The Repair LibraryRead · Learn · Master

BGA Repairability Assessment

The catalog told you what killed the package; this section decides what — if anything — the bench should do about it. The instrument is the repair envelope: each intervention reaches some failure classes and not others, and the map is learnable. Reflow genuinely repairs only what never formed correctly — head-in-pillow, cold, and unwetted joints can be completed by a disciplined second reflow with fresh flux, because finishing an unfinished joint is a real repair. Reballing renews the ball metallurgy entirely — the genuine answer to fatigue-cracked fields when the pads are intact and the silicon is alive. Replacement renews the package — the only reach into die-side death. And below all of them sits the board's own bench: cratered pads, torn laminate, and broken fanout vias are board-side damage that no ball-level intervention touches, handed off to the pad-repair discipline Volume 4 built or priced as a board replacement. Against the envelope stands the internet's favorite counterfeit: the blind reflow — the oven trick, the heat-gun resurrection — which briefly re-fuses crack faces or accidentally completes a pillowed joint, then relapses on schedule because the fatigued metallurgy, the standoff, and the root cause all remain. The section's second half turns the envelope into arithmetic: success probability against labor, the collateral risk a heavy rework imposes on neighbors and board, replacement and donor availability, and the data-versus-device question that reprices everything. It closes with the decision discipline — class, envelope, price, then commit, decline, or hand off — and the honest quote that names what the bench can and cannot promise. Sometimes the most professional repair is the one turned down well.

AdvancedLow Risk23 min read

What You Will Learn

  • You will learn the repair envelope — which failure classes reflow, reballing, and replacement each genuinely reach.
  • You will learn the unreachable tier — cratered pads, torn laminate, broken fanout vias, and die-side death that no ball-level repair touches.
  • You will learn why blind-reflow resurrections are temporary — and what separates the oven trick from a repair.
  • You will learn the pricing arithmetic — success probability, labor, collateral risk, replacement availability, and the data-versus-device question.
  • You will learn the decision discipline — class, envelope, price, then commit, decline, or hand off, with the honest quote.

What You Will Be Able To Do

  • You will be able to map each failure class onto the interventions that genuinely reach it.
  • You will be able to identify board-side and die-side damage outside every ball-level envelope and name the handoff each requires.
  • You will be able to explain why a blind reflow resurrects briefly and relapses on schedule.
  • You will be able to price a BGA job honestly — probability, labor, collateral risk, and the replacement alternative.
  • You will be able to run the decision discipline end to end and write the quote that names what the bench can and cannot promise.

Required Tools

  • The four-class table and stress maps from last section — the assessment's inputs are the catalog's outputs
  • A notebook — this section builds the envelope map and the quote templates the whole volume will price with
  • Donor boards and the case files of real dead devices where the bench has them — decisions practice best on real stakes
  • Parts-availability sources — the replacement question is half the arithmetic, and it is looked up, not guessed
  • No hot tools — this section decides; the chapters ahead act

When NOT to Attempt This

Do not attempt this section if any of the following apply to you:

  • You are not comfortable working with small surface-mount components.
  • You have not completed the prerequisite sections for this skill.
  • You do not have the required tools in working condition.

Section Overview

The catalog named the deaths; this section decides the bench's answer, and the instrument is the repair envelope (bga-failure-modes). Each intervention reaches what it renews. Reflow completes joints that never formed — pillowed, cold, unwetted; reballing renews fatigued ball metallurgy over intact pads; replacement renews the package and is the only reach into die-side death (what-is-bga). Below the envelope sits the unreachable tier. Cratered pads, torn laminate, broken fanout vias — board-side damage handed to the board's own bench or priced as replacement (repairability-assessment). Against the envelope stands the counterfeit. The blind reflow — the oven trick — resurrects briefly and relapses on schedule, because the root cause survives the miracle. The arithmetic prices the decision. Success probability against labor, collateral risk against the neighbors, replacement availability against the hours — repriced entirely when the payload is data (when-to-remove-a-chip-vs-replace-the-board). And the discipline closes it. Class, envelope, price — commit, decline, or hand off, with the quote that names what the bench can and cannot promise. The envelope, the unreachable, the counterfeit, the price, the discipline — the assessment entire.

