ignore it, and the smallest dimension on the drawing becomes the most expensive one to fix.

Root Face by Process and Joint Type: SMAW, GTAW, GMAW, FCAW, and SAW

A root face that works beautifully under a TIG torch can be a liability under a fast-freeze cellulose electrode. The number on the drawing may stay the same, but the process changes how much metal the arc can melt, how quickly the root opens, and how much control the welder has when things start to move. Process selection is therefore not separate from root-face selection. The two decisions belong together.

SMAW: Deep Bite, Fast Freeze

Shielded metal arc welding remains a pipeline and structural mainstay because cellulose electrodes such as E6010 produce a forceful, deep-penetrating arc and a fast-freezing slag. That combination lets a welder work a keyhole around a pipe without the root collapsing. The classic technique is a whip-and-pause motion: advance to open the keyhole, then return to fill it before it grows. Cellulose electrodes tolerate a slightly thicker land and a wider gap than many low-hydrogen electrodes because they dig aggressively.

Low-hydrogen electrodes such as E7018 behave differently. They deposit cleaner, tougher metal, but their arc does not excavate the root the same way. A heavy land paired with a tight gap is a common route to incomplete root penetration. For a starting point, many procedures use a 1.5–2 mm (1/16–5/64 in) land and a 2–3 mm (5/64–1/8 in) gap, then adjust to the qualified procedure. The broader reasons stick welding still dominates field work are covered in the advantages of stick welding.

GTAW: The Pipe Root Standard

Gas tungsten arc welding is the most common open-root pipe process for good reason: it gives the welder independent control of arc current, filler addition, and travel speed. On standard-wall pipe, a typical starting point is a 1/16 in (1.6 mm) land with a 1/16–3/32 in root gap. The arc is precise but not especially powerful, so a thick land simply stalls penetration rather than being burned through.

Fit-up discipline matters more here than in almost any other process. The land should be consistent around the circumference, the gap should be checked in at least four quadrants, and the purge should be established before the arc starts. On stainless steel, poor purge produces sugaring and oxide scale at the root that is difficult to remove later. GTAW rewards patience: a steady torch angle, controlled filler dipping, and a root that is inspected before the hot pass goes on.

GMAW and FCAW: Speed With a Narrower Window

Short-circuit GMAW is the forgiving option for fit-up. It can bridge a slightly uneven gap and works well on thinner material and less precise joints. The trade-off is a higher risk of lack of fusion at the root sidewalls, especially when the land is thick and the welder travels too quickly. Spray and pulsed-spray transfer are more productive and penetrate better, but they need tighter gap control because the arc is energetic enough to open a wide root quickly.

Flux-cored welding is fast and tolerant of outdoor conditions, but the slag introduces its own root risk. If the molten slag runs ahead of the arc, the root bead can sit on slag rather than fused metal. Root-face thickness should be matched to the process and position, and the root should be cleaned thoroughly before the next pass. Neither process rewards a “set it and forget it” approach on open roots.

SAW: Backing, Both Sides, or No Root At All

Submerged arc welding is normally performed with backing, from both sides, or on a closed root rather than as an open-root pass. The process is designed for high deposition rates in the flat position, not for controlling

What the Codes and Standards Actually Say About Root Face

Ask three people on a job site for the “correct” root face and you may get three numbers. One quotes the drawing, another quotes a procedure, and a third quotes a standard nobody has actually opened. All three may be partly right. Codes rarely hand you one universal land dimension because the right root face depends on the joint, the process, and the procedure that was qualified for the work.

The Hierarchy: Design Code, WPS, and Inspector

Think in three layers. The design code sets acceptance criteria and, in some cases, preparation requirements. The welding procedure specification and the drawing translate those rules into qualified dimensions and variables. The inspector then verifies the joint against those documents, not against a personal preference.

That hierarchy matters. A design code may permit a range of joint details, but it does not give your welder permission to invent a land on the pipe. The qualified WPS defines the root face, gap, and bevel that were tested. The drawing shows what the designer wants. When those documents disagree, stop and get it resolved before anyone strikes an arc.

The WPS also ties the qualified root face to position and technique. A procedure qualified flat does not automatically cover 6G, and consumable choices follow the same logic as all-position welding practice. Position changes how the pool behaves, which changes how much land the welder can safely leave.

