The Repair LibraryRead · Learn · Master

BGA Removal with IR Rework Station

The chapter's second removal craft belongs to a machine built for exactly this work, at Professional depth: the IR rework station — a bottom-side infrared array that founds the whole board, a top-side emitter that heats the target through radiation instead of airflow, a fixture that holds the board in rails with the sag supports a hot board needs, and thermocouple feedback that turns the profile from a hand-flown pass into a closed-loop program. The station's gifts are real: no airflow means no scatter, no spilled plume, and no neighbor blown sideways; radiant zones heat broad targets evenly; and a profile that ran correctly once can run correctly again, board after board, because the machine flies it from the thermocouple's feedback rather than an operator's wrist. The section's Professional depth lives in what radiation changes: IR heats surfaces by what they are — differential absorption means the dark overmold of the target drinks the emitter's energy while shiny cans and bright connectors reflect it away, a physics that makes some shields easier (foil reflects IR excellently) and some targets harder (a reflective lid couples poorly until matte tape darkens it), and that puts the thermocouple's testimony above any assumption about what the emitter's power is doing. The fixturing earns its own discipline: rails hold the board, sag supports stand under it, because a large board at profile temperature droops under its own weight between unsupported spans — and a sagged board is a warped board wearing the station's fixtures as an alibi. The operation itself keeps every law the hot-air section wrote: the patient inversion, the bottom-founded profile, the molten-field check — run through the station's access or with its lift assist — and the zero-force lift, now often the machine's own vacuum arm doing what the law demands. What the station adds is repeatability with accountability: programs saved per board class, logs the case file can cite, and an operator whose craft has moved from the wrist to the setup, the verification, and the honest judgment of what the machine's log actually proves.

ProfessionalMedium Risk23 min read

What You Will Learn

  • You will learn the station's anatomy — bottom array, top emitter, rails, feedback — and what each part buys.
  • You will learn differential absorption — why IR heats by surface, and how tape and foil manage what airflow never asked.
  • You will learn the fixturing discipline — rails, sag supports, and the droop a hot board writes without them.
  • You will learn the closed-loop profile — programs flown from thermocouple feedback, verified rather than trusted.
  • You will learn the operation under the standing law — check, zero-force lift, and the logs that make repeatability accountable.

What You Will Be Able To Do

  • You will be able to set up an IR station — fixture, supports, thermocouples, shields — for a real board.
  • You will be able to manage differential absorption with matte tape on targets and foil on neighbors.
  • You will be able to build and verify a closed-loop removal profile rather than trusting the emitter's dial.
  • You will be able to run the removal under the full law — check, lift, aftermath — through the station's workflow.
  • You will be able to read the station's log honestly and write what it proves into the case file.

Required Tools

  • The IR rework station — bottom array, top emitter, rails, and its thermocouple feedback loop
  • Sag supports for the board's unsupported spans — the fixture's discipline, not its accessory
  • Matte-dark kapton for reflective targets and foil for IR-reflective shielding — absorption managed, not assumed
  • The station's vacuum lift arm or the pen and tweezers — the zero-force law's instruments either way
  • The case file, profile programs, and the station's logs — repeatability is recorded or it is folklore

When NOT to Attempt This

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

  • You are not experienced with the specific repair type described here.
  • You do not have professional-grade equipment for this procedure.
  • The device has sentimental or high monetary value and you cannot afford a mistake.
  • You have not successfully completed this repair on a sacrificial device first.

Section Overview

The dedicated machine arrives at Professional depth: the IR rework station — radiation replacing airflow (bga-removal-with-hot-air). The trade is real. No scatter, no plume, no blown neighbors, broad even zones — bought with radiation's physics (hot-plates-and-ir-preheaters). Differential absorption is the new management. Dark targets drink, shiny surfaces reflect — matte tape couples a reflective lid, foil shields excellently, and the thermocouple outranks every assumption (thermal-cameras-hotspot-detection-in-repairs). The fixture has its own law. Rails hold, sag support stands under the heavy spans — a hot board droops between them, and sag is warp wearing an alibi. The closed loop flies the profile. Programs from thermocouple feedback — repeatable, logged, verified rather than trusted (temperature-controlled-vs-fixed-temperature-stations). And the law survives the machine. Check, zero-force lift, honest aftermath — with logs the case file can cite. Trade, absorption, fixture, loop, law — the station entire.

