Laser soldering quality inspection is a three-layer process: closed-loop thermal and placement data captured by the soldering cell during the 0.3–3 s heating cycle, 100 % optical inspection (CCD/AOI) of the finished join…
Laser soldering quality inspection is a three-layer process: closed-loop thermal and placement data captured by the soldering cell during the 0.3–3 s heating cycle, 100 % optical inspection (CCD/AOI) of the finished joint, and periodic X-ray plus destructive testing that proves the process is still in control. Because the heat-affected zone is only a few millimetres wide and the joint solidifies before visible oxidation can develop, defects such as cold joints, incomplete wetting and solder-ball satellites are not reliably found by eye — inspection has to begin inside the machine, not at the end of the line. The sections below set out what to measure, which method catches which defect, and what acceptance limits experienced lines actually use.
What in-process signals act as the first inspection layer?
A laser soldering cell with closed-loop temperature control already generates most of the evidence you need. The pyrometer reading is used to modulate laser power in real time, so the recorded curve — not the setpoint — is the primary quality record. Log and SPC these values per joint:
- Peak joint temperature: typically 225–245 °C for SAC305 (melting 217–220 °C), measured at the joint, not at the laser spot.
- Time above liquidus: commonly 0.3–1.0 s. Shorter risks incomplete wetting; longer widens the heat-affected zone and raises intermetallic growth.
- Ramp rate: 100–250 °C/s is typical. A sudden drop in achievable ramp for the same power indicates a contaminated window, a drifted focus, or a changed absorber on the pad.
- Laser power / duty trend: plot it. A slow upward drift over a shift is the earliest warning of lens contamination or feeder misalignment.
- Placement data: solder ball or paste deposit position from CCD alignment, typically held to ±30–50 µm for fine-pitch work; wire feed length repeatability for wire-based processes.
Set a Cpk target of 1.33 or better on peak temperature per joint type. When Cpk falls, the cause is usually mechanical (nozzle wear, ball feeder drift) rather than optical, and it is far cheaper to fix at this stage than after AOI.
Which optical inspection method catches which laser soldering defect?
Post-solder optical inspection is the workhorse: it is the only method fast enough for 100 % coverage. Two decisions matter — sensor type and pixel resolution.
- 2D AOI (typically 15–25 µm/pixel) is sufficient for satellite balls, bridging and gross placement errors, but cannot judge fillet height or wetting angle.
- 3D AOI or confocal scanning (5–10 µm/pixel for 0402/0201 and fine-pitch connectors) measures fillet profile, coplanarity and wetting angle, which is what distinguishes a good joint from a cosmetically similar cold joint.
- Lighting: low-angle ring light for fillet shape and wetting angle; coaxial light for coplanarity and missing solder; a separate UV or colour channel helps separate flux residue from contamination.
| Defect | Optical symptom | First check |
|---|---|---|
| Cold / insufficiently wetted joint | Convex fillet, wetting angle >90°, dull surface, visible pad edge | Peak temperature log and TAL for that joint; pad finish and flux activity |
| Solder-ball satellites | Isolated spheres 30–200 µm within 1–2 mm of the joint | Ball feed alignment and nozzle cleanliness; spatter from excessive power |
| Bridging (0.3–0.5 mm pitch) | Solder web between adjacent leads | Deposit volume, dwell time, and preheat of the connector body |
| Tombstoning / drawbridge | Component lifted at one termination | Asymmetric heating; check spot centring and pad thermal mass difference |
| Heat damage to neighbours | Discoloured or deformed plastic housing within 0.5–1 mm | Spot size, shielding gas coverage, and HAZ verification on a cross-section |
| False call on a good joint | Flagged for residue, shadowing or reflection | Re-teach the library with real golden samples; adjust low-angle light gain |
Keep a golden-board set of at least 10 known-good and 10 known-bad samples per product and re-verify the AOI library after every program change. A false-call rate above roughly 5 % is a process signal, not an AOI problem — it usually means the soldering window is too wide.
When is X-ray or CT genuinely necessary?
X-ray earns its place wherever the joint is hidden: power-device thermal pads, through-hole barrel fill, joints under shields or connector bodies, and package-level soldering in semiconductor assembly. Transmission X-ray at 5–15 µm resolution is adequate for void percentage; CT is a first-article and failure-analysis tool, not a production gate.
Agree void criteria with the customer up front rather than citing a generic figure — 25 % of pad area is a common limit for thermal pads, with no single void crossing the die centreline. Voiding in laser-soldered joints is often low because time above liquidus is short, but it is still driven by paste chemistry, pad design and via-in-pad structure, so it must be measured rather than assumed.
How do you validate the process with destructive testing?
Destructive testing is your calibration of everything above. Run it at first article, at every reel or lot change, after any consumable change (lens, nozzle, feeder, wire spool) and on a fixed interval per shift — typically five joints per critical component type.
- Shear or pull test: establish limits from a first-article study of at least 30 joints, then set the lower control limit at three standard deviations below the mean. Record the failure mode, not just the number — a joint that peels cleanly off the pad failed at the interface, which points to pad finish or flux, not to laser energy.
- Cross-section: verify fillet geometry, wetting angle below 90°, and intermetallic thickness in the 1–3 µm range with no Kirkendall voids or dewetting.
- SEM/EDS: use when a fracture looks brittle or a discoloured joint fails mechanically, to confirm the intermetallic species and rule out contamination.
- Ionic contamination: ROSE test below 1.56 µg/cm² NaCl equivalent per IPC-TM-650 method 2.3.25 when flux chemistry or cleaning changes.
Finally, verify the temperature measurement itself: at the start of each production shift, run a Type-K thermocouple coupon through the cell and compare it with the pyrometer reading. A 10–15 °C disagreement silently invalidates every thermal record you are collecting.
FAQ
Can AOI replace X-ray for laser-soldered joints?
No. AOI only sees surfaces. Any joint under a component body, a shield or a connector housing needs X-ray or electrical test, and through-hole fill can only be judged radiographically.
How often should the closed-loop pyrometer be verified?
Compare against a thermocouple coupon at each shift start, and do a full calibration quarterly or after any optical-path service. Keep the coupons and the comparison records — they are the evidence behind your thermal SPC data.
Why does our AOI reject joints that shear-test perfectly?
Usually flux residue, specular reflection from a shiny fillet, or a library trained on samples that do not match current production. Re-teach with current golden samples, then re-check the false-call rate; if it stays high, the soldering window is probably too wide and producing genuinely variable fillets.
Does laser soldering produce fewer voids than reflow?
Localised heating with a short time above liquidus often gives lower voiding on thermal pads, but the result depends heavily on paste chemistry, via-in-pad design and preheat. Measure it on your own boards before writing a specification.
What is the minimum inspection set for a new automotive product?
Thermal SPC on every joint, 100 % 3D AOI, X-ray on first article and on any hidden-joint feature, plus shear testing at first article and per shift — then document the sampling plan in the control plan so it survives an audit.
If you share your component mix, pitch and joint types, we can map these inspection layers onto a specific laser soldering cell configuration for your line.