Fillet Weld Spec Table

In-Depth Guide

Leg, throat and effective throat — where a = 0.7s comes from and when it fails

For an equal-leg, right-angle fillet, the throat a is the height of the isosceles triangle inscribed in the weld cross-section: a = s/√2 ≈ 0.7s (s = leg length). Strength is calculated on this throat section, so you must know whether a drawing gives a (throat) or z/s (leg) — reading the same number as one versus the other changes the section area by about 1.4× (~40%) and throws off the strength badly. The 0.7 factor holds only for equal-leg right angles; for unequal-leg fillets or obtuse/acute joints you must recompute the actual inscribed height. Deep-penetration automatic welds count the penetration as effective throat, so a can be larger.

Reading the weld symbol — arrow side vs other side, below/above the line (KS B ISO 2553)

When the fillet triangle sits below the reference line, weld the side the arrow points to (arrow side); when it sits above, weld the opposite side. Symbols on both sides mean a double fillet. Reading above/below the wrong way and welding the wrong face is a classic shop error, and it is critical on joints welded from one side only. The tail of the symbol carries the process, leg size and other notes. The ISO/KS arrow-side convention can differ from the US system (AWS), whose above/below habit is reversed, so confirm which standard the drawing follows first.

Continuous vs intermittent fillets — the 6-50@150 notation and where intermittent is banned

The intermittent fillet notation 6-50@150 means a 6 mm leg, 50 mm of weld, repeated on a 150 mm pitch (center to center). Offsetting the welds on the two sides gives a staggered intermittent weld. Intermittent welding cuts heat input, distortion and weld volume, so it suits statically loaded structures. But every start and stop of an intermittent bead creates a stress raiser and fatigue crack origin, so continuous fillets are mandatory on members under cyclic load and on sealed tanks or pressure vessels that must be leak-tight. In corrosive service, moisture collects in the gaps and causes crevice corrosion, so continuous is safer there too.

Minimum and maximum leg size — why too small, and too big on thin plate, both fail

The minimum leg is a lower limit for crack prevention, not strength. Thick base metal pulls heat away fast (quenching) and can crack before the bead solidifies, so the thicker the base metal, the larger the minimum leg needed to put in enough heat. The maximum leg, by contrast, is set at the edge of thin plate: if the leg approaches the edge thickness, the base metal melts away (edge melt-down) and the corner collapses. So for a fillet running along a member edge, the maximum leg is capped at the plate thickness (about plate thickness − 1.5 mm on thin plate). In short, leg size is not a number you raise for strength alone — it is set between quench cracking (lower) and melt-down (upper).

Defects and inspection — root gap, undercut, convex/concave and the fillet gauge

A large-looking leg is not automatically strong. If the root gap (the space between the faying surfaces) opens up, the surface looks big but the real throat is short and strength drops. Pouring in too much heat to build a bigger leg increases angular distortion, while undercut (a groove melted into the base metal at the toe) and overlap (the toe folding over without fusion) are both fatigue crack origins — grinding the weld toe smooth extends fatigue life. To inspect, set a fillet (weld) gauge against the bead to measure leg and throat and whether the face is convex or concave. Concave beads tend to be short on throat and excessive convexity raises toe stress concentration, so aim for a slightly flat to mildly concave profile.