How to Judge Laser Welding Seam Quality
Judge a laser weld against its specified surface profile, fusion geometry and service requirements. Start with the drawing and process record, inspect the full seam, then use sections, suitable non-destructive testing and performance tests where required. A smooth bead alone cannot establish internal quality.
1. Define what this joint must achieve
The seven checks below organize the inspection decision; they are not seven mandatory tests for every weld. A decorative enclosure, a loaded bracket and a sealed tank need different evidence. Decide the required result before selecting the inspection method.
Write down the exact material and condition, thickness of each member, joint detail, required penetration or fused width, permissible surface profile and relevant product tests. Add the inspection method, acceptance limits, sampling plan and person authorized to release the work. “Looks good” leaves too much open to interpretation.
Keep three documents distinct: the drawing defines the joint; the welding procedure specification (WPS) describes how to make it; the inspection plan states how compliance will be checked. Where procedure qualification is required, its test record provides supporting evidence for the qualified range.
For procedure qualification, ISO 15614-11:2025 covers electron and laser beam welding of metallic materials. Its ISO record lists a corrected version dated May 2026. Apply the edition and corrections required by the governing contract or product rules.
2. Check the joint and process record
Before adjusting laser power, confirm that the part and setup match the qualified procedure. A wrong coating, shifted joint or contaminated protective window can change the weld while the displayed power setting stays the same.
Before welding
- Identify the grade, material lot, condition and coating.
- Check edge preparation, surface cleanliness and the overlap interface.
- Measure gap, mismatch and part position against the specified limits.
- Confirm fixture, clamp sequence, backing and access to the seam.
During welding
- Record program version, power or pulse settings, travel speed and focus.
- Record wobble settings, filler wire and shielding where used.
- Retain alarms, tracking data and significant maintenance changes.
- Link the record to the actual part, seam and production time.
Use the evidence to narrow the cause. If a defect starts after an optics change, verify that change. If it occurs only at one fixture position, check alignment and fit-up there. Changing several variables together makes the next result harder to explain.
3. Inspect the whole visible seam
A visually acceptable laser seam follows the intended path and meets the specified profile and continuity limits. Examine the start, steady portion, corners, overlaps and stop, plus the root where accessible. Do not judge from one attractive section.
Make observations repeatable
Use an approved cleaning method that preserves indications. Provide suitable lighting and access, and use the required gauges or calibrated measuring system. ISO 17637:2016 covers visual examination of fusion welds and can also support pre-weld joint examination.
Record actual dimensions and locations: width, undercut depth, underfill, reinforcement, mismatch and start/stop features as required by the drawing. A weld toe is the boundary where the visible weld surface meets the parent material.
Color and spatter are clues. Their meaning depends on the material and application. Neither a silver surface nor an even ripple pattern gives a universal pass/fail verdict.
| Observation | What to establish | Useful next action |
|---|---|---|
| Fine line or crack-like mark | Whether it is a discontinuity and how far it extends. | Hold the affected work for evaluation; use appropriate surface testing or sectioning. |
| Pits or holes | Whether the opening connects to a larger void or lost section. | Map the positions and select internal examination where required. |
| Undercut or underfill | Undercut is a groove at an edge; underfill is a low weld face. Measure the specified dimensions. | Compare with the limits; check section geometry if the effective joint size is uncertain. |
| Irregular width, missed path or mismatch | Where geometry departs from the approved condition. | Compare part location, tracking, speed and fit-up records at that position. |
| Crater or abrupt restart | Continuity and local fusion at the transient. | Include that location in the section or functional-test plan, as applicable. |
On a small screen, swipe the table sideways to read all columns.
A symptom does not identify one cause. TWI’s laser-weld defect guidance describes cracking related to alloy and weld shape, and porosity associated with contamination, shielding and keyhole instability. Zinc-coated lap joints can also develop blowholes from trapped coating vapor. For a detailed profile investigation, use the laser-welding undercut guide.
4. Verify penetration and fusion at the required interface
A prepared cross-section is a direct way to examine the joint during development, qualification or scheduled verification. It shows the sampled plane. One section cannot prove that the same condition continues along the entire seam.
Penetration and fusion answer different questions. The required depth comes from the joint design. Fusion concerns whether the intended surfaces actually melted together. On a lap joint, a deep-looking bead in the upper sheet may still leave the lower sheet unjoined.
Conceptual sections, not measured welds or acceptance templates. Colors show the melt zone and parent sheets; the diagrams do not show the heat-affected zone or a specified coating thickness.
Record the section, not just its photograph
Identify its location and orientation. Measure the specified depth, interface width or effective weld size; record the root shape and observed pores, cracks or unfused regions. Distinguish the melted fusion zone from the heat-affected zone (HAZ), which was heated without melting.
Prepare the specimen without masking the feature being examined. Cutting damage and uneven polishing can distort interpretation. Macro examination addresses overall geometry; micro examination and hardness testing answer different material-property questions. See Struers’ weld-section guidance and ISO 17639:2022 for their respective preparation and examination scopes.
For full versus partial joint penetration and measurement terminology, see what weld penetration means. A deliberately specified partial-penetration joint should not be confused with failure to achieve a full-penetration requirement.
