PCB Cold Solder Joint — Failure Analysis & 8D

What Is a Cold Solder Joint?

A cold solder joint is a connection that formed without enough heat or with contaminated surfaces, so the intermetallic compound (IMC) layer never developed properly. The joint may pass a functional test and then fail intermittently in the field as thermal cycling opens a microscopic crack. Cold solder is a common cause of no-fault-found returns in automotive and industrial electronics, and it is almost always a process-control problem rather than a component problem.

Symptoms

• Dull, grainy or porous joint surface compared with a shiny, smooth good joint
• Intermittent electrical connection — passes at test but fails in the field
• Joint resistance above 50 mΩ
• Cracks at the pad-to-solder or lead-to-solder interface under a microscope
• Fillet that does not wet properly or pulls away from the pad

Affected Components

• PCBAs in automotive, industrial and consumer electronics
• Through-hole and surface-mount (SMT) solder joints
• BGA, QFN and SOIC packages and discrete component joints
• ENIG, HASL, OSP and immersion-tin pad finishes

Detection Methods

• Visual inspection (10-40x): check gloss, wetting angle and fillet shape
• X-ray: reveals hidden cold solder under BGA and QFN packages
• Automated optical inspection (AOI): machine vision checks shape, colour and reflectivity
• ICT / flying probe: measures joint resistance and detects intermittency
• Cross-section plus SEM: measures the IMC layer thickness and integrity

Possible Causes

• Insufficient heat: peak reflow temperature too low or time above liquidus (TAL) too short, so the IMC layer never forms fully.
• PCB contamination: copper pad oxidation, especially on OSP finishes, prevents proper wetting.
• Component lead oxidation: terminals that have been in stock too long oxidise and resist wetting.
• Thermal mass imbalance: large copper planes act as heat sinks, so small pads never reach reflow temperature.
• Vibration during cooling: conveyor vibration or handling while the joint is still plastic.
• Insufficient flux activity: low-activity no-clean flux cannot remove surface oxides.

5-Why Example

Problem: QFN-32 device failed intermittently after 500 thermal cycles.

1. Why was it intermittent? → Cold joint on pin 15 — incomplete IMC
2. Why was the IMC incomplete? → Peak temperature at the component body was only 225 °C versus the 235 °C required
3. Why was the temperature low? → The large ground plane under the QFN acted as a heat sink
4. Why was the thermal mass effect not accounted for? → The profile was built on a standard test board, not the production board
5. Why was it not validated on the production board? → The NPI process did not require production-board profiling

TRC: Low joint temperature caused by the PCB thermal mass effect. MRC: Reflow profile validation used non-representative test boards.

Fishbone (6M)

• Man: Operator increased conveyor speed for throughput without re-profiling
• Machine: Zone 3 thermocouple offset by 5 °C
• Method: Profile checks only every six months
• Material: OSP coating on the PCB was older than its six-month shelf life
• Measurement: Profiled on a thin test board, not the actual product
• Environment: Winter shift temperature drop affected the preheat zone

D3 Containment

• 100% X-ray inspection of the affected lot; quarantine all QFN-32 assemblies
• Raise the reflow peak setpoint by 10 °C as a temporary measure
• Verify the profile on the actual PCB at the start of every shift

D5 Permanent Fix

• Build a production-representative profiling board with thermocouples at every critical thermal mass
• Install real-time reflow SPC with a ±3 °C alarm
• Upgrade AOI with QFN / BGA joint inspection algorithms
• Require production-PCB profiling in the NPI checklist before mass production

D7 Prevention

• NPI process: the thermal profile must be validated on the production PCB
• Quarterly reflow oven calibration and profile verification
• Keep OSP PCB inventory under three months
• DFMEA: add thermal relief design for QFN ground pads

Key Takeaways

1. A cold joint can pass test and still fail in the field — verify the process, not just the sample.
2. Thermal mass, not the oven setpoint, is the usual culprit on large copper planes.
3. Always profile the production board, never a standard test coupon.
4. X-ray is the only reliable way to inspect hidden joints under BGA and QFN.
5. The management root cause is a validation gap in the NPI process.