Why This Matters

The IR station is what the removal craft looks like when a bench does this work often enough to buy repeatability, and knowing its physics is what separates operating it from owning it (bga-removal-with-hot-air). This matters because the airflow problems genuinely vanish: the §5.1 scatter, spill, and plume disciplines exist because air moves — radiation does not, and the boards that fought the nozzle's compromises get an even, still heat the plume could never offer (hot-plates-and-ir-preheaters). This matters because absorption is the new spill: IR heats by surface — a matte black SoC and a bright shield can sit side by side in the same zone at wildly different temperatures, and the operator who assumed the emitter's power reached everything equally has cooked one neighbor and starved the target (thermal-cameras-hotspot-detection-in-repairs). It matters because sag is the station's quiet failure: rails at the edges, heat in the middle, gravity all the while — the unsupported span droops exactly when the laminate is most compliant, and the droop freezes in as the warp the preheat law existed to prevent. It matters because closed-loop is not self-verifying: the loop controls what its thermocouple sees — one point, standing for a hundred joints — and the log records measurements, not achievements: the check still authorizes the lift, exactly as the honesty law demands (temperature-controlled-vs-fixed-temperature-stations). And it matters because repeatability is the Professional product: a profile proven once and saved per board class turns the next removal from a craft performance into a verified procedure — which is what customers with fleets of identical boards are actually buying. The machine changes the physics, not the law — learn both, and the station earns its floor space.

Required Prerequisites

  • BGA Removal with Hot Air — the removal law entire: patient inversion, shield thinking, the check, and the zero-force lift the station must still obey.
  • Hot Plates and IR Preheaters — the IR foundations: radiant heat's behavior, and the bottom-side founding this station builds into its frame.
  • Matte-dark kapton for reflective targets — to couple the emitter into lids that would reflect it away (thermal-cameras-hotspot-detection-in-repairs)
  • Aluminum foil for IR shielding — to reflect the emitter off neighbors that need nothing from it (hot-plates-and-ir-preheaters)
  • Thermocouple tape and spare junctions — to keep the loop's one honest sense placed and verified (temperature-controlled-vs-fixed-temperature-stations)
  • The profile-program notebook or the station's library — to save what worked per board class, because repeatability unrecorded is folklore (bga-removal-with-hot-air)
  • The IR station itself, or bench time on one — to learn the machine on donors before a customer board rides its rails (hot-plates-and-ir-preheaters)
  • Donor boards spanning sizes, including one large enough to sag — to see the droop the supports exist to prevent (bga-removal-with-hot-air)
  • A board with mixed surfaces — dark packages, bright cans, shiny connectors — to watch differential absorption do its uneven work (thermal-cameras-hotspot-detection-in-repairs)
  • The thermal camera where the bench owns one — to image the zone's actual evenness against the emitter's claim (temperature-controlled-vs-fixed-temperature-stations)
  • The bench camera and the station's logs — to file setups and profiles beside the boards they served

Real-World Applications

The station earns its place on the boards and the volumes that reward it. A bench that fought a server-class board's mass with hot air twice moves the third to the IR station: the bottom array founds the whole plane evenly, the top emitter works a still zone no plume disturbs, and the removal that took two tense passes becomes one logged profile (bga-removal-with-hot-air). A technician setting up a board with a bright shield can beside the target manages absorption instead of assuming it: matte tape darkens the target's reflective lid, foil tents the can — and the thermocouple confirms the target climbing while the neighbor idles (thermal-cameras-hotspot-detection-in-repairs). A shop with a fleet contract — forty identical gateway boards, same failed package builds the profile once, verifies it on the first article with the full check-and-lift law, saves it per the class, and runs the fleet as procedure with each board's log in its file (temperature-controlled-vs-fixed-temperature-stations). And a bench that watched a wide board droop between the rails on a practice run stands sag supports under the heavy spans before the next first degree — the droop that would have frozen in as warp now never happens (hot-plates-and-ir-preheaters). The confusions this prevents: a reflective lid starving under an emitter the dial swore was delivering, a bright neighbor cooked by assumption, a fleet flown by memory instead of program, and a board warped by its own weight between two proud rails.

