Skip to main content

5.5 Data Logging & Repair Tickets

Capturing failures, routing to rework/MRB, and closing CAPA (Corrective and Preventive Action) loops

A good data

Data logging and repair tickettickets are the nervous system makesof electronics manufacturing, turning every failure traceableinto structured information that can be acted upon. Instead of vanishing into sticky notes or informal fixes, each defect is captured with full traceability—serial numbers, defect codes, images, and fixable,process capturingcontext—so no issue gets lost. Repair flows, MRB decisions, and CAPA actions all keydraw detailsfrom automaticallythe sosame nothingrecords, isensuring lost.consistent Standardcontainment, defectsafe codesrework, routeand problemsdata-driven prevention. With this discipline, the factory not only repairs boards efficiently but also learns from each failure to thereduce right owner,recurrence and boards move from quarantine through rework or MRB (Material Review Board, which decides whether to repair, scrap, or return) with re-inspection before release. Strong rework discipline and consistent MRB decisions keep fixes safe and repeatable. Key metrics such as MTTR (Mean Time to Repair), repeat-fail rate, and CoPQ (Cost of Poor Quality) guide upstream process improvements and ensure consistent communication across manufacturing and test. The result is a closed-loop system that contains, corrects, and prevents defects efficiently.

cost.

5.5.1 What “good” looks like (one page)

A fail becomes a ticket in seconds—auto-filled with facts (who/what/when/where/how). The board is quarantined by ID, the right people get pinged, repair techs follow a standard work recipe, retest proves the fix, and the data flows into Pareto and CAPA without copy-paste. No sticky notes. No mystery boards.



5.5.2 The record you must capture (every time)

Make tickets boringly complete. Your MES/traceability system should auto-populate most fields:

Identity & context

  • Board Serial Number / Panel ID / Work Order
  • Product/Rev, BOM Rev, ECN (if in effect)
  • Station (SPI/AOI/AXI/ICT/FCT), recipe rev, line/machine ID
  • Date/time, operator/tech badge

Failure detail

  • Defect code (from a controlled list)
  • Refdes / Location (X/Y if available)
  • Side (Top/Bottom), layer (if via/PTH)
  • Evidence: SPI/AOI/AXI images, reflow profile plot, ICT/FCT measurements

Materials hooks

  • Stencil ID, paste lot, component lot/date code, feeder ID, fixture ID
  • Optional: reel splice timestamp, program rev (PnP)

Routing

  • Containment state (Quarantine → Rework → MRB)
  • Disposition (Rework / Scrap / Use-as-is / RTV to supplier)
  • Repair action (WIs used, parts replaced), tech, ret test result

If your software can’t store attachments, fix the software. Pictures and plots end debates.



5.5.3 Defect codes that drive action (keep it small)

Start with 15–25 codes, not 200. Codes should point to a process knob.

Family

Code

What it means

First owner

Print

PVL-LOW / PVL-HIGH

SPI volume out of band

Printing team

Reflow

BGA-HIP

Head-in-pillow

Process/oven


QFN-VOID

Thermal voids over limit

Process/oven

Placement

ORIENT

Polarity/pin-1 error

PnP/DFx

Assembly

BRIDGE-FP

Fine-pitch bridge

Stencil/print

Materials

COMP-BADLOT

Component lot suspect

SQE/Supplier

Test

ICT-OPEN / ICT-SHORT

Structural fault

DFT/ICT

Functional

FCT-COMMS / POWER

Interface/power fail

Test Eng

Damage

MECH-CRACK / ESD

Handling/ESD

Ops/QE

Never put “Other” without a free-text note. Free-text alone is how problems hide.



5.5.4 Flow that never loses a board (golden path)

  1. Fail at station → Auto-ticket. MES locks the panel/SN; NG diverter sends it to quarantine.
  2. Triage (cell lead/QE): confirm code, attach evidence, choose Rework vs MRB.
  3. Rework (if reversible): tech opens standard work (WI) by code, performs fix, logs parts used and cycles.
  4. Verify: required re-inspection/re-test in the same station that failed plus any downstream gates affected. No bypass.
  5. Close: ticket gets root cause (5-Why field), action taken, and sign-off.
  6. MRB (if not reworkable/suspicious): board, parts, and paperwork to Material Review Board for disposition (Scrap / Use-as-is / RTV).
  7. CAPA (for recurring/critical): open 8D; assign owner, due dates, and effectiveness check.