Why This Matters

BGA work is the most expensive bench time this handbook teaches, and the assessment is what keeps that time from being spent on jobs physics has already decided (what-is-bga). This matters because the envelope prevents category errors: a fatigue-cracked field reflowed instead of reballed relapses; a cratered pad reballed instead of repaired fails at first power; the class-to-intervention map is the difference between repair and ritual (bga-failure-modes). This matters because the counterfeit is everywhere: the oven trick and the heat-gun resurrection circulate endlessly, customers arrive asking for them by name, and the bench that can explain the relapse — briefly risen, root cause intact — earns more trust than the one that just says no (when-to-remove-a-chip-vs-replace-the-board). It matters because the arithmetic protects both parties: an honest success probability and a named collateral risk produce quotes customers can actually decide on — and protect the bench from owning a board it could never have saved (repairability-assessment). It matters because data reprices everything: a dead phone worth fifty dollars carries photographs worth anything, and the assessment that notices the payload changes the whole decision. And it matters because declining is a professional skill: the volume ahead teaches removal, reballing, and replacement — and this section teaches when not to use them, which is the judgment that makes the skills safe to own. Decide before heating — every chapter ahead assumes this section happened first.

Required Prerequisites

  • BGA Failure Modes — the four-class catalog whose output is this section's input: no class call, no assessment.
  • What Is BGA? — the package anatomy the envelope reasons over: balls, substrate, die, and the board underneath.
  • Nothing consumed — the section spends judgment, not materials (what-is-bga)
  • A notebook page for the envelope map — to build the class-to-intervention grid the volume will consult before every job (bga-failure-modes)
  • A page for the quote templates — to draft the commit and decline languages once, calmly, before a customer is waiting (when-to-remove-a-chip-vs-replace-the-board)
  • Case-file cards for the bench's real dead devices — to practice decisions on stakes that already happened (repairability-assessment)
  • The donor bin with its flagged BGAs and the catalog's stress maps — to ground every envelope call in a package that exists (bga-failure-modes)
  • A parts-availability search on a live device — to practice the replacement question the arithmetic requires (when-to-remove-a-chip-vs-replace-the-board)
  • One genuinely dead device with a known story — to run the full discipline end to end on real stakes (repairability-assessment)
  • The notebook and quote templates — to write decisions down, because an assessment that lives in the head drifts (what-is-bga)
  • The bench camera — to file the envelope maps and practice quotes in the record

Real-World Applications

The assessment pays off at the moment of commitment, and first in the jobs taken well. A bench quoting a game console with the classic warm-up video fault runs the discipline in order: cycling-crack class from the interview, reball as the envelope's genuine answer, pads assumed intact until inspection says otherwise — and quotes the reball with its real success probability instead of the oven trick the owner found online (bga-failure-modes). A shop handed a dashcam that went through a windshield and into the rain prices the stack honestly — probable cratered pads under the impact, corrosion already creeping under the water, two envelopes exceeded before the quote is written — and declines with a written explanation that earns the next job (when-to-remove-a-chip-vs-replace-the-board). A technician holding a dead phone full of un-backed-up photographs reprices the whole decision around the payload: the device is scrap arithmetic, but the data makes a controlled, one-shot repair worth a specialist's rate — and worth saying so plainly (repairability-assessment). And a hobbyist tempted to practice a first reball on a customer's board reads the collateral line of the arithmetic and routes the learning to the donor bin instead — the practice patients the study sections flagged exist exactly for this (what-is-bga). The confusions this prevents: a reball sold where a reflow completion would do, an oven trick sold as a repair, a cratered board heated toward a failure it was always going to reach, and a data-bearing device treated as device-priced scrap.

Common Challenges

  • The envelope is invisible at quote time. The class call comes from interview and catalog, not from opening the patient — the difficulty is quoting on a hypothesis honestly: name the class you believe, the envelope it implies, and the inspection that could change both (bga-failure-modes).
  • The counterfeit has better marketing. The oven trick is free, famous, and occasionally works for a month — the difficulty is explaining the relapse without condescension, because the customer who understands the trap becomes the customer who pays for the repair (when-to-remove-a-chip-vs-replace-the-board).
  • Declining feels like failure. The bench learned the skills and wants to use them — the difficulty is internalizing that the decline is a use of the skills: the judgment chapters ahead assume, applied at the only moment it is cheap (repairability-assessment).

