Judging Laser Weld Appearance: What Passes and What Does Not

Appearance is the first quality gate a laser weld meets, and the most argued-about one. A seam can look imperfect and hold full strength; it can look beautiful and hide a lack of fusion. Without written criteria, acceptance drifts toward whoever inspects last - the operator, the inspector, or worst of all the end customer. This article sorts the surface defects you will actually see into cosmetic and rejection categories, and shows how to write the thresholds down so the argument ends before it starts.

The Defect Families

Defect	What it looks like	Usually means
Undercut	A groove melted into the parent metal at the bead edge	Power or speed mismatch; can reduce fatigue life - measure it
Porosity (surface)	Pinholes along or beside the bead	Gas, contamination or keyhole instability - check depth
Spatter	Droplets frozen on the surface around the bead	Parameter window or shield gas issue; usually cosmetic if bonded weakly
Lack of fusion	A seam present but not bonded through	Rejection every time - appearance can be deceptively good
Excessive penetration / sinkage	A heavy bead or sunken root	Over-power; check for burn-through on thin sheet
Cracks	Fine lines at bead centre or toe	Rejection every time, regardless of size

Cosmetic versus Strength-Relevant

The sorting rule that survives audits: classify defects by what they do to the joint, not by how they look. Spatter that sits on the surface and flicks off is cosmetic, however ugly. Undercut is structural - it removes parent metal exactly where bending stress concentrates. Porosity is a depth question: surface pinholes may be cosmetic, but porosity that reaches into the load path is a rejection. Heat tint sits on the cosmetic side unless the finish tier forbids it, and even then it is a surface specification rather than a strength one. And lack of fusion plus cracks are always rejections, because appearance tells you nothing about how bad they are. When in doubt, a cross-section settles the argument in twenty minutes.

Writing the Thresholds Down

A usable appearance standard is one page per part family, expressed in measurable terms. A typical set of entries:

Undercut: none deeper than 10% of sheet thickness on functional welds; any depth on pressure-tight joints is a rejection.
Porosity: no open pore larger than 0.3 mm diameter; no more than 3 pores per 100 mm of seam.
Spatter: no spatter bonded to functional sealing faces; loosely attached spatter acceptable if it passes the specified wipe test.
Bead width and reinforcement: within the qualified range recorded at process sign-off.
Heat tint: light straw acceptable; heavy blue/black tint on stainless visible surfaces is a rejection where the finish tier demands it.

Two rules make the standard real: every limit must be measurable with the tools the line actually has (loupe, gauge, lamp), and every limit must be the same in the work instruction, the training and the customer agreement.

Lighting and Tools: Half the Judgement

Most appearance disputes are lighting disputes in disguise. A seam judged under a single overhead strip light looks different from the same seam under a directional lamp held at a shallow angle - which is exactly how undercut and fine porosity reveal themselves. The standard should therefore fix the viewing conditions, not just the limits: a defined lamp angle and intensity, a defined viewing distance, a loupe of a stated magnification for borderline calls, and a flat reference board for judging heat tint. Photographs of accepted and rejected seams, taken under the same lighting, convert the standard from an opinion into a gallery - and new inspectors learn it in an afternoon instead of a quarter. A shared visual reference also protects the standard when personnel change, which over the life of a line is the more common event.

Sampling versus Full Inspection

Appearance inspection is fast enough to run 100% on critical parts and sampled elsewhere. A pragmatic allocation: 100% visual on visible-surface and safety-relevant welds; a fixed sample (first article plus hourly checks) on functional internal welds backed by per-weld monitoring data; and cross-section audits on a weekly or per-batch rhythm to confirm that what looks right under the lamp really is fused through. Shift the allocation only on evidence - a defect escape tightens the sample, a month of clean records relaxes it.

Rework Criteria and Records

Decide rework rules before the first dispute. Grinding a cosmetic seam smooth is one thing; re-welding over a rejected joint is another, because each thermal cycle alters the joint. Sensible defaults: surface-only defects (spatter, light tint) may be reworked cosmetically with a record; structural defects (undercut beyond limit, porosity in the load path, cracks, suspected lack of fusion) are either cut out and re-welded per a qualified repair procedure, or scrapped. Every rework gets a mark and a record row - parts that fail twice do not get a third attempt on the line.

The records matter as much as the rules. A simple tally - defect type, part, shift, disposition - becomes a defect map within weeks, and the map points at its own causes: one joint family, one material batch, one time of day, one electrode or optic approaching the end of its life. Fix the cause and the appearance argument disappears along with the defect, which is the only outcome everyone involved actually wants.

Conclusion

Laser weld appearance is a language, and an appearance standard is its dictionary: defect families, measurable thresholds, controlled viewing conditions, a sampling plan and rework rules on one page. Write it once, train it, and apply it the same way on every shift - then the strongest opinion on the line stops being the acceptance criterion, the standard does, and everyone downstream of the welds can trust what they see.

Judging Laser Weld Appearance: What Passes And What Does Not