House rule: a board cannot leave repair without passing the failing step again. “Looks good” isn’t a result.



5.5.5 Rework discipline (fast, safe, repeatable)

  • ESD controls on every bench; log strap tests.
  • Work Instructions (WIs) per code: tools, temps, nozzle sizes, paste/flux, max heat cycles.
  • Rework limits per part: e.g., BGA ≤ 2 reflows, 0402 ≤ 3 touch-ups—ticket must count attempts.
  • Verification: AXI after BGA/QFN work; AOI for fine-pitch; ICT/FCT for electricals; update genealogy if parts replaced (lot/date link).



5.5.6 MRB & dispositions (decide once, apply always)

MRB = QE + PE/ME + Materials/SQE + sometimes Design. Inputs: ticket, evidence, cost, risk.

  • Rework (safe/standardized)
  • Use-as-is (documented risk; customer waiver if needed)
  • Scrap (record CoPQ hit)
  • RTV (Return to Vendor) with photos, fail data, and lot trace → start SCAR (supplier CAPA)

MRB outcomes should update AVL status or process settings when patterns appear.



5.5.7 CAPA that actually closes

  • Trigger: threshold on a code (e.g., >0.5% of boards for 2 days, or any safety/field risk).
  • 8D skeleton in the system: D1 team, D2 problem statement, D3 containment, D4 root cause (5-Why/Fishbone), D5 corrective, D6 preventive, D7 effectiveness, D8 closure.
  • Effectiveness: define the metric up front (e.g., BRIDGE-FP ≤ 0.1% for 4 lots) and check after release. No metric = no closure.



5.5.8 Metrics to run the floor (and reviews)

  • MTTR (mean time to repair) by code/station
  • Repeat-Fail Rate (same SN, same code)
  • Top-10 Pareto (last 7/30 days)
  • Aging tickets (open >24/48 h)
  • Disposition mix (Rework/Scrap/RTV/Use-as-is)
  • DPPM by source (Process / Materials / Design / Test)
  • Cost of Poor Quality (CoPQ) = Scrap + Rework labor + Parts + RTV costs
  • Supplier score: defects per lot, SCAR on-time closure

Weekly ops/QE look at Pareto + MTTR; monthly MRB/CAPA review checks effectiveness and CoPQ trend.



5.5.9 Glue to upstream & downstream (close the loop)

  • Upstream: recurring BRIDGE-FP → stencil/aperture change (7.4) and printer setup (7.5); BGA-HIP → profile/TAL + N₂ (8.x); COMP-BADLOT → SQE + incoming screens (6.5).
  • Downstream: FCT fails feed back to ICT/BSCAN adds (11.1–11.3) or firmware self-tests; RMA field returns enter the same code system so factory and field speak one language.



5.5.10 Pocket checklists

At the station (operator)

  • Ticket popped with SN; evidence attached (image/plot/data)
  • Board quarantined (NG lane / rack); no pass-around
  • Correct defect code selected; free-text note if edge case

At repair (tech)

  • ESD check OK; WI opened for this code
  • Parts/flux/temps logged; rework attempt count updated
  • Required re-inspection/re-test passed; genealogy updated if parts swapped

At MRB/CAPA (QE/PE/SQE)

  • Disposition chosen with data; CoPQ recorded
  • Supplier notified (if RTV) with full pack of evidence
  • CAPA opened if threshold hit; metric for effectiveness defined
  • Dashboard reflects change; “why” written in release notes



Bottom

A line:robust make tickets automatic, evidence mandatory,ticketing and routingrepair predictable.system Standardizeensures rework,defects useare MRBcontained, tofixes makeare one clear decision,verified, and driveroot CAPAcauses are closed with metricsmeasurable youeffectiveness. checkThis later—notstrengthens promises.both Whenimmediate datayield flowsand cleanlylong-term fromreliability, failwhile tokeeping fixteams toaligned dashboard,around qualityclear, improvesfactual and arguments disappear.

data.


Back to top