ASME B16.25 and the Standard Butt-Weld End

For standard butt-welding ends, ASME B16.25 specifies a 37.5° ± 2.5° bevel and a 1.6 mm ± 0.8 mm root face. That is a nominal 75° included angle with a land of roughly 1/16 in, plus a tolerance band that can make a visible difference on small pipe.

ASME B31.3 requires end preparation consistent with the qualified WPS. In plain terms, the fitting may arrive with a B16.25 end, but the procedure still governs what gets welded. If the WPS calls for a different land or gap, the fitter needs to know that before the bevel is cut or accepted.

Hi-Lo: The Misalignment That Changes Both Sides

ASME B31.3 paragraph 328.4.3 addresses misalignment, and common practice limits hi-lo to the lesser of 1.6 mm (1/16 in) and 25% of nominal wall thickness. On thin wall, the percentage usually controls. On heavy wall, the fixed 1.6 mm limit often does.

Hi-lo quietly wrecks root-face assumptions because the two sides of the joint no longer behave the same way. One side may present a tighter land and gap; the opposite side may open up. The pool favors the open side, while the tight side risks incomplete fusion. A root face that measured perfectly on the bench can become wrong the moment the pipe is clamped off-center.

AWS D1.1, API 1104, and ASME Section IX

AWS D1.1 sets fit-up tolerances for structural groove welds and requires groove dimensions to match the approved detail. It does not give one root face for every joint. API 1104 governs pipeline girth welds and includes specific root bead acceptance criteria, which is why pipeline inspectors focus so heavily on root appearance and profile.

ASME Section IX adds the qualification layer. The procedure must be qualified, and joint design details are controlled by that qualified WPS. Change the joint design beyond the qualified limits and the WPS may need requalification. That is not a paperwork technicality; it is the reason a “small” field change to the land can become a compliance problem.

Standard	What it fixes	What it leaves to the WPS or drawing
ASME B16.25	Bevel angle and root face for standard butt-weld ends: 37.5° ± 2.5°, 1.6 mm ± 0.8 mm	Actual gap, fit-up, and joint-specific details where the piping specification differs
ASME B31.3	Acceptance, end preparation consistent with the qualified WPS, and misalignment guidance in 328.4.3	The qualified land and gap, welding variables, and repair decisions
AWS D1.1	Structural groove details, fit-up tolerances, and groove dimensions matching the approved detail	WPS essentials and position-specific technique
API 1104	Pipeline girth weld acceptance, including root bead criteria	Procedure dimensions, gap and land values, and field technique
ASME Section IX	Procedure qualification and the limits of qualified joint designs	Production fit-up within the qualified ranges

What This Means at the Fit-Up Table

• Confirm which document governs: design code, project specification, drawing, or WPS.
• Measure the root face parallel to the bevel face, not along the plate surface.
• Check hi-lo and record it before the root pass, because misalignment changes the effective land on both sides.
• Do not “improve” a qualified joint by adding land or gap without approval.

Code compliance is not about finding a magic number. It is about proving that the joint in front of you matches the joint that was qualified. Control the root face within the documents that govern it, and the root has a fair chance of passing inspection.

Cutting, Machining, and Measuring the Root Face

The root face is not made by the welder. It is made by whoever runs the torch, beveler, or lathe, and it is accepted or rejected before the first arc is struck. A perfect travel angle cannot rescue a land cut 3 mm heavy on one side of a pipe.

Thermal Cutting vs. Machining

Two routes lead to a beveled end: thermal cutting and machining. Both can produce a good root face, but each leaves a different signature for the welder.

Method	Good for	Watch out for
Oxy-fuel	Carbon steel, field work	Oxide layer, notches, dross
Plasma	Stainless, aluminum, faster cuts	Contaminated edge, dross
Pipe beveler / end prep	Repeatable shop work	Tool drift, dull inserts
Clamshell lathe / portable mill	Large or heavy-wall pipe	Setup time, alignment

Oxy-fuel cutting is fast and portable, but the cut face carries an oxide layer and a heat-affected zone, and a sloppy cut can leave a notch at the root edge. Notches act as stress risers, so flame-cut bevels should be ground back to sound, shiny metal. Dross and heavy oxide must go too, since a light wire brushing will not remove tightly bonded scale. Plasma-cut edges follow the same rule: free of notches, dross, and heavy oxide before welding (Hobart Brothers).

Plasma cutting is quicker and handles stainless and aluminum, but it can leave dross and a contaminated edge. On stainless, grind away the plasma-cut layer and use dedicated stainless abrasives, because carbon steel dust embedded in the surface starts corrosion. Aluminum needs its own routine, covered in our guide to aluminum cleaner for welding.

