The Weld Inspection Guide: Telling a Good Weld From a Pretty One

There is a particular kind of disappointment reserved for the welder who has finally learned to lay a bead that looks like a stack of dimes, and then cuts one open. Evenness is a measure of hand control. It is not a measure of fusion, of size, of filler choice, or of whether the weld reached the root of the joint. Inspection exists because those are different questions, and because the answer to the important ones is frequently not visible from outside.

What inspection is actually for

Inspection answers one question: does this weld meet the requirement? That framing matters, because it means there is no such thing as a good weld in the abstract. There is a weld that satisfies a specified requirement, or one that does not. A bead that would be rejected on a pressure vessel might be entirely appropriate on a garden gate.

The requirement comes from somewhere: a drawing that specifies weld size and type, a welding procedure that specifies how it is made, and a code that specifies what is acceptable. Inspection compares the weld in front of you against that requirement. Without a stated requirement, inspection collapses into opinion, which is roughly where most internet weld critique lives.

Before the arc is struck

It surprises people that a substantial portion of professional inspection happens before welding starts. The reason is economic: a problem caught at fit-up costs a few minutes, and the same problem caught after the assembly is welded, painted and delivered costs considerably more.

• Is the material what the drawing says it is? Wrong grade is not visible and not recoverable.
• Is there a procedure and does it match this joint? Process, filler, gas, position, preheat requirement.
• Is the welder qualified for it? Process, position and thickness range.
• Is the joint preparation correct? Included angle, root face, root gap, alignment. Poor preparation limits what any welder can achieve.
• Is the joint clean? Mill scale, rust, oil, moisture, paint and previous slag all produce defects that get blamed on technique.
• Are the consumables correct and properly stored? Particularly for low-hydrogen electrodes, where storage is a requirement rather than housekeeping.
• Is preheat required, and has it been achieved and measured? Through the thickness, not just on the surface.

For a home shop this translates into a short discipline: know your material, prepare the joint properly, clean it, check the fit before you tack rather than after, and write down what you did when a result comes out well. That last habit is an informal welding procedure and it is what turns a lucky result into a repeatable one.

Visual inspection: what it catches

Visual examination is the most-used inspection method in the world by a very wide margin, and it deserves more respect than its simplicity suggests. It catches a large fraction of the defects that actually occur, it costs almost nothing, and it can be applied to every weld rather than to a sample.

Good visual inspection needs adequate light, a clean weld — slag and spatter hide exactly what you are looking for — and something to measure with. Doing it properly means removing slag first, then looking, with a gauge in hand rather than an impression in mind.

Reading a fillet weld

The fillet is the most common weld in fabrication and the one most worth being able to read. The vocabulary is small and, once you have it, it makes every inspection conversation legible.

• Leg length — the distance from the joint root along each fused face. Both legs are measured; an unequal fillet where the drawing calls for equal is a nonconformity.
• Throat — the shortest distance from the root through the weld to the face. This is the dimension that carries load, and a concave face reduces it even when the legs measure correctly.
• Toe — the junction of weld face and base metal. Undercut and overlap both live here, and a sharp toe transition is a stress concentration regardless of whether either defect is present.
• Root — where the two members meet at the back of the fillet. Whether the weld actually reached it is invisible from the face.
• Face profile — flat, convex or concave. Codes commonly limit both reinforcement and concavity, because both extremes cause problems.

What visual inspection cannot see

This is the part that matters most and gets least airtime. A number of the most serious weld defects produce no surface indication at all.

Destructive testing you can do on scrap

The single most useful thing a self-taught welder can do is destroy their own test welds. It costs an offcut and twenty minutes, and it answers the questions visual inspection cannot. The key is that these are coupons — deliberately made to be wrecked — welded on the actual material, at the actual thickness, in the actual position as the real job.

Break test. Weld a fillet on a T-joint coupon, then bend the upstanding member over until the joint fails. Look at the fracture surface. You are looking for evidence that the weld penetrated to the root, for unfused areas showing as smooth regions rather than torn metal, and for porosity or inclusions. This is the fastest, crudest and most informative test available to a home shop.

Bend test. Prepare a coupon containing a groove weld, dress the reinforcement flush, and bend it so the weld goes into tension. Defects open up. The formal version specifies coupon dimensions, bend radius and acceptance criteria; the informal version still shows you lack of fusion at the toe or root with brutal clarity.

Macro-etch. Cut a section through the weld, polish the cut face, and apply an appropriate etchant. The weld profile, penetration depth, fusion at the root and the extent of the heat-affected zone become visible. Etchants are corrosive chemicals — read the Safety Data Sheet, use appropriate protection and ventilation, and dispose of them properly.

Dye penetrant and magnetic particle

These are the two non-destructive methods genuinely accessible outside an industrial setting, and both find surface-breaking defects that the eye misses.

Dye penetrant works on any non-porous material, including aluminium and stainless. The surface is cleaned thoroughly, penetrant is applied and left for a dwell time so capillary action draws it into any surface-breaking discontinuity, the excess is carefully removed, and a developer is applied that draws the trapped penetrant back out as a visible indication. Cleaning discipline is everything — inadequate cleaning before or over-aggressive cleaning after both destroy the result.

Magnetic particle works only on ferromagnetic materials. A magnetic field is induced in the part and fine magnetic particles are applied; where a discontinuity disturbs the field, particles gather and reveal it. It finds surface-breaking and some slightly subsurface defects, and it is faster than penetrant on steel — but it cannot be used on aluminium, austenitic stainless or other non-magnetic materials.

Ultrasonic and radiographic testing

These are the volumetric methods — the ones that actually see inside the weld — and both are firmly in qualified-technician territory. They are worth understanding because they explain what serious inspection involves and why it is not something a home shop replicates.

Ultrasonic testing sends high-frequency sound into the material and interprets the reflections. Discontinuities reflect sound differently from sound metal, so internal defects can be detected and located. It requires calibrated equipment, appropriate couplant, and considerable interpretive skill — the instrument produces signals, and turning signals into conclusions is the qualification.

Radiographic testing passes ionising radiation through the weld onto a detector, producing an image where internal features appear as density differences. It is excellent at finding volumetric defects such as porosity and inclusions. It also involves radiation sources subject to strict licensing, controlled areas and specialist training — there is no informal version of it.

Building an inspection habit

For everyday shop work, inspection does not need to be a formal process. It needs to be a routine that runs automatically, so that problems surface while they are still cheap.

1. Clean before you look. Slag and spatter hide the exact features you are inspecting. Chip and brush first.
2. Light it properly. Most missed surface defects are missed because nobody put a decent light on the joint at a low angle.
3. Measure, do not judge. A gauge against the drawing beats an impression, every time.
4. Inspect between passes, not just at the end. A defect buried under three subsequent passes is a grinding job; the same defect caught immediately is a thirty-second fix.
5. Look at the toes and terminations first. Undercut, overlap and crater cracks cluster there, and those are the defects most likely to matter.
6. Keep test coupons. Break one whenever you change process, material, thickness or a significant setting.
7. Come back the next day on anything critical. Delayed cracking is real and a same-day inspection cannot rule it out.
8. Write down what worked. Material, thickness, process, filler, gas, settings, position. That record is what makes a good result reproducible.

And keep the underlying distinction in view. Consistency of appearance is genuine evidence of hand control, and hand control does correlate with sound welding — it is not nothing. But it is evidence about the welder, not proof about the weld. The point of learning inspection is not to become suspicious of your own work. It is to know which questions your eyes have answered, which ones they have not, and when the answer is worth paying someone qualified to find.