Common Challenges

  • The machine invites trust it has not earned. A closed loop feels like a guarantee — it controls one thermocouple's point; the check still authorizes the lift, and the log records measurements, not achievements (temperature-controlled-vs-fixed-temperature-stations).
  • Absorption is invisible until measured. Two neighbors in one zone at wildly different temperatures look identical to the eye — the thermocouple and the thermal camera are the only honest witnesses (thermal-cameras-hotspot-detection-in-repairs).
  • The fixture's gaps hide in plain sight. Rails look like support — they hold edges; the middle spans sag at temperature, and the supports go under before the first degree or the warp explains it after (hot-plates-and-ir-preheaters).

Safety Notes

Risk Level: Medium. A radiant machine with a founded board — the burn and fume laws in full, plus radiation's own etiquette.

Professional Tips Before Starting

  • Walk the zone with the camera before the first real run. The emitter's evenness imaged, not assumed — ten minutes of thermal portraits teaches a station's personality for good (thermal-cameras-hotspot-detection-in-repairs).
  • Build the first-article habit. Fleet or not, the first board of any class runs under the full manual law — check, lift, site verdict — and only its verified profile earns the program library (temperature-controlled-vs-fixed-temperature-stations).
  • Stage the supports with the fixture, not after it. Sag supports live in the setup checklist — the droop happens exactly once per forgotten support (hot-plates-and-ir-preheaters).

The Station — Trade, Absorption, Fixture, Loop, Law

Recap and Frame

The hot-air section taught the removal law by hand — patient inversion included; this section hands the profile to a machine and keeps the law (bga-removal-with-hot-air). The IR foundations arrive laid. The lab volume introduced radiant heat's behavior and the bottom-side founding — the station builds both into a frame with feedback (hot-plates-and-ir-preheaters). The closed-loop idea arrives proven. Temperature control by feedback earned its case at iron scale in the stations chapter — the IR station is the same honesty applied to a whole board's profile (temperature-controlled-vs-fixed-temperature-stations). The emissivity lesson arrives trained. The thermal camera chapter taught that surfaces lie to radiation — dark and shiny read differently — and this section meets the same physics from the emitter's side (thermal-cameras-hotspot-detection-in-repairs). What is new is craft relocated. The wrist's orbit becomes a program; the operator's skill moves to setup, absorption management, verification, and the honest reading of logs — Professional depth as judgment over motion. And the order is the machine's order: fixture and supports, absorption managed, profile built and verified, then the operation under the standing law. Hold the frame — same law, new physics, relocated craft — and the station earns its floor space.

The Trade and the Fixture — What Radiation Buys and Demands

The station's first lesson is what changed when airflow left the operation (hot-plates-and-ir-preheaters). Radiation buys stillness. No plume, no scatter, no neighbor blown sideways, no spill hunting the unshielded connector — the top emitter heats a zone that simply does not move, and the §5.1 disciplines built around moving air relax into radiant equivalents. The zones are broad and even. The bottom array founds the whole board at once — the mandatory preheat built into the machine's floor — and the top emitter covers targets that made nozzles choose between coverage and flood (bga-removal-with-hot-air). The fixture holds and must support. Rails grip the board's edges; the spans between them hang free — and a board at profile temperature is compliant laminate that droops under its own weight exactly where nothing stands under it: the sag supports go beneath the heavy spans, touching bare board rather than components, standing before the first degree (temperature-controlled-vs-fixed-temperature-stations). Sag is warp with an alibi. The droop that happens at temperature freezes in at cool-down, and the station's own fixtures take the blame the missing support earned — the §5.1 warp law, arriving through gravity instead of gradient. The underside of the zone gets its own survey. On double-sided assemblies, components hanging inverted beneath the target can reach liquidus through the laminate at peak and drop into the bottom array — they are noted before fixturing, secured where the risk is real, or the profile's bottom contribution bounded to spare them. The aperture is part of the setup. The top emitter's zone sized or masked to the target's neighborhood — radiant coverage chosen the way nozzles were, deliberately, and walked with the camera, because tall neighbors cast real shadows inside a radiant zone: line-of-sight is radiation's one airflow-like caveat, and the walk finds the shadows before a profile meets them (thermal-cameras-hotspot-detection-in-repairs). Still, broad, held, supported — the trade and fixture entire. The station removes the air and keeps the gravity — support the spans, and the machine's stillness becomes the board's friend.