Machining gives the most repeatable result. Pipe bevelers cut a clean land and bevel in one setup, clamshell lathes square and bevel large pipe, and portable mills handle heavy wall. The trade-off is setup time and tool drift: a dull insert can rub a tapered land that looks fine until a gauge goes on it.

Setup: Angle, Land, Squareness, and the Ovality Trap

Set the bevel angle first, then the land. If the angle is wrong, the land width reads correctly only at one point. Use a bevel gauge or the machine scale, and check after the cut, not only during setup.

Squareness matters as much as angle. An end out of square by a degree or two shifts the gap around the circumference and changes how much land the root pass sees. Out-of-round pipe and wall thickness variation make it worse. Seamless pipe can vary in wall thickness within permitted tolerances, so two pipes of the same schedule may not match at the root: a fitter sets a perfect gap at top dead centre and still finds it tight at the bottom.

For ASME butt-weld ends, ASME B16.25 sets the familiar 37.5° ± 2.5° bevel and 1.6 mm ± 0.8 mm root face, though the WPS and drawing govern the actual joint. Checking both catches mismatched assumptions before the root is welded.

Measurement That Actually Catches a Bad Land

Eyeballing a land is how good joints become rework. A root face gauge, V-weld measuring gauge, or bridge cam gauge gives a direct reading; calipers work for land width and gap when geometry allows. Check the tool against a known reference and retire worn gauges, since a dropped gauge can be off by enough to matter.

Measure the land parallel to the bevel face, not along the pipe surface, then repeat at four quadrants: top, bottom, and both sides. One reading says the joint is right at that point; four say whether the root has a chance around the whole circumference. Record low and high readings, not a tick in a box.

Visual inspection before the root pass commonly checks groove angle, root opening, root face, and joint alignment (AWS). If any one is out, the welder is being asked to fix a fit-up problem with arc force.

Documentation and a Pre-Weld Checklist

Fit-up sheets and pre-weld sign-off turn a verbal “looks good” into a record. Note the measured land, gap, bevel angle, and hi-lo at each quadrant, plus date, joint number, and inspector. Photograph the prepared end with a gauge in frame; photos do not replace measurement but settle arguments later.

1. Confirm the WPS and drawing dimensions before cutting.
2. Verify pipe schedule, wall thickness, and material certificate.
3. Cut and bevel with the land width controlled.
4. Grind thermal-cut faces to remove oxide, dross, and notches.
5. Clean the joint faces before fit-up.
6. Measure bevel angle, root face, gap, and hi-lo at four quadrants.
7. Record readings and flag anything out of tolerance.
8. Get pre-weld sign-off, then protect the end from moisture and damage.

A root face that is cut, checked, and documented gives the welder a fair chance. One that is guessed turns the root pass into a repair job.

Welding the Root Pass: Technique That Matches Your Root Face

Everything before this was measurement and preparation. The root pass is where those numbers turn into heat, and where a correctly sized land gets thrown away by a machine set for a different one. The land decides how much steel the arc must melt before the keyhole opens. Amperage, arc length, travel speed, and work angle decide whether that melting happens at the right moment, in the right place, at the right rate. Match the two and the root comes in clean. Mismatch them and you get the glowing hole from the opening story.

Setting the Machine Around the Land

Amperage is not a number you inherit from the last job. A thick land needs more energy to break through, but extra energy should come from arc time and manipulation before it comes from the dial. Crank the amps to chew through a heavy land and you undercut the bevel faces and burn through on the thin side of a hi-lo joint. A long arc does similar damage more slowly: it widens the cone, spreads heat, and gives you a mushy keyhole you cannot read.

Travel speed is the quiet variable. Move too fast on a thick land and the arc never establishes a keyhole, so you get incomplete penetration that looks fine from outside. Move too slowly and the puddle grows until the root drops out. Let the keyhole tell you the speed rather than deciding it in advance. On many qualified procedures the finished root bead stays around 3 mm (1/8 in) or less, which keeps fusion controllable and leaves room for the hot pass to work.

SMAW Open Root: Reading the Keyhole

With cellulosic electrodes you control a small keyhole at the leading edge of the puddle and let it close behind you. Whip and pause is the classic pipe manipulation: a short forward push to open the keyhole, then back into the puddle to fill it. Pause length is your penetration control. A thick land wants a fraction more pause; a thin land wants you moving or the puddle drops.

Electrode angle matters more on pipe than on plate. On vertical and overhead sections, keep the arc pushing slightly into the groove so the puddle does not sag ahead of you. If you are still choosing electrodes, this comparison of stick welding advantages explains why SMAW remains the default for open-root pipe work. Low-hydrogen electrodes behave differently: gentler arc, tighter puddle, stricter storage.

GTAW Open Root: Angle, Timing, and Purge

GTAW dominates pipe roots because you can watch the keyhole and control filler down to a single dip. Torch angle around 10–15° from perpendicular, pushed slightly forward, keeps the arc focused on the land. Dip technique gives precise, small additions and suits tight lands; lay-wire feeds continuously and suits thicker lands and faster travel. Pulsing helps: the background current holds the puddle while the peak gives fusion at the bevel.

Purge quality is not optional on stainless. Most procedures target oxygen below roughly 100 ppm inside the pipe to prevent oxidation and heat tint, and reactive metals such as titanium need far lower. Use dams or purge paper, monitor with an oxygen analyser, and vent properly. Over-purging fails too: excessive purge pressure or amperage is a recognised cause of root concavity, or suck-back.

Inspect Before the Hot Pass

Look at the root before you cover it. A hot pass can smooth a slightly high or uneven root and burn out small bits of trapped material. It cannot fix lack of fusion, because that is a bond problem, not a shape problem. If the root is cold-lapped or sitting proud enough to trap slag, grind it now.

Positions and Heat Input

1G forgives because gravity holds the puddle. 2G lets it sag toward the lower plate. 5G and 6G are where land control earns its keep, since the same root face has to behave flat, vertical, and overhead in one pass. For travel direction and rod choice across positions, this all-position welding rod guide is a useful companion. On stainless and high-strength steels, keep heat input inside the WPS range; sensitisation and hydrogen-assisted cracking are both heat-driven. Clean between passes, and never carry slag into the next bead.

Root bead appearance	What it tells you	What to change next time
Narrow, high crown, no keyhole visible	Land too thick for the heat, or travel too fast	Reduce land, tighten arc, slow slightly
Sunken, concave, dark edges	Excess heat, over-purge, or too much gap	Lower amps, reduce purge flow, tighten gap
Blowhole or enlarged keyhole	Thin or knife-edged land, wide gap, hi-lo	Add land, close gap, fix alignment
Cold lap at one bevel face	Angle error or arc favouring one side	Correct work angle, check bevel symmetry

For groove geometry and technique fundamentals, Hobart Brothers publishes a useful groove weld technical article, and ASME B31.3 is where the qualified procedure requirement ultimately sits — see the B31.3 process piping standard.

Root Face Defects and Troubleshooting: Burn-Through, Suck-Back, LOF, and LOP

Almost every root defect traced back to the land started as a dimension nobody re-checked. The arc does not care about intentions; it responds to how much metal sits under it and how much space it has to push through. This section is meant to be read with a grinder in one hand and a gauge in the other: identify the appearance, work backwards to the cause, then fix the variable rather than the symptom.

Burn-Through and Ice Holes

Burn-through looks exactly like it sounds: a ragged opening punched clean through the root, often with a stalactite of metal hanging on the inside. It happens when arc energy exceeds what the root face and gap can support. A land that came out at 0.8 mm instead of 1.6 mm, paired with a gap that drifted wide, leaves nothing for the arc to sit against. High amperage, long arc length, slow travel, and small-diameter pipe all push the same direction. Detection is visual inside the joint or by borescope; radiography confirms the size. The fix is mechanical before it is electrical: cut back to sound metal, re-bevel to restore the specified land, re-fit the gap, then run cooler with a shorter arc and slightly faster travel.

Suck-Back and Root Concavity

Suck-back is burn-through’s quieter cousin. The root bead sinks below the inside surface, leaving a smooth hollow where reinforcement should be. The cause is usually excess heat against too little filler, or a thin land over a generous gap. Unlike a hole, suck-back passes a casual look and fails on measurement. Root concavity reduces the load-bearing throat, and acceptance criteria tie the allowable amount to wall thickness, as set out in ASME B31.3. Correct it by adding filler earlier, reducing current, tightening the gap, or increasing travel speed.

Incomplete Root Penetration and Lack of Root Fusion

Incomplete penetration shows as a root bead that never reached the inside surface, or reached it on one side only. The land was simply too thick for the heat available. Lack of root fusion is more subtle: metal was deposited, but it sat on the bevel without melting into it. It is often associated with excessive land, a tight gap, low amperage, fast travel, or contamination on the groove face. Both are planar defects and both concentrate stress. The land fix is to thin the face or open the gap within WPS limits; the process fix is to restore current and slow down. Electrode selection matters here too, and a look at this all-position welding rod guide helps match rod behavior to position.

Reinforcement, Porosity, Tungsten, Sugar, and Hi-Lo Steps

Excessive root reinforcement is a land-and-gap mismatch in the other direction: wide gap, thin face, too much filler, slow travel. It creates a notch and a slag trap for the hot pass. Root porosity is almost always gas or contamination, not geometry. Tungsten inclusion is a GTAW problem from dipping the electrode or running too hot. Sugar on stainless is purge failure, visible as a dark, rough oxide inside the pipe. A misalignment step comes from hi-lo, which leaves one side of the joint effectively thicker than the other.

The Land-to-Gap Seesaw

A thin land with a wide gap and a thick land with a tight gap can look almost identical on a drawing, yet they behave nothing alike. The first invites burn-through and suck-back; the second invites lack of fusion and incomplete penetration. Treat land and gap as a matched pair. When one moves, adjust the other and re-qualify the change.

Repair Without Making It Worse

Grind or gouge out the defect to sound metal. Carbon arc gouging is fast but must be followed by grinding to remove the carbon-affected layer, as covered in this air arc gouging guide. Re-prepare the joint, weld with a qualified repair procedure, re-inspect, and record what was done. Repeated repair welding on the same location can locally alter hardness and toughness, and most specifications cap the number of attempts.

Defect	Primary Root-Face Fix
Burn-through	Restore specified land; reduce gap
Suck-back / concavity	Thicker land, tighter gap, more filler
Incomplete penetration	Thin the land or widen gap within WPS
Lack of root fusion	Reduce land, clean face, restore current
Excessive reinforcement	Widen gap, reduce filler deposit
Hi-lo root step	Realign pipe; re-check land on both sides

Short Note: fix the fit-up before you touch the machine. Most root defects are preparation failures wearing a welding disguise.

Common Mistakes, a Shop-Floor Checklist, and the Bottom Line

Everything in this article comes down to a handful of habits repeated on every joint. The welder who checks the land in four places, records the number on the fit-up sheet, and refuses to arc up on bad prep will beat the welder with twice the arc time almost every shift. Here is where crews go wrong, and how to catch the error first.

The Mistakes That Keep Showing Up

• Measuring the land in the wrong direction. Root face is measured parallel to the bevel face, not along the plate surface; a land reading 2 mm one way and 3.5 mm the other is a burn-through waiting to happen.
• Checking one point on the circumference. Out-of-round pipe and wall variation mean the land at 12 o’clock says nothing about 6 o’clock — check all four quadrants.
• Trusting a worn bevel gauge. Gouged, bent or cheap gauges drift; compare against a known reference and replace them.
• Leaving a knife edge. A feathered land looks neat and burns through the moment the keyhole opens.
• Over-landing “to be safe.” A heavy land feels conservative but steals penetration and pushes the welder to over-amp, causing undercut and burn-through elsewhere.
• Ignoring hi-lo. Misalignment changes the effective land on one side of the joint, whatever the drawing says.
• Mixing pipe schedules. Two pipes alike on the rack can differ in wall thickness, so the prep is no longer symmetrical.
• Skipping the fit-up sign-off. Undocumented fit-up means nobody owns the dimension when the root fails.
• Over-purging stainless. Excess flow and long purge times cost money and can cause turbulence and coking at the root.
• Blaming the welder for a fitter’s error. Most root-pass failures trace back to fit-up decisions made before the arc starts.

The Checklist, Stage by Stage

1. Drawing and WPS review: confirm bevel angle, root face, root gap, backing and acceptance criteria agree. If they disagree, stop.
2. Material verification: check grade, schedule, wall thickness and heat number against the drawing; measure wall thickness rather than trusting the stencil.
3. Cutting and beveling: verify end squareness, bevel angle and land width; grind or machine away thermal-cut oxide.
4. Cleaning: remove oil, paint, rust and moisture at least 25 mm (1 in) back from the groove on both faces.
5. Fit-up and measurement: set gap with proper spacers, check land at four quadrants, measure hi-lo, confirm uniform root opening.
6. Pre-weld inspection: sign the fit-up sheet, photograph the joint, get inspector buy-in before the root pass.
7. Root pass: follow WPS parameters, watch the keyhole, keep the arc on the leading edge of the puddle. Rod choice and angle control matter here — this all position welding rod guide covers the basics well.
8. Post-root inspection: check suck-back, burn-through, lack of fusion and oxidation before the hot pass.
9. Documentation: record dimensions, welder ID, consumables, purge readings and repairs.

Training That Actually Sticks

Skill comes from coupons, not conversation. Cut test plates with deliberately varied root faces and gaps, weld them, section them, and study the penetration profile. Destructive testing teaches faster than any chart. Shadowing an inspector for a shift is just as useful, because it shows how a small fit-up shortcut looks from the other side of the clipboard. On repairs, learning to gouge cleanly with carbon arc before re-preparing the joint prevents the classic mistake of grinding a defect deeper instead of removing it — see this guide to air arc welding and carbon arc gouging.

Safety at the Bevel and the Root

Weld prep generates as much exposure as welding. Grinding dust, metal fume and shielding gas all accumulate in the same space. The UK Health and Safety Executive changed its enforcement expectation in 2019 after classifying mild steel welding fume as a human carcinogen, and expects exposure controls even outdoors. In the US, OSHA 29 CFR 1910.252 covers ventilation and protection for welding, cutting and brazing, and NIOSH publishes criteria for a recommended standard on welding, brazing and thermal cutting exposure. Practically: extract or ventilate at the arc, wear a respirator where extraction is not possible, protect eyes from grinding and arc radiation, and treat argon and nitrogen purges as asphyxiation hazards in enclosed spaces.

The Bottom Line

Root face is a two-millimetre detail with an outsized vote. It governs how much metal the arc must melt, how much fill the joint demands, and whether the root fuses or falls through. Control the root face and you control the root — the pass every inspector, every radiograph and every pressure test ultimately cares about.

Frequently Asked Questions

What is a root face in welding?

The root face is the flat, unbeveled portion of the groove face next to the joint root. It gives the root pass a controlled edge to fuse into and acts as a heat sink that resists burn-through. It is sometimes called the land, especially on older drawings and pipeline work.

What is the difference between root face and root gap?

The root face is metal left in place at the bottom of the bevel; the root gap (formally root opening) is the space between the two parts at the root. The face resists penetration, while the gap encourages it, so the two dimensions must be balanced against each other and against the bevel angle.

What is a typical root face thickness?

For standard ASME B16.25 pipe ends the root face is 1.6 mm ± 0.8 mm (1/16 in ± 1/32 in). Plate work often starts around 1.5–3 mm (1/16–1/8 in) depending on thickness and process, and thin sheet may use a feather edge. The qualified WPS or drawing always governs.

What happens if the root face is too thick?

Too much land means the arc cannot melt through to the root, which produces incomplete penetration and lack of root fusion. Welders often respond by adding amperage, which can cause undercut, distortion, or burn-through elsewhere in the joint.

What happens if the root face is too thin?

A thin or knife-edge land gives the molten pool very little support, so burn-through, suck-back, and an uneven root bead become likely. It also makes the joint sensitive to small variations in gap and travel speed.

Do fillet welds have a root face?

No. The root face is a groove-weld dimension. Fillet welds are defined by leg size, throat, and convexity rather than by a bevel and land, although fillet welds do have a joint root where the two members meet.

How do inspectors measure the root face?

They use a root face or V-weld measuring gauge, bridge cam gauge, or calipers, and they check several points around a pipe circumference rather than a single spot. The measurement is taken along the bevel face, not vertically, and it is recorded on the fit-up or pre-weld inspection sheet.

Can you weld with zero root face?

Yes, on thin material and some square-groove joints a feather edge is normal and works well. On thicker open-root joints a zero land removes the safety margin against burn-through and demands very tight control of gap, amperage, and travel speed.

Does root face affect how much filler metal you use?

Indirectly, yes. The root face sets how much bevel depth remains, and bevel area grows with the square of that depth, so a smaller land means a larger groove to fill. Groove angle has an even stronger effect on fill volume.

Is root face the same as root penetration?

No. Root face is a preparation dimension on the base metal before welding; root penetration describes how far the weld metal extends through the joint root. The root face strongly influences penetration, but they are different things.