Logic board 🔬 Technical investigation June 2026 14 min read

MacBook „burned CPU" — why this diagnosis is wrong in 95% of cases

You hear from a shop: „the CPU is burned, you need a new board" — followed by an estimate that climbs sharply to high figures. Statistically, in 95 out of 100 cases the diagnosis is fundamentally wrong. Complete technical article: the 5 chips that actually fail (PMIC, CD3215, ISL9239, T2, MOSFET), correct 6-step diagnostic procedure, when CPU IS actually burned (rare, ~2-3% of cases) and how to verify the diagnosis you received.

Alin Marin — CEO Radical Service
Alin Marin CEO Radical Service

Peste 12 ani experiență în reparații Apple · 25.000+ dispozitive reparate · ASE București

MacBook burned CPU — board-level diagnostic under microscope
95%
„CPU burned" diagnoses are wrong
5 chips
that actually fail
6 steps
of correct diagnosis
~2-3%
cases where CPU IS actually burned
⚠️

Quick test for you: if diagnostic took under 10 minutes, it\'s suspect

A real board-level diagnostic for MacBook takes 30-90 minutes and includes 6 systematic steps (microscope, LCI, multimeter on rails, thermal camera, oscilloscope, BoardView). If the shop told you „CPU is burned" in 5-10 minutes without showing thermal camera photos or measured voltage values — they didn\'t do diagnostic. They looked at the board with naked eye and drew the commercially advantageous conclusion: complete board replacement.

Part 1 — Why „CPU burned" is universal answer for shops without equipment

On MacBook, when it doesn\'t power on at all, observed symptoms from outside are identical regardless of real cause: no LED, no fan, no chime, no response to USB-C charger. That\'s exactly why „CPU burned" becomes universal label — it sounds credible, sounds technical, and justifies the high cost of board replacement.

Technical reality is different. A MacBook doesn\'t power on if any of the following chips fail: PMIC, charging IC, USB-C PD controller, power MOSFET, T2 security chip, or in rare cases, the SoC itself. All produce same external symptom, but each requires different repair, at different cost, in different time.

The difference between correct diagnosis and generic „CPU burned" shows through 5 specific instruments: stereo microscope, multimeter, thermal camera, oscilloscope, BoardView + schematics. Shops without these equipment can\'t do board-level diagnosis. They can only swap parts: display, battery, keyboard, maybe SSD. When a MacBook comes that won\'t power on, their solution is: „board dead, buy another" or „complete board swap at us" — at a significant cost, often exceeding the real cost of a correct board-level repair.

Part 2 — The 5 chips that actually fail (in order of frequency)

Real statistics of MacBook „no power" defects arriving at the workshop, compiled from our reports:

1

PMIC (Power Management IC)

Intel code: U7000 / U7100 / U7800 (Intel)

Apple Silicon: SMC integrated in SoC (Apple Silicon) + discrete PMIC for USB-C

~35% of "no power" cases

Role: Distributes base voltages: PP3V3_G3H, PPBUS_G3H, PP1V8, PP1V05. The heart of board power supply.

Typical symptom: MacBook doesn't power on at all, no LED, no chime, no thermal hotspot — or hotspot localized on PMIC.

How we diagnose: Fluke multimeter on PP3V3_G3H (3.36V ideal). Missing = burned PMIC or shorted output capacitor.

Repair: Microsoldering PMIC replacement + adjacent capacitor check. Much lower cost than board replacement.

2

CD3215 / CD3217 (USB-C PD Controller)

Intel code: CD3215C00 / CD3215B03 (Touch Bar 2016-2019)

Apple Silicon: Apple Silicon: proprietary chips for USB-C PD, similar function

~25% of MacBook 2016-2019 cases

Role: Negotiates USB-C Power Delivery protocol with charger. Failure = MacBook stops charging.

Typical symptom: Won't charge on any USB-C port, or charges only on one. Typical after overvoltage, cheap charger or fall.

How we diagnose: Thermal camera over CD3215 — hotspot when charger connects = chip with internal short. USB-C analyzer test (AVHzY).

Repair: Microsoldering CD3215 replacement. Part itself costs little, labor is what matters.

3

ISL9239 / ISL9240 (Charging IC)

Intel code: ISL9239HRZ / ISL9240HRZ

Apple Silicon: Similar — Intersil chips for MacBook

~15% of "battery not charging" cases

Role: DC-DC converter that transforms USB-C input into battery charging voltage. Works at high current (3-5A) and naturally heats up.

Typical symptom: Battery won't charge even though MacBook powers on AC. Or charges inconsistently with random cutoffs.

How we diagnose: Rigol oscilloscope to see PWM at switch pins. Missing = burned chip. Adjacent MOSFET verification.

Repair: Microsoldering ISL9239 replacement.

4

T2 Security Chip (Intel 2018-2020) / Secure Enclave (Apple Silicon)

Intel code: Apple T2 — ARM SoC with Secure Enclave

Apple Silicon: Secure Enclave integrated in SoC M1/M2/M3/M4

~10% of Intel 2018-2020 cases

Role: Controls SSD, Touch ID, FileVault encryption, microphone, camera. T2 defects = bricked MacBook or missing functions.

Typical symptom: MacBook powers on but doesn't detect SSD, or doesn't detect Touch ID, or won't boot from any media. "No bootable device" error.

How we diagnose: Apple Configurator 2 + DFU. If DFU Revive fails, T2 has physical issue.

Repair: Controlled T2 reflow or, in severe cases, T2 replacement with re-pairing. Difficult but feasible.

5

Input/output MOSFETs (multi-channel)

Intel code: Q7050, Q7051 (USB-C input), Q7700-Q7710 (output)

Apple Silicon: Similar MOSFETs on M-series boards

~10% of complex cases

Role: Semiconductor switches that route voltages. First to fail from mains overvoltage or non-original chargers.

Typical symptom: Random behavior: powers on sometimes, suddenly shuts off, one USB-C port dead but others OK, or complete "no power".

How we diagnose: Multimeter in ohm mode: short resistance source-drain = burned MOSFET. BoardView verification for exact identification.

Repair: Individual MOSFET replacement with microsoldering.

Total: these 5 categories cover ~95% of real „MacBook won\'t power on" cases. ~5% remain for: actually burned SoC, display board defect, delaminated internal connectors, delaminated RAM (BGA), or other rare causes.

Part 3 — Comparison table: what the backstreet shop says vs technical reality

❌ What the backstreet shop says

„The CPU is burned, you need a new board"

Cost: Complete board swap (high cost) or „buy another MacBook"

✓ Technical reality

95% of the time it's PMIC, CD3215, ISL9239 or MOSFET

Real cost: Board repair (board-level intervention) — significantly cheaper than board swap

Why they\'re wrong: They don't have microscope to see the board, no BoardView/schematics, no thermal camera. „CPU burned" = universal label for „I don't know how to diagnose".

❌ What the backstreet shop says

„Apple Silicon M1/M2/M3 can't be repaired"

Cost: Send to Apple, very high cost

✓ Technical reality

Board-level repairs on Apple Silicon are possible for ~60% of defects: discrete PMICs for USB-C, power MOSFETs, internal connectors, display board, WiFi/BT antennas

Real cost: Much lower than Apple board swap

Why they\'re wrong: Smaller BGA + more integrated architecture = more difficult, but NOT impossible. Requires professional equipment + experience.

❌ What the backstreet shop says

„Nothing we can do, data is lost"

Cost: External specialized data recovery firm — separate cost

✓ Technical reality

On pre-T2 Intel MacBook: detachable SSD, direct recovery. On T2/Apple Silicon: soldered SSD, but if we repair the board, data remains intact in ~90% of cases.

Real cost: Included in board repair if data is accessible

Why they\'re wrong: They don't have capability for proper DFU Revive, don't know how to use Apple Configurator 2.

❌ What the backstreet shop says

„The CPU is physically broken"

Cost: Unusable, must be thrown away

✓ Technical reality

Physically broken processor is EXTREMELY rare (direct fall on exact CPU spot + major impact). Typical „broken" is delaminated BGA pads — recoverable through reballing.

Real cost: Reballing possible if BGA isn't totally destroyed

Why they\'re wrong: Confuses „delaminated BGA pads" (recoverable) with „physically broken chip" (rare and irreversible).

❌ What the backstreet shop says

„It went into short, everything is burned"

Cost: „Total loss" — recommends complete board swap

✓ Technical reality

Typical short is on a single burned capacitor or MOSFET. Thermal camera identifies it in minutes. Repair = just the defective component.

Real cost: Just the defective component repair, part of „board repair" service

Why they\'re wrong: No FLIR/Seek thermal camera + lack of experience tracing shorts on specific rails.

Part 4 — Our 6-step diagnostic procedure

Here\'s exactly what we do when we receive a MacBook „won\'t power on". Procedure takes 30-90 minutes, cost 180 RON, included in repair price if you continue with us:

  1. 1

    Visual inspection with microscope (5-10 min)

    4-40× stereo microscope to identify oxidation marks, burned SMD capacitors, visibly delaminated BGA pads, tilted connectors. If we see oxidation = liquid, it's a different story (we acknowledge honestly it's not PMIC).

    → Output: Confirmation of presence/absence of visible external causes.

  2. 2

    LCI (Liquid Contact Indicator) check

    LCI indicators are 4-6 points on the board that turn red on liquid contact. If triggered = device contacted liquid, not spontaneous processor defect. Repair becomes different category (ultrasonic cleaning + reconstruction).

    → Output: Clear answer to „liquid or not?".

  3. 3

    Multimeter on key rails (10-15 min)

    With Fluke 87V multimeter we systematically check voltages: PP3V3_G3H (3.36V), PPBUS_G3H (12.5V), PP1V8_S5 (1.8V), PP1V05_S0 (1.05V). Each missing rail indicates a specific chip. We directly attribute the defective chip.

    → Output: Exact identification of defective chip — 80% of cases solved here.

  4. 4

    FLIR thermal camera for hotspots (5 min)

    We put the board under power (even with external sources) and scan with FLIR One Pro thermal camera. Hotspots = chip with internal short (shorted capacitor, defective MOSFET). We identify without touching anything — purely observational.

    → Output: Thermal map of board — visual confirmation of defective chip.

  5. 5

    Oscilloscope for complex analyses (15-30 min)

    For cases where rails are present but behavior is intermittent: Rigol oscilloscope to see PWMs, I²C communications between SMC and PMIC, boot signals. Especially useful for T2 defects or intermittent issues.

    → Output: Signal-level diagnosis, not just static voltage.

  6. 6

    BoardView + schematics confirmation (10 min)

    With BoardView (3D board visualization) and official Apple schematics we identify: exact required part, BGA pad count, affected traces, adjacent capacitors. Then we prepare the estimate with part cost + labor for the specific intervention.

    → Output: Written report with technical defect description and transparent estimate.

Part 5 — When the CPU IS actually burned (honesty)

Intellectual categoricity: there are real cases of burned CPU/SoC. They\'re rare (~2-3% of total „won\'t power on") and have verifiable characteristics. Here are the 4 scenarios when „CPU burned" diagnosis is correct:

1

Direct fall on CPU body + major impact

~1-2% of cases

MacBook dropped from 1.5m+ directly on CPU side (rare, but possible). Verifiable: MacBook lid deeply bent physically exactly on CPU zone, impact marks on body.

2

Corrosive liquid directly on CPU + prolonged time

~1% of cases (with triggered LCI + confirmed diagnostic)

Acid from swollen battery that directly contacted CPU BGA pads, or Coca-Cola/sugar that created permanent conductive bridge. Visible through deep green oxidation under CPU.

3

Extreme overvoltage (direct lightning strike)

~0.5% of cases

Very rare: lightning that exceeded bypass capacitor protections and reached CPU/SoC. Typically associated with multiple chips burned simultaneously (PMIC + CD3215 + CPU).

4

Intrinsic factory defect (known gates)

~2% on models affected by public gates

Only real technical category: mass defects known on certain generations (see Apple Hardware Defects Timeline 2010-2026). E.g.: GPU-gate MacBook Pro 2011 — AMD chip actually burned due to abnormal temperatures.

In these cases, our diagnostic includes: confirmation through thermal camera (or absence of hotspot on SoC + absence of all rails means complete death), test with external sources (we power the board directly, bypassing PMIC, to isolate SoC), correlation with history (confirmed physical fall, triggered LCI liquid, etc.). Only then can we affirm with certainty: „CPU/SoC compromised, board can no longer be recovered".

Part 6 — How YOU verify the diagnosis received from another shop

  1. Ask for thermal camera photo of board under power. Serious shop does this test mandatorily. If they don\'t have thermal camera, they can\'t diagnose board-level.
  2. Ask them to tell you the exact missing rail. Acceptable answer: „PP3V3_G3H missing" or „PPBUS_G3H under 5V". Suspect answer: „has no voltage" (generic) or „processor doesn\'t receive power" (vague).
  3. Ask for the exact model of the defective chip. „CD3215C00 with delaminated pads" vs „component on the board" — first is diagnosis, second is words.
  4. Ask explicitly: is it „board repair" or „board swap"? Two completely different things. Board repair = board-level intervention with microsoldering on the defective chip, your original board stays (with T2 pairing preserved, data intact, resale value retained). Board swap = complete board replacement, costs much more, requires re-pairing and you lose device continuity. Ask for the price for board repair, not for board swap. If the shop tells you the repair isn\'t an option, it means they can\'t do it — not that it isn\'t possible. Check our real prices for MacBook board repairs on the /en/macbook-repair-bucharest/ page or via online calculator.
  5. Ask for a second opinion. Bring MacBook to us with the report from the other shop. We redo the diagnostic independently, with photographic reports. Compare the two opinions — difference between concrete diagnosis and generic „CPU burned" is evident even for a non-specialist.

FAQ

How can I verify myself if the „CPU burned" diagnostic is correct or wrong?
Three free checks before paying anything: (1) Ask for thermal camera photo of board under power — serious shop has thermal camera and shows it to you. (2) Ask them to tell you the exact missing rail (PP3V3_G3H? PPBUS_G3H? PP1V05_S0?) — shop that correctly diagnoses identifies the rail, not just „CPU burned". (3) Ask them to show you the schematics for your board (Apple BoardView circulates in the industry) — if they don't have access, they can't do correct board-level diagnostic.
Why would a shop say „CPU is burned" if it's actually PMIC?
Two commercial reasons and one technical. (1) „CPU burned" diagnostic justifies the recommendation for complete board swap — a service significantly more expensive and more profitable for them than a board-level microsoldering repair. (2) Shop doesn't have equipment (microscope, thermal camera, oscilloscope, BoardView) to do real diagnostic, so „CPU burned" is universal answer that doesn't require technical justification. (3) Technically: for someone without training, burned PMIC symptoms (no power, no fan, no LED) look like „burned CPU" at first glance.
On Apple Silicon (M1/M2/M3/M4) is nothing really repairable anymore?
False. On Apple Silicon the SoC (CPU+GPU+RAM+NPU integrated) is more difficult to replace individually — practically the SoC component itself isn't replaced, but the problems that arise aren't typically in the SoC. Most frequent Apple Silicon defects: discrete PMICs for USB-C (repairable), power MOSFETs (repairable), internal connectors (repairable), display board (replaceable), WiFi/BT antennas (repairable). Only if the SoC itself is physically dead (extremely rare) then yes, board becomes economically non-repairable.
How long does a real diagnostic take, not „eyeballed"?
At us: 30-90 minutes for complete board-level diagnostic. Includes the 6 steps described above (microscope, LCI, multimeter, thermal camera, oscilloscope, BoardView). Diagnostic cost: 180 RON, included in repair price if you choose to continue. At shops that say „CPU burned" in 10 minutes: means they did NOT do diagnostic — they looked at the board with naked eye and drew the universal conclusion.
What are my chances to recover data if the CPU really is burned?
Depends on MacBook model. (1) Pre-2018 Intel with detachable SSD: data is safe on SSD even if board is dead — we extract SSD and read with external adapter. (2) 2018-2020 Intel with T2: soldered SSD + hardware encrypted through T2. If T2 is dead = data is lost. If T2 lives but something else is defective = we can repair board and data remains intact. (3) Apple Silicon M1+: SSD in SoC package + Secure Enclave. Typical defects (discrete PMICs, USB-C, MOSFET) do NOT affect SSD. We repair and you return with all data.
Shop X told me „we have 99% repair rate, it's definitely CPU burned". How do I verify?
Ask for concrete data: how many board-level boards do they repair monthly? Do they have equipment in workshop they can show (microscope, thermal camera, hot air station, BGA station)? Do they have microsoldering certifications (IPC-7711)? Look for Google Reviews where customers mention successful board-level repairs. Shop that only swaps parts (display, battery) has NO real microsoldering experience — they can't correctly diagnose a board defect.
Real difference: board repair vs board swap?
Two distinct services, NOT synonyms. (1) Board repair (board-level) = intervention on the defective chip with microsoldering, your original board stays. Advantages: T2 pairing preserved (data remains accessible), original device serial preserved (AppleCare+ continuity where applicable), resale value retained. We bill „board repair" as a single service with total price (part + labor + testing + VAT), not itemized by components. (2) Complete board swap = full replacement with new/refurbished board. Much more expensive, requires T2/Apple Silicon re-pairing, macOS re-installation, data loss if SSD was hardware-encrypted through the old T2. For exact prices of board repairs on your model, check the online calculator on the MacBook repair page or ask for written estimate.
How do I recognize shops that actually do microsoldering?
Simple checks: (1) Do they have Instagram/Facebook profile with real photos from microscope, boards in work, SMD capacitors replaced? Serious shop documents. (2) Do they have clear description of equipment (AmScope/Vision Engineering microscope, JBC/Quick hot air, MLink/Achi BGA station)? Facade shops can't detail. (3) Do they have technical articles written (not just marketing) about specific defects — PMIC, CD3215, T2, etc.? (4) Will they let you see the physical workshop before handing over the device.
If your diagnostic says „not CPU burned" but the previous shop insists it is — who is right?
Ask for both written reports. Shop that says „CPU burned" must be able to support diagnosis with: (a) specific missing voltage measurement on SoC, (b) thermal camera hotspot on SoC not on another chip, (c) BoardView for identification. We give you written report with exact defective chip, measured values, thermal camera photos. Compare the two reports — difference between a concrete diagnosis and generic „CPU burned" is obvious.

Useful resources

Second opinion with written report

Got „CPU burned" as diagnosis? Verify it before paying.

Bring MacBook to us (or via courier), we redo the diagnostic in 30-90 minutes with microscope, thermal camera, oscilloscope, BoardView. We give you written report with exact defective chip and transparent estimate. Diagnostic 180 RON, included in repair price if you choose to continue with us. If we confirm „CPU burned" — we give you detailed technical report so you can decide informed.