Safety Notes

Risk Level: Low. Nothing is heated and nothing is opened — this section decides, and the standing law holds around the deciding.

Professional Tips Before Starting

  • Write the envelope map before you need it. Class down the side, interventions across the top, reach marked honestly — the grid drawn calmly today beats the same grid improvised over a customer's board (bga-failure-modes).
  • Draft both quote templates now. One for commit — class, plan, probability, collateral named; one for decline — reason, alternative, respect — the language written in advance is the language that stays honest under pressure (when-to-remove-a-chip-vs-replace-the-board).
  • Ask about the payload first. Data changes the price, the method, and sometimes the practitioner — the question costs nothing and reprices everything (repairability-assessment).

The Assessment — The Envelope, The Unreachable, The Counterfeit, The Price, The Discipline

Recap and Frame

The chapter has built package, metallurgy, and catalog; this section is where the theory becomes a decision (what-is-bga). The catalog's output is the input. Class, mechanism, and stress geography from the last section arrive here as the facts the assessment reasons over — the interview and the four-class card are already half the quote (bga-failure-modes). The decision has ancestors. Volume 3 taught chip-versus-board arithmetic and Volume 4 taught board-level repairability — the same honest weighing, now aimed at the most expensive package class the bench takes on (when-to-remove-a-chip-vs-replace-the-board). The stakes are the volume's highest. BGA interventions cost hours, risk neighbors, and get one or two attempts at most — the price of deciding badly is the board, which is why the deciding gets its own section before any chapter teaches the heat (repairability-assessment). What is new is the envelope. Not whether the bench can repair — whether this intervention reaches this failure: a map of genuine reach, drawn class by class. And the order is deliberate: envelope first, the unreachable tier, the counterfeit that pretends the envelope away, then the arithmetic and the discipline that turn all of it into a quote. Hold the frame — theory become judgment, judgment before heat — and the section does its work.

The Envelope — What Each Intervention Genuinely Reaches

Every intervention repairs by renewing something, and its reach ends where its renewal does (bga-failure-modes). Reflow renews formation. A disciplined second reflow with fresh flux completes joints that never formed correctly — the pillowed ball merged at last, the cold or disturbed joint carried to true liquidus, the unwetted land finally wetted — when the non-wet was flux or oxide rather than the pad's own finish, for a black-pad or contaminated land is board-side, outside reflow's reach — a genuine repair, because the failure was an unfinished process and finishing is the fix. Reflow does not renew metallurgy. A fatigue-cracked field re-melted is old solder refrozen: grain structure aged, voids in place, standoff unchanged — the crack's causes all present at the next cold morning (what-is-bga). Reballing renews the ball metallurgy. Old alloy stripped, fresh spheres attached, the field's mechanical life restarted — the genuine answer to cycling-cracked fields, provided the pads beneath are intact and the silicon above is alive, and the answer the reballing chapter will teach hands-on. Replacement renews the package. A donor or new part swaps silicon, substrate, and balls together — the only reach into die-side death, package-internal damage, and the failures no ball work touches — at the price of the full removal-and-replacement craft ahead (when-to-remove-a-chip-vs-replace-the-board). Shorts split across the envelope. A process bridge is a removal-and-redo story; a migration short is cleaning and condition removal, the one repair in the catalog that may need no package work at all — though the standoff limits what cleaning reaches in place, and the honest path is sometimes removal to clean (repairability-assessment). Formation, metallurgy, package — reflow, reball, replace — the envelope entire. Match the renewal to the failure — an intervention aimed outside its envelope is effort spent rehearsing the next failure.