The Absorption — Managing What Surfaces Do to Radiation

IR heats surfaces by what they are, and managing that is the section's distinctly Professional skill (thermal-cameras-hotspot-detection-in-repairs). Dark drinks, shiny reflects. The matte black overmold of the usual target couples the emitter's energy beautifully; bright shield cans, shiny connector shells, and reflective lids send it away — differential absorption, the same emissivity physics the camera chapter read, now writing real temperature differences across one zone. The target is coupled deliberately. A reflective lid under-heats while the dial claims delivery — matte-dark kapton on the target's top converts the claim into coupling, the two-minute fix for the starving-target mystery (hot-plates-and-ir-preheaters). The neighbors are shielded in IR's own terms. Aluminum foil, a mediocre argument against hot air's convection, is an excellent IR shield — it reflects the emitter away — while bright neighbors partially self-protect by their own reflection: the shield map survives, redrawn for radiation (bga-removal-with-hot-air). Conduction still travels. The §5.1 warning holds under any heat source — shields manage radiation, not the board-borne heat that reaches adjacent hidden-joint packages through the laminate — and the flanking BGA's protection remains awareness, coverage discipline, and stillness until cool. The thermocouple outranks the dial, doubly. Under differential absorption, emitter power and target temperature diverge by surface — the taped junction at the package is the only witness that saw the target, and the camera's zone portrait is the only witness that saw the neighbors (temperature-controlled-vs-fixed-temperature-stations). Coupled, shielded, conducted, witnessed — the absorption entire. Radiation asks a question air never did — what color is your heat going to — and the bench that answers it on purpose owns the station.

The Loop and the Law — Programs, Verification, and the Lift

The closed loop is the station's signature, and its honesty rules are the section's close (temperature-controlled-vs-fixed-temperature-stations). The program flies the profile. Thermocouple feedback drives the emitters through the card's stages — base, soak, climb, bounded peak — the same profile law the volume has flown by hand, now executed by a controller that does not get tired, distracted, or optimistic. The loop controls what it sees. One junction's point stands in for a hundred hidden joints — the loop holds that point to the program, and everything beyond the junction is inference: the closed loop is honest about its thermocouple, and silent about everything else (bga-removal-with-hot-air). The check survives automation. The molten-field check still authorizes the lift — run through the station's access with the same tweezer-light law, or trusted to the station's lift assist only after the bench has verified what that assist actually senses — because no program's completion certifies a hundred joints the loop never saw (hot-plates-and-ir-preheaters). The lift keeps its law. Zero force governs the machine's vacuum arm exactly as it governed the pen — the arm that lifts on the program's schedule rather than the check's answer is automation performing the exact mistake the law exists to prevent. The log is read, not worshiped. The station records what was measured — junction temperatures against time — and the case file cites it as exactly that: evidence of the profile, never proof of the joints, with the check, the lift, and the site verdict still writing the lines the log cannot (thermal-cameras-hotspot-detection-in-repairs). First articles earn the library. A profile proven under the full manual law — check, lift, site verdict clean — gets saved per board class, and the fleet that follows runs as verified procedure with every board's log in its file. Programmed, bounded, checked, lifted, logged — the loop and law entire. The machine automates the profile and the operator keeps the verdicts — that division is the whole Professional grade.

