Quick Root Cause Record

_NCMR-compatible nonconformance record_

NC Description

Cold solder / insufficient wetting on QFN-48 ground-pad thermal via array joints on BCN-IoT-4800 Rev C, produced on Reflow Oven #1913, against the internal target of <0.5% defect rate (IPC-A-610 / J-STD-001, Class 2). Detected by routine AOI (Dan Kim), confirmed by X-ray. First detected June 2, 2026.

NC owner: Anna Novak, SMT Process Engineer (CAPA owner, NCR#26-0244).

Scope / Quantity at Risk

• Product: BCN-IoT-4800 Controller PCBA, Rev C
• Equipment: Reflow Oven #1913
• Population inspected: 2,000 QFN-48 joints, June 2–10, 2026, on affected PN, Oven #1913 output
• Defects found: 84 of 2,000 (4.2%)
• Comparison population: Ovens #1911/#1912, same paste lot and window — 6 of 2,000 (0.3%)
• Containment population: all Oven #1913 output, affected PN, June 2 onward — 100% X-rayed before release
• NCR#26-0244, opened June 5, 2026

Disposition

• 84 confirmed defective joints: Rework. Hand-reworked by Raj Patel; re-inspected (X-ray + AOI) by Dan Kim; confirmed conforming post-rework.
• Remainder of Oven #1913 output, affected PN, June 2–10 (X-ray passed): Use-as-is. Justification: every unit individually screened by 100% X-ray, not statistically sampled; confirmed free of the defect morphology under IPC-A-610 Class 2. Approving authority for this use-as-is disposition: TBD — no formal MRB/engineering sign-off role named beyond Anna Novak (CAPA owner); record this as an open item.
• New-build WIP on Oven #1913, affected PN: On hold, pending Zone 7 repair, board-level profile validation at the QFN pad, and re-verification (see Immediate Correction).
• Paste lot PST-26-0339: not rejected. Within J-STD-006 spec (Ag 2.5–4.0%; this lot 2.8%); ran clean on Ovens #1911/#1912. Confirmed contributing factor, not standalone cause — see Root Cause(s).

Containment

Verification: 100% X-ray screening confirms no unscreened product has shipped or can ship from the suspect population. No customer notification or regulatory report required — Class 2, no safety-critical application, zero field exposure, no customer complaints as of June 10, 2026.

Decisive Contrasts (IS / IS-NOT)

Decisive contrasts:

1. Same paste lot ran on all three ovens, but defects appeared only on #1913 — this rules out paste lot alone as the cause and points to something oven-specific.

2. Zone 7 is the peak/reflow zone, the exact point in the profile where the board must cross liquidus — a shortfall there hits hardest at the highest-thermal-mass joint.

3. Defects are isolated to the QFN ground pad only, not distributed across joint types — consistent with a marginal, not gross, thermal shortfall that only bites the joint with the least margin.

6M Fishbone (breadth)

Machine

• CANDIDATE (RC-1): Oven #1913 Zone 7 heater running ~5°C below programmed setpoint in the peak/reflow zone, week of June 1, 2026 — confirmed contributor.
• RULED OUT: Zones on #1911/#1912 all within ±2°C of setpoint — no comparable equipment deviation on the other two ovens.

Material

• CANDIDATE (RC-2): Paste lot PST-26-0339, SAC305 liquidus 219°C vs. 217°C standard (+2°C thermal demand) — confirmed contributor.
• RULED OUT (as standalone cause): Same lot ran on #1911/#1912 at 0.3% defect rate (normal) — falsifies paste-alone hypothesis.

Method

• CANDIDATE (systemic escape point): No interim escalation threshold for single-zone deviations between 90-day profile verifications — WI-SMT-014 (current rev) contains no such threshold.
• CANDIDATE (systemic escape point): Profile verification uses gold-board fixture; thermocouple not placed at QFN ground pad — per KIC-1913-20260428 and the thermocouple placement map in WI-SMT-014.
• CANDIDATE (systemic escape point): No requirement to re-verify thermal profile when a paste lot change shifts liquidus demand — confirmed by Anna Novak, not addressed in WI-SMT-014.

Measurement

• TBD: No board-level peak-temperature measurement at the QFN ground pad exists from the defect window — direct confirmation of the mechanism is not available; action-2 profile validation will close this.
• CANDIDATE: Zone 7 thermocouple history log exists and shows the ~5°C deviation — indirect but consistent measurement evidence.

Man (People)

• RULED OUT (as individual cause): Raj Patel observed the Zone 7 deviation the week of June 1 but had no documented threshold to act against — the oven had passed its April 28 verification and no interim trigger existed. System gap, not individual error.

Environment (Mother Nature)

• Considered — no environmental differences reported or identified between ovens, zones, or shifts. Only one disposition here because no mechanism connecting ambient/environmental factors to an oven-specific, zone-specific defect was proposed or observed.

5-Why Chain

Why 1: Why did the QFN-48 ground-pad thermal via array joints exhibit insufficient wetting on Oven #1913?

> The joint did not reach or sustain sufficient temperature above liquidus during reflow.

> Evidence: Zone 7 thermocouple history log; COA liquidus 219°C; defect isolated to highest thermal mass joint.

Why 2: Why did the joint not reach sufficient temperature above liquidus?

> Zone 7 (peak zone) ran ~5°C below setpoint (RC-1) at the same time paste lot PST-26-0339 required 2°C more thermal energy to wet (RC-2, liquidus 219°C vs. 217°C standard). The QFN ground pad — slowest joint to heat — was the only location where the combined margin loss crossed the wetting threshold.

> Evidence: Zone 7 thermocouple history log; COA PST-26-0339; COA standard lot; defect isolated to QFN ground pad only, all other joints wetting normally.

Why 3: Why was the Zone 7 deviation not caught and corrected before it produced defects?

> No documented escalation threshold exists for single-zone deviations between scheduled 90-day profile verifications. Raj Patel observed the deviation but had no procedural trigger to act — the oven had passed its April 28 verification.

> Evidence: WI-SMT-014 (current rev) — confirmed by Anna Novak to contain no such threshold.

Why 4: Why did the 90-day profile verification not detect this as a risk?

> Verification is performed on a gold-board thermocouple fixture that does not place a thermocouple at the QFN ground pad — the highest thermal mass, worst-case location. A single drifting zone can remain within the verification envelope while still producing insufficient peak temperature at the joint that needs it most.

> Evidence: KIC-1913-20260428 (April 28 verification, passed); thermocouple placement map in WI-SMT-014.

Why 5: Why does the verification method not account for the worst-case joint or for paste changes that shift thermal demand?

> WI-SMT-014 was designed to confirm the oven's general thermal output against a fixed fixture — not to measure worst-case production joints, trigger re-verification on incoming material changes, or provide interim monitoring between 90-day cycles.

> Evidence: WI-SMT-014 (current rev) — confirmed by Anna Novak.

Root Cause(s)

RC-1 — Oven #1913 Zone 7 thermal deviation (co-equal contributing cause).

Mechanism: Zone 7 (peak/reflow zone) heater output drifted to ~5°C below programmed setpoint → reduced peak temperature delivered to the board at the moment of liquidus crossing → insufficient thermal energy reaching the highest-mass joint.

Verdict: PROBABLE. Strong, consistent evidence (thermocouple log, oven-to-oven contrast, timing) and the paste-alone alternative is ruled out; direct board-level confirmation at the QFN pad (Action 2, due June 12, 2026) is the step that would move this to CONFIRMED.

RC-2 — Paste lot PST-26-0339 higher liquidus (co-equal contributing cause).

Mechanism: SAC305 liquidus at 219°C (vs. 217°C standard) → +2°C higher thermal energy required for full wetting → reduced margin at the joint with the least thermal margin to begin with (QFN ground pad).

Verdict: PROBABLE. Confirmed via COA; consistent with defect isolation to the highest-thermal-mass joint; ruled out as a standalone cause because the same lot ran clean on #1911/#1912. Moves to CONFIRMED with the same Action 2 board-level profile data.

Systemic root cause — verification method escape point.

Mechanism: WI-SMT-014's 90-day gold-board verification neither instruments the worst-case joint nor triggers on paste-liquidus changes nor sets an interim zone-deviation threshold → a drifting Zone 7 heater combined with a marginally higher-liquidus paste lot went undetected until defects appeared in AOI.

Verdict: CONFIRMED — directly evidenced by WI-SMT-014 (current rev, no threshold and no QFN-pad thermocouple) and KIC-1913-20260428 (verification passed without measuring the failing location).

Ruled-Out Hypotheses

• Paste lot PST-26-0339 alone. Ruled out — the same lot ran on Ovens #1911/#1912 in the same window at 0.3% (normal). If paste were the sole cause, all three ovens would show elevated defects.
• Oven #1913 Zone 7 deviation alone (independent of the paste change). Not currently supported as sufficient by itself — defects appeared June 2, the day after the paste lot changed, not before. The prior (217°C) paste lot ran on #1913 with the same zone condition present and no elevated defect rate reported.
• Stencil / print variation. Ruled out — same stencil across all three ovens; no print-related defect pattern reported.
• Placement variation. Ruled out — same placement program and feeders across all three ovens; no placement offset reported.
• Generic voiding / head-in-pillow. Ruled out — X-ray and AOI confirmed cold-joint / insufficient-wetting morphology, not voiding or HiP failure signatures.
• Environmental factors. Not currently supported — no environmental differences identified between ovens or shifts.
• Operator error (Raj Patel not escalating the Zone 7 reading). Ruled out as an individual cause — no documented threshold existed for him to act against; this is a system gap (see Method, Root Cause(s)).

Evidence

Immediate Correction

Escalation path: if the Zone 7 heater cannot be restored to within ±2°C of setpoint by June 11, 2026, Oven #1913 remains out of service for BCN-IoT-4800 Rev C pending equipment repair or replacement.

Approval / Sign-off

This report has been generated and captured here; it is not locked, closed, or signed off. Sign-off occurs through the Review & Sign-off step in the tool, where each named person above records their approval.

Escape Point

Profile verification (WI-SMT-014, 90-day cadence) does not instrument the QFN-48 ground pad — the board's worst-case thermal location — and carries no interim escalation threshold for single-zone deviations or requirement to re-verify on paste-liquidus changes. This is why a drifting Zone 7 heater and a marginally higher-liquidus paste lot combined and produced defects for several days before detection by AOI, rather than being caught at the equipment or incoming-material stage. Formal escape-point remediation (procedure revision, threshold-setting, fixture upgrade, effectiveness verification) is out of scope for this record — see Open Items & Gap Impact and route to CAPA/Full 8D if escalated.

Open Items & Gap Impact