The Unreachable Tier — Board-Side and Die-Side

Two territories sit outside every ball-level envelope, one below the package and one inside it (what-is-bga). Board-side damage is below. Cratered pads from drop, laminate torn under a pad, a fanout via cracked in the escape routing — the failure catalog's board-side entries live in the board, and no reflow, reball, or replacement touches them: new balls on a torn pad are a joint to nothing. The handoff exists. Volume 4 built the pad-repair bench — lifted pads rebuilt, fine-pitch lands remade, repairs verified — and board-side damage under a BGA is that discipline at its hardest, priced honestly against board replacement because under-package pad repair multiplies every difficulty the pad bench knows (repairability-assessment). Die-side death is inside. A cracked die, broken package-internal bumps from a careless profile, silicon killed by the drop the balls survived — package-internal failures leave replacement as the only envelope, and when no replacement part exists, the honest answer is that no repair exists (bga-failure-modes). The diagnosis is subtractive. Die-side death is suspected when the joints check out — continuity good, X-ray clean in the chapters ahead — and the package still will not live: the conviction arrives by eliminating everything reachable (when-to-remove-a-chip-vs-replace-the-board). Below the balls or inside the body — the unreachable tier entire. Knowing where the envelopes end is what keeps the bench from selling repairs that physics already declined.

The Counterfeit — Blind Reflow and the Resurrection Trap

The internet's favorite BGA repair is heat without diagnosis, and the section names it honestly: the blind reflow (bga-failure-modes). The trick has real mechanisms. An oven's soak or a heat gun's pass can re-fuse the touching faces of a fatigue crack, or accidentally complete a pillowed joint that sat a breath from merging — the device genuinely rises, which is exactly why the trick will not die. The relapse is scheduled at the miracle. The re-fused crack is old, voided, grain-coarsened solder across an unchanged standoff under an unchanged die shadow — every cause of the first crack intact — and the accidental completion cured one joint of a field whose warpage and profile problems went unaddressed: weeks or months, and the symptom returns (what-is-bga). The costs are hidden. An uncontrolled bake stresses every plastic, connector, and neighbor on the board — collateral spent on a non-repair — and the second death often arrives worse: more cracks, moved neighbors, a board now harder to save than before the first miracle (repairability-assessment). The counterfeit prices the genuine article. What separates a blind reflow from the reflow the envelope honors is the class diagnosis and the discipline around the heat — known failure, chosen profile, verified outcome — the difference between finishing an unformed joint and briefly bribing a cracked one (when-to-remove-a-chip-vs-replace-the-board). And the customer conversation is the skill. Explain the mechanism, the relapse, and the real options without scorn — the owner who understands the trap becomes the customer who chooses the repair. Real mechanisms, scheduled relapse, hidden costs — the counterfeit entire. Heat without a class call is a delay purchased with trust — the envelope's reflow and the oven's reflow share a temperature and nothing else.

The Arithmetic and the Discipline — Pricing the Decision

The envelope says what is possible; the arithmetic decides what is wise, and it has named terms (when-to-remove-a-chip-vs-replace-the-board). Success probability is a fraction, stated. A reflow completion on a fresh process failure runs high; a reball over suspected-but-uninspected pads runs lower; anything after a previous blind reflow runs lower still — the number is an estimate, and an estimated number beats an implied certainty. Labor is the volume's most expensive. Removal, reballing, and replacement are hours of skilled bench time with setup, consumables, and verification — the arithmetic charges them against the outcome honestly (repairability-assessment). Collateral risk is the term benches forget. Heavy rework heats neighbors, stresses the board, and risks the exact under-package pads whose damage would end every envelope — one attempt's collateral can convert a repairable board into an unreachable one, which is why the risk is named in the quote, not discovered after (what-is-bga). The replacement alternative anchors everything. A donor board at half the repair's price ends the repair conversation; an unobtainable board removes the anchor and raises what a repair is worth — the availability search is part of the assessment, not an afterthought. Data reprices the whole equation. When the payload is photographs or work, the device's scrap value stops mattering: the job becomes a controlled, conservative, one-shot recovery priced on the data — and sometimes routed to a specialist, which is itself an assessment outcome (bga-failure-modes). And the discipline orders it all. Class from the catalog, envelope from the class, price from the arithmetic — then commit with a quote that names probability and collateral, decline with a reason and an alternative, or hand off to the board bench or the specialist. Probability, labor, collateral, anchor, payload — the arithmetic entire. The declined job costs a quote; the misjudged one costs the board, the hours, and the trust — decline well, and the decline pays.