Common Mistakes

  • Trusting the dial through a reflective lid. Emitter power read as target temperature — differential absorption divorces them; the taped junction is the only witness that saw the target (thermal-cameras-hotspot-detection-in-repairs).
  • Rails without supports. Edges held, spans hanging, heat rising — the droop happens at peak compliance and freezes in as warp; supports stand before the first degree (hot-plates-and-ir-preheaters).
  • Lifting on the program's schedule. The profile complete, the arm commanded, the check skipped — no program certifies joints the loop never saw; the check authorizes, the program only heats (bga-removal-with-hot-air).
  • Reading the log as proof of joints. 'Profile nominal' cited as 'removal verified' — the log proves the junction's history; the check, lift, and site verdict prove the operation (temperature-controlled-vs-fixed-temperature-stations).
  • Fleet-running an unproven profile. Board two through forty on a program board one never verified — first articles earn the library under the full manual law, or the fleet inherits a folklore profile at scale.

Troubleshooting Guidance

The station troubleshoots by its own physics: starving, uneven, drooping, disagreeing. If the target climbs slower than the program expects: look at its surface before its zone — a reflective lid decouples the emitter, and matte tape fixes in two minutes what power escalation would fix by cooking the neighborhood (thermal-cameras-hotspot-detection-in-repairs). If neighbors run hot in a zone the map called safe: re-shield in IR's terms — foil reflects where kapton only insulates — and re-walk the zone with the camera, because absorption differences the eye cannot see the portrait can (hot-plates-and-ir-preheaters). If a board comes off with new warp: audit the spans, not the profile — the supports' placement against the board's heavy regions is almost always the answer, and the profile takes blame the fixture earned (bga-removal-with-hot-air). If the loop and a second measurement disagree: verify the junction before believing either — placement, tape, and contact first; a lifted junction under-reads and the loop obediently overheats to chase it (temperature-controlled-vs-fixed-temperature-stations). If the check argues after a completed program: the program was a hypothesis and the check just falsified it — extend under the card's bounds, re-check, and revise the saved profile, because the library holds what the law verified, not what the controller hoped. The throughline: the station's mysteries resolve at surfaces, spans, and junctions — the three places radiation's honesty lives.

Verification & Testing Methods

Confirm the station craft before the profile-design section builds on it:

  • [ ] I can set up the IR rework station — rails, aperture, thermocouples verified — for a real board.
  • [ ] I can place sag support under the heavy spans, on bare board, before the first degree.
  • [ ] I can manage differential absorption — matte tape coupling targets, foil reflecting neighbors, the camera walking the zone.
  • [ ] I can build a closed-loop profile, verify it as a first article under the full manual law, and save it per board class.
  • [ ] I can run the removal through the station — check authorizing, zero force governing the arm, log cited as evidence, not proof.

Then try the practice exercises below — station work on donors; scenarios differ from the quiz.

Practice Exercises

  1. Fixture and walk the zone (6 minutes, a donor board, the station, camera if owned). Rail the board, stand sag supports under its heavy spans on bare laminate, tape and verify the junction at the package edge, size the top aperture — then walk the zone's evenness with the thermal camera or a moved junction, filing the station's portrait before any profile runs (hot-plates-and-ir-preheaters).
  2. Manage the absorption (7 minutes, a mixed-surface donor, tape and foil). Survey the zone's surfaces — dark targets, bright cans, reflective lids — couple the target with matte-dark tape where it needs it, foil the neighbors that want the emitter reflected away, and verify with junction or camera that the zone now heats what the plan intended (thermal-cameras-hotspot-detection-in-repairs).
  3. Build and fly a first article (7 minutes, the fixtured donor). Program the removal profile from the card — base, soak, climb, bounded peak — fly it on the loop while watching the junction's narration, run the molten-field check at liquidus-plus-margin through the station's access, and execute the zero-force lift with arm or pen only on the check's answer (bga-removal-with-hot-air).
  4. Judge the log and close the record (5 minutes, the station's log, the case file). Read the run's log as evidence — junction history against the card — write what it proves and what it cannot, enter the site verdict and exposure per the standing law, and save the verified profile to the library with its board class named (temperature-controlled-vs-fixed-temperature-stations).