5. Select NDT for the flaw, material and geometry
Non-destructive testing (NDT) examines the part without cutting out a test section. Choose it for the expected flaw and demonstrate that the technique can detect that flaw in the actual joint. Access, thickness, surface condition, orientation and detection sensitivity all matter.
ISO 17635:2025 provides general selection guidance. Keep the testing technique, testing extent and acceptance criteria explicit; a quality-level label alone does not define a complete examination.
| Method | Useful evidence | Main limitation |
|---|---|---|
| Visual testing (VT) | Accessible surface condition and measured geometry. | Cannot establish hidden fusion, internal pore content or material properties. |
| Penetrant testing (PT) | Discontinuities open to a clean, nonporous surface. | Does not reveal closed internal flaws. ISO 3452-1 gives principles, not weld acceptance criteria. |
| Magnetic particle testing (MT) | Surface indications in suitable ferromagnetic material. | Confirm magnetic material suitability; do not select it merely because the material is called “steel.” See ISO 17638. |
| Radiographic testing (RT) | Internal conditions such as porosity, where access and sensitivity are suitable. | Tight planar flaws can be difficult to detect at unfavorable beam angles. |
| Ultrasonic testing (UT) | Internal reflectors examined with a validated probe and setup. | Thin, narrow or complex joints can require specialized techniques; do not assume conventional UT covers them. |
| Prepared section (destructive) | Fusion geometry and visible imperfections at the cut plane. | Consumes a specimen and samples a location, rather than the full seam volume. |
Swipe horizontally on mobile. Method suitability does not itself establish an acceptable result.
Why conventional UT is not a default for thin seams
ISO 17640:2018 describes manual ultrasonic testing for low-attenuation metallic welds at least 8 mm thick, primarily full-penetration ferritic joints. It also provides for specified special applications. This scope is not a physical claim that ultrasound cannot inspect thinner material; it means another application-specific technique must be justified rather than assumed.
For RT, TWI’s radiographic detectability research shows why flaw opening, orientation and penetrated thickness affect detection. “Use X-ray” is therefore incomplete advice without a defined technique and required sensitivity.
6. Test the property the product depends on
Good geometry and an acceptable NDT result support quality assessment, but they do not directly measure every service property. Select the test that represents the product’s load or function, with a defined method and acceptance criterion.
- Static load: use the appropriate tensile, shear or joint test. Record specimen geometry, loading direction, result and fracture location.
- Ductility or exposed imperfections: a specified bend test can reveal openings or limited ductility. Specimen orientation and bend conditions matter; see TWI’s bend-test explanation.
- Metallurgical behavior: use the required microstructure, hardness or other material tests where the alloy and service make them relevant.
- Sealing: define the leak-test method, test conditions and permissible leak rate for the completed assembly.
- Electrical or cyclic service: evaluate the specified resistance, temperature rise or fatigue performance under representative conditions.
A high tensile result does not establish leak tightness, and a leak test does not establish fatigue life. Likewise, failure in the parent material is evidence from that specimen and loading arrangement; it is not proof of every other service condition.
7. Confirm repeatability before production release
One successful coupon demonstrates one tested condition. Production also needs control over material, fit-up, equipment and the verification plan. Check repeated parts and the relevant tolerance conditions before treating a setting as a reliable production window.
Validate what the monitor means
Photodiodes and cameras can observe light signals or melt-pool features and correlate them with weld behavior. Their output is process evidence. TWI explains these monitoring principles and how quality is inferred from the observed signals.
Validate the system against appropriate sections, NDT or performance results for the actual process. Include relevant unacceptable conditions, review missed indications and false calls, and reassess the relationship after material, geometry, optics or sensor changes. An unchanged signal should not override a failed inspection.
Define the response to an abnormal result
The control plan should identify the stop or hold conditions, affected production interval, confirmatory examination, release authority and repair rules. Preserve the suspect part and its data. Investigate the changed variable before accepting further output.
Minimum information for a useful weld-quality record
- Part and seam ID; material lot, coating and thickness stack.
- Drawing and WPS revision; actual setup and program.
- Inspection method, coverage, locations, equipment and limits.
- Measured results, section or test references, and deviations.
- Disposition, responsible person and any follow-up verification.
Sampling and requalification triggers belong to the applicable specification and validated control plan. There is no universal “test every N welds” interval.
Questions when the result is unclear
Can I grind or polish the weld before inspection?
Only as permitted by the manufacturing and inspection sequence. Finishing can remove profile information, smear a surface opening or alter weld size. Preserve the required as-welded observations first, and evaluate the finished condition too where the specification requires it.
What if the weld passes visual inspection but fails a leak or strength test?
Keep the failed result and identify the affected work. Verify the test setup, then investigate joint geometry, internal condition and material properties as relevant. Passing the visual check does not cancel a separate required test failure. Release needs an authorized disposition against the actual requirements.
To discuss a laser-welding application with Oceanplayer Laser, send the material and coating, thickness stack, joint drawing, expected gaps, seam path and acceptance criteria. Include failed samples or existing inspection results. Agree which sections and product tests the evaluation should include.