Common Mistakes

  • Selling the intervention the bench wants to perform. A reball quoted because reballing is the new skill — the envelope matches renewals to failures; a reflow completion or a cleaning may be the honest job (bga-failure-modes).
  • Reballing over unexamined pads. Fresh spheres committed to a drop-death board before anyone looked underneath — cratered pads end the ball-level envelope; the board-side question comes first (repairability-assessment).
  • Performing the oven trick with better equipment. A profiled, controlled, documented reflow of a fatigue-cracked field — discipline does not move the envelope: old metallurgy refrozen is still old metallurgy, however good the profile (what-is-bga).
  • Quoting certainty. 'We can fix that' where the honest number is sixty percent — the stated fraction and the named collateral are what make the quote a decision the customer actually gets to make (when-to-remove-a-chip-vs-replace-the-board).
  • Ignoring the payload question. A data-bearing phone priced as a fifty-dollar device — ask what is on it first; the answer reprices the method, the caution, and sometimes the practitioner.

Troubleshooting Guidance

The assessment troubleshoots decisions, and its failures have signatures too: misclass, mismatch, misprice. If a committed repair relapsed: re-run the discipline backward — was the class call wrong, or the envelope mismatched to a right call? A reflow that bought a month says fatigue was treated as formation; the relapse is data, and it points at the reball or replacement conversation (bga-failure-modes). If a job died on the bench: audit the collateral line — did the attempt's heat or handling exceed what the quote priced? The postmortem that names the overrun is what keeps it out of the next quote (repairability-assessment). If quotes keep losing to the oven trick: the explanation is underperforming, not the price — walk the relapse mechanism once with the customer, briefly risen and rescheduled, and let the trick price itself (when-to-remove-a-chip-vs-replace-the-board). If the decision will not settle: the missing fact is usually one of three — the class call, the pad condition, or the replacement price — name which, get it, and the decision usually makes itself (what-is-bga). If every assessment lately says decline: check the inputs before the courage — a bench that cannot yet inspect under the package is assessing blind, and the X-ray and inspection chapters ahead widen what can be known and therefore what can be honestly committed. The throughline: decisions are repairable too — audit them with the same class-mechanism honesty the boards get.

Verification & Testing Methods

Confirm the assessment before the survey closes the chapter:

  • [ ] I can map every failure class onto the interventions whose renewal genuinely reaches it — the repair envelope, drawn from memory.
  • [ ] I can name the unreachable tier — cratered pads, torn laminate, broken fanout vias, die-side death — and the handoff or verdict each requires.
  • [ ] I can explain the blind reflow's brief resurrection and scheduled relapse, and what separates it from the envelope's genuine reflow.
  • [ ] I can price a job with stated success probability, labor, collateral risk, and the replacement anchor — repriced when the payload is data.
  • [ ] I can run class → envelope → price → commit, decline, or hand off, and write both quote languages.

Then try the practice exercises below — desk and judgment work only; scenarios differ from the quiz.

Practice Exercises

  1. Draw the envelope map (5 minutes, desk). Grid the catalog's classes down the side — mechanical both faces, process, wetting, short both members, plus a die-side and package-internal row — against reflow, reball, replace, board-bench handoff, and decline across the top; mark genuine reach, temporary-only, and unreachable in your own notation, with one-line reasons on the contested cells and the die-side row honest about its single reach — replace, or decline when no part exists (bga-failure-modes).
  2. Sort four case files (5 minutes, desk, the notebook). Decide commit, decline, or hand off — with class call and envelope reasoning — for: an e-reader with warm-up flakiness and a pristine history; a laptop that drowned a week after a drop; a set-top receiver dead since a violent power event, whose joints verify clean at every reachable test and whose package has no obtainable replacement; and a console with classic cycling symptoms whose replacement board costs less than an hour of bench time (when-to-remove-a-chip-vs-replace-the-board).
  3. Write the resurrection explainer (5 minutes, desk). Draft the customer-facing paragraph: why the oven trick rises, why it relapses, what it quietly costs the board, and what the envelope's genuine options are — in words a non-technician would keep reading (what-is-bga).
  4. Price two quotes (5 minutes, desk). Write one commit quote — class, plan, stated success fraction, named collateral, replacement anchor — for a cycling-cracked automotive infotainment unit with intact pads; and one decline — reason, alternative, respect — for a drop-and-drown laptop; then reprice the decline's device assuming it carries the owner's un-backed-up business records, and note what changed (repairability-assessment).

These core steps — the envelope map, the case sort, the explainer, and the paired quotes — are tested in the Chapter Quiz at the end of this chapter, where a score of 80% is required to continue.

Key Takeaways

  • The repair envelope maps renewals to failures: reflow completes what never formed, reballing renews fatigued ball metallurgy over intact pads, replacement renews the package — and each reach ends where its renewal does (bga-failure-modes).
  • The unreachable tier is real: cratered pads, torn laminate, and broken fanout vias belong to the board's bench or the replacement decision, and die-side death leaves replacement — or an honest 'no repair exists' (repairability-assessment).
  • The blind reflow rises briefly and relapses on schedule — old metallurgy refrozen under an unchanged standoff — and what separates it from genuine reflow is the class diagnosis, not the temperature (what-is-bga).
  • The arithmetic has named terms — stated success probability, the volume's heaviest labor, collateral risk that can convert repairable to unreachable, the replacement anchor, and the data payload that reprices everything (when-to-remove-a-chip-vs-replace-the-board).
  • The discipline is the section: class, envelope, price — then commit with honesty, decline with a reason, or hand off — because the declined job costs a quote and the misjudged one costs the board.

Skills Learned

  • You can now map failure classes onto the interventions that genuinely reach them.
  • You can now identify board-side and die-side damage outside every ball-level envelope and name the handoffs.
  • You can now explain the blind reflow's resurrection and relapse to a customer without scorn.
  • You can now price a BGA job with stated probability, labor, collateral, and the replacement anchor.
  • You can now run the decision discipline end to end and write both quote languages.

Glossary Additions

  • repair envelope — the set of failure classes a given intervention can genuinely reach, drawn from what the intervention renews: reflow renews joint formation — completing head-in-pillow, cold, and unwetted joints where the non-wet was flux or oxide rather than a compromised pad finish, a real repair because the failure was an unfinished process — but not metallurgy, so fatigue-cracked fields re-melted are old solder refrozen; reballing renews the ball metallurgy over intact pads, the genuine answer to cycling-cracked fields; replacement renews the whole package, the only reach into die-side death; and board-side damage — cratered pads, torn laminate, broken fanout vias — sits outside every ball-level envelope, belonging to the board's own repair bench or the replacement decision. The envelope turns 'can we repair this?' into the answerable 'does this intervention's renewal match this failure's mechanism?' — and an intervention aimed outside its envelope is effort spent rehearsing the next failure.
  • blind reflow — heat applied to a BGA without a class diagnosis — the oven trick, the heat-gun resurrection — which genuinely revives devices briefly by re-fusing the touching faces of fatigue cracks or accidentally completing near-merged pillowed joints, and relapses on schedule because every root cause survives the miracle: the metallurgy stays aged and voided, the standoff and die-shadow stresses stay unchanged, and the uncontrolled bake spends collateral on every plastic and neighbor for a non-repair. What separates a blind reflow from the repair envelope's genuine reflow is not temperature but diagnosis and discipline — known failure class, chosen profile, verified outcome — and the bench skill it demands is the customer conversation: explaining the brief rise and the scheduled relapse without scorn, so the trick prices itself against the real options.
  • collateral risk — the chance that a rework attempt damages what it did not target: heavy BGA interventions heat neighboring components, stress connectors and plastics, flex the board, and risk the under-package pads and fanout vias whose damage would move the job from repairable to unreachable — one attempt's collateral can end every envelope the assessment had open. Collateral risk is a named term in the honest quote — stated alongside success probability so the customer decides with both numbers — and it is the line of the arithmetic that routes learning to donor boards, keeps first attempts off data-bearing devices, and prices a second attempt more skeptically than the first, because collateral accumulates even when repairs do not.

Suggested Next Sections

Must read next:

  • BGA in Modern Consumer Electronics — Section 2.5 closes the theory chapter with the survey: where BGA and its shrinking cousins actually live in phones, laptops, consoles, and the wider bench queue, what underfill and soldered-down-everything mean for the envelopes just drawn, and how the device landscape sets the strategy for the hands-on chapters ahead.

Recommended: