Types of Welding Joints and When to Bevel Them

The Five Basic Types of Welding Joints

There are five basic types of welding joints: the butt joint, the T-joint (tee joint), the corner joint, the lap joint and the edge joint. AWS A3.0, the American standard for welding terms, defines each one by how the members sit relative to each other, not by the weld that joins them.

The five joints differ in where the parts touch and at what angle. The diagram of the five welding joint types on this page shows them side by side in cross-section.

• Butt joint: two members lie in roughly the same plane and meet edge to edge.
• T-joint: the edge of one member meets the face of the other at roughly a right angle, forming a T.
• Corner joint: two members meet at or near their edges at roughly a right angle, forming an L.
• Lap joint: two members overlap in parallel planes.
• Edge joint: the edges of two or more parallel or nearly parallel members are joined.

Drawings also show variants of these five. A skewed T-joint has its stem at an angle other than 90°, and two T-joints on opposite faces of one plate form a cruciform joint.

The table below summarizes the five joints of welding, the welds most often made in them and where they are typically used. Short definitions of the terms are in the welding terms glossary.

Joint Type vs Weld Type

A joint type describes how the parts are arranged; a weld type describes the weld that joins them. A T-joint, for example, can be welded with two fillet welds, with a bevel groove weld from one side, or with a double-bevel groove weld that reaches complete joint penetration. The joint stays the same, while cost, strength and edge preparation change with the five weld types below.

• Groove welds: made in a groove between the members or in a prepared edge. The groove can be square, V, bevel, U, J, flare-V or flare-bevel, single or double, and designed for complete or partial joint penetration.
• Fillet welds: roughly triangular in cross-section and laid in the corner between two surfaces. They need no edge preparation.
• Plug and slot welds: made through a round hole or an elongated slot in one member of a lap or T-joint.
• Spot and seam welds: made between overlapping members, usually by resistance welding on sheet.
• Edge welds: made on the edges of an edge joint or a flanged corner.

The structural difference between a groove weld and a fillet weld is the throat, the dimension that carries the load. A groove weld with complete joint penetration has a throat equal to the thickness of the thinner part. An equal-leg fillet between members at 90° has a theoretical throat of about 0.7 times its leg size. ISO drawings can state that throat directly as the a dimension, while AWS drawings give the leg size.

Watch the wording on international drawings. ISO 2553 and European practice call a groove weld a butt weld even when it sits in a T-joint, so a single-bevel butt weld may join a stiffener to a flange. AWS keeps butt for the joint and groove for the weld. The guide to weld bevel and groove types covers the groove shapes and their dimensions.

Butt Joint Welding: From Square Edges to Double-V

A butt joint joins two members that lie in approximately the same plane, and it is welded with a groove weld that passes the force straight through. Welded butt joints with complete joint penetration, ground flush where fatigue matters, are the standard splice for plate, beams and vessel shells.

Plate thickness decides whether the edges stay square or are beveled. An arc fuses only a limited depth into a square edge, so beyond a few millimeters the edges are opened into a groove that lets the arc reach the root. ISO 9692-1, the international reference for joint preparation of steel, lists typical preparations for each thickness range. The table below condenses those ranges into a first selection guide for butt joint welding in steel.

The ranges overlap on purpose, because process, position, access and code decide between them. They are guidance, not a procedure: the welding procedure specification (WPS) and the applicable code govern the final groove angle, root face and root opening. The reasoning behind the thickness limits is explained in why weld beveling is necessary for thick plate welding.

Access is the second decision. A single-V groove is welded from one side, usually onto a backing bar or with a controlled open root, and it tends to pull the plates out of flat as the groove fills. A double-V groove needs access to both faces and a back-gouged root, but it uses less weld metal and the shrinkage of one side offsets the other. When butt-welded parts differ in thickness, the thicker part is usually tapered down to the thinner one, with the slope set by the design code.

T-Joint Weld: Fillet or Bevel Groove

A T-joint, or tee joint, is welded with fillet welds in most cases and with a bevel groove weld when the stem must carry its full strength across the joint. The stem is the member whose edge meets the face of the other member: the web of a girder or the gusset of a bracket. Fillets need no preparation, which makes them the default for stiffeners, brackets and most static connections.

Fillet welds have a cost limit. The weld metal in a fillet grows with the square of its leg size: a 6 mm (1/4 in) fillet holds about 18 mm² of weld metal in cross-section, a 12 mm (1/2 in) fillet about 72 mm², four times as much. On thick stems, the fillets needed to develop the stem strength become so large that a groove weld uses less metal and less heat. Double fillets, one on each side of the stem, are also far more reliable than a single fillet, which can open at its root when the stem bends.

For a groove-welded T-joint weld, only the stem is prepared. A single-bevel groove is welded from one side, as a partial joint penetration weld or onto backing for complete penetration. A double-bevel (K) groove is beveled from both faces of the stem and welded from both sides with the root back-gouged. It is the usual detail for a complete joint penetration T-joint in thick plate.

Groove welds in T-joints are commonly finished with a small reinforcing fillet that eases the change of section at the weld toe. Whether a tee needs complete or partial penetration is a design decision, covered in the guide to complete vs partial joint penetration. Skewed T-joints need extra care on the acute side, where access to the root is limited.

Corner Joint Welding: Open and Closed Corners

A corner joint joins two members that meet at roughly a right angle with their edges together, forming an L. Unlike a T-joint, the weld sits at the outside corner of the assembly, which makes corner joints the basis of box sections, frames, tanks and sheet metal enclosures. How the edges are placed decides whether corner joint welding needs any edge preparation.

• Open corner: the plates touch only at their edges and leave a V-shaped groove on the outside. The weld fills that natural groove without beveling, which suits sheet and light plate.
• Half-open corner: one plate overlaps part of the other plate's edge. It is easier to fit than a full open corner.
• Closed corner: the edge of one plate sits flush against the face of the other. The outside is flat, so the joint is welded with a fillet on the inside or, on thicker plate, with a bevel groove on the outside.

On thick plate, a closed corner is prepared like a T-joint: the member whose edge butts against the other gets a single-bevel or J groove. Corner joints in box sections are often reachable from the outside only, so they are designed as single-sided groove welds with backing or as partial penetration welds. Weld shrinkage pulls the corner out of square, so tacks, fixtures and weld sequence are planned before the first pass.

Lap Joint Welding

A lap joint joins two members that overlap in parallel planes, and it is usually welded with fillet welds along the edges of the overlap. It needs no edge preparation and tolerates variation in part size, so lap joint welding is common in sheet metal assemblies, gussets, repair plates and splice straps. Plug and slot welds through holes in the top member add weld area where the edges alone are too short.

The weakness of a lap joint is its load path. The two plates are offset by their own thickness, so a tension load bends the joint, and a lap welded on one edge only can open like a hinge. Fillets on both edges of the overlap restrain that rotation and are standard for load-carrying laps. For the same reason, lap joints rate well below ground butt joints under fatigue loading.

A few rules of practice apply to most structural laps. They come from US structural practice and are worth checking against the code edition you work to.

• Overlap length: AWS D1.1 and the AISC Specification call for an overlap of at least five times the thickness of the thinner part, and not less than 25 mm (1 in), in load-carrying lap joints.
• Edge fillets: a fillet along the edge of the top plate is limited by that plate's thickness. On thicker plate it is kept slightly smaller, so the original edge stays visible for inspection.
• Crevices: the gap between overlapping plates can trap moisture, pickling acid or cleaning fluid. Seal-weld the overlap where corrosion or coating matters, and vent large overlaps before hot-dip galvanizing.

Edge Joint Welding

An edge joint joins the edges of two or more parallel or nearly parallel members, and the weld is made along those edges. It is a sheet metal joint more than a structural one: flanged edges of thin sheet, ducts, light containers and seams that only need to seal. On thin sheet, the edges are often melted together by TIG welding with little or no filler metal.

Thicker members in an edge joint can be grooved with a square, bevel, V, J or U preparation along their edges. Even so, the weld sits at the edge while the members continue side by side, so any load that pulls the plates apart puts the weld root in tension. Edge joints are therefore avoided where the joint carries significant tension, bending or cyclic load. If an edge joint turns up in a load-carrying position during design review, a butt, T- or lap joint usually carries the load more directly.

How to Choose a Welding Joint

Choose a welding joint by working through the load, the plate thickness, access to both sides, distortion and the cost of preparing the edges. The part geometry often fixes the joint family: members in one plane meet in a butt or lap joint, members at an angle in a T- or corner joint. Within that family, load and thickness decide the weld and the edge preparation.

The ratings below are relative and assume steel plate joined by arc welding. The table compares the joints on the criteria that usually decide the choice.

A practical sequence works for most fabrication drawings. Each step narrows the choice made in the step before.

1. Define the load: magnitude, direction relative to the weld, static or cyclic.
2. Let the geometry choose the joint family, then pick the weld that carries the load with the least weld metal.
3. Compare thickness with the square-edge limit to see whether a bevel is needed.
4. Check access: with one side only, plan backing or a single-sided groove.
5. Check inspection: a weld that needs ultrasonic or radiographic testing needs a geometry that allows it.
6. Weigh the cost of beveling against the extra welding time of a larger fillet.

Edge Preparation by Joint Type

Butt joints above the square-edge limit, groove-welded T-joints and thick closed corners need a beveled edge, while lap joints and most edge joints need only clean, straight edges. The table below shows which member is prepared in each case and when a bevel becomes necessary.

On single-bevel and J grooves, the drawing must show which member gets the bevel. In AWS A2.4 the arrow points to that member, with a break in the arrow line where it is not obvious, as explained in the welding symbols chart. Beveling the flange instead of the web is a mistake that is expensive to correct after cutting.

Edge condition matters as much as geometry. Thermally cut edges carry slag, dross and an oxide layer that can end up in the root, so plasma and oxy-fuel cut parts are cleaned before or during beveling. Straight bevels for V, Y, single-bevel, double-V and double-bevel grooves can be produced on a plate beveling machine with a repeatable angle and root face along the full edge.

TFON Bevel Wizard machines bevel plate and sheet from 4 to 100 mm thick at a working angle of -30° to +80°. The Bevel Wizard TF-C 8010 L has a sliding table for parts up to 12,000 mm long, and the TF-F 8030 L has a fixed table for parts 60 to 3,000 mm long. J and U grooves need a root radius, which is usually machined, so discuss them with TFON before planning them on a straight-bevel machine. For the process overview, see weld bevel preparation.

Common Joint Design Mistakes

Most welded joint problems are fixed on the drawing, not in the shop. The mistakes below appear again and again in fabrication drawings, and each one costs weld metal, distortion or a failed inspection.

• Specifying CJP by default: a complete penetration groove weld where a fillet or partial penetration weld would carry the load adds beveling, back gouging, weld metal and testing.
• Single-sided welds under bending: a single fillet on a T-joint or a lap welded on one edge can open at the root. Weld both sides where the joint bends.
• Single-V on thick plate that can be turned: it needs nearly twice the weld metal of a double-V and distorts more.
• No access for the welder: a groove inside a closed box or a fillet in a tight acute angle cannot be welded properly. Check torch access at the design stage.
• Bevel angle confused with groove angle: beveling each plate of a 60° V-groove to 60° doubles the groove angle and roughly triples the weld metal. State which angle the drawing means.
• Through-thickness pull on thick plate: large welds on T- and corner joints load the face plate through its thickness, which can cause lamellar tearing in susceptible steel. Keep welds no larger than needed and consider steel with guaranteed through-thickness properties.
• Loose fit-up on fillet welds: a gap between stem and face reduces the effective throat. Codes such as AWS D1.1 limit that gap and require a larger leg when it exceeds a small value.

Every weld also leaves residual stress and some distortion. Where dimensional stability or service conditions demand it, the guide on stress relief after welding explains the options.