These core steps — the fixtured zone, the managed absorption, the verified first article, and the honest log — are tested in the Chapter Quiz at the end of this chapter, where a score of 80% is required to continue.

Key Takeaways

  • The IR rework station trades airflow for radiation — no scatter, no plume, broad even zones, built-in founding — and keeps every law the hot-air removal wrote (bga-removal-with-hot-air).
  • Differential absorption is the new management: dark drinks and shiny reflects, matte tape couples starving targets, foil reflects the emitter off neighbors — and the taped junction outranks the dial, doubly (thermal-cameras-hotspot-detection-in-repairs).
  • The fixture demands sag support under the heavy spans before the first degree — a hot board droops where nothing stands, and sag is warp wearing the fixtures as an alibi (hot-plates-and-ir-preheaters).
  • The closed loop flies the profile from one junction's testimony — repeatable and logged — but controls only what it sees: the check still authorizes, zero force still governs the arm, and the log is evidence, never proof (temperature-controlled-vs-fixed-temperature-stations).
  • First articles earn the library: a profile verified under the full manual law gets saved per board class, and the fleet that follows runs as accountable procedure — the craft relocated from wrist to judgment.

Skills Learned

  • You can now fixture a board with the supports its hot weight demands.
  • You can now manage what surfaces do to radiation instead of assuming delivery.
  • You can now build, verify, and library a closed-loop removal profile.
  • You can now keep the check and the zero-force law in charge of an automated lift.
  • You can now cite a station log for exactly what it proves.

Glossary Additions

  • IR rework station — the dedicated board-rework machine that replaces hot air with radiation: a bottom-side infrared array founding the whole board — mandatory preheat built into the floor — a top-side emitter heating the target zone without airflow, rails fixturing the board with sag supports under its spans, and thermocouple feedback flying programmed profiles through a closed loop that logs every run. The station buys stillness — no scatter, plume, or blown neighbors — plus broad even zones and repeatability — profiles proven as first articles under the full manual law and saved per board class — at the price of radiation's own physics: differential absorption must be managed, the loop controls only the point its junction sees, and every law the hot-air removal wrote — the check, the zero-force lift, the site verdict — survives the automation intact, because the machine automates the profile, never the responsibility.
  • differential absorption — the radiant-heat physics that makes IR heat surfaces by what they are: matte dark surfaces — the usual target package's overmold — couple the emitter's energy strongly, while bright cans, shiny connector shells, and reflective lids send it away, writing real temperature differences across a single zone that the eye cannot see and the emitter's dial cannot report. The management is deliberate: matte-dark kapton couples a reflective target that would otherwise starve under a dial claiming delivery, aluminum foil — mediocre against convection — shields excellently under IR by reflecting the emitter away, and the taped junction plus the thermal camera's zone portrait are the only honest witnesses to what the radiation actually did. Conduction is unmoved by any of it: board-borne heat still reaches adjacent hidden-joint packages through the laminate, and their protection remains awareness and stillness.
  • sag support — the fixture discipline the IR station's rails make necessary: supports standing under a board's unsupported spans — placed on bare laminate, never on components, before the first degree — because a board at profile temperature is compliant laminate that droops under its own weight exactly where nothing holds it, and the droop that forms at peak compliance freezes in at cool-down as permanent warp. Sag is the station's quiet failure mode — warp wearing the fixtures as an alibi, blamed on profiles that never caused it — and the discipline is structural: supports live in the setup checklist beside the junction check, sized to the board's heavy regions, standing before any heat exists, because the droop happens exactly once per forgotten support.

Suggested Next Sections

Must read next:

  • Thermal Profile Design for BGA Reflow — Section 5.3 builds the profiles both removal crafts have been flying: Professional profile construction from alloy and board upward — stages, targets, ramps, and the verification that turns a recipe into a measured reality.

Recommended: