Board 820-02521 — MacBook Air 13" M1 2020

The 820-02521 is the primary logic board for the MacBook Air 13" M1 (2020, A2337), featuring Apple's first in-house M1 SoC paired with integrated 8GB unified memory and the CD3217 dual USB-C power delivery controller. This board introduced the M1 thermal ecosystem to the MacBook Air line and is notably sensitive to liquid ingress along the PPBUS_G3H distribution rail, which runs directly under the keyboard deck and into the power domain near the trackpad connector. Repair requires familiarity with M1-specific power sequencing, the ISL9239 main PMIC, and eDP display flex routing that runs tightly beneath the keyboard assembly.

Key Test Points & Voltages

Net / Test Point	Power State	Expected Voltage	Notes
`PPVRTC_G3H`	Sleep / Standby	3.65–3.85 V	Real-time clock rail. Sourced from coin cell + external power. First rail to measure when board shows zero signs of life. Isolate coin cell if stuck low.
`PP3V42_G3H`	Sleep	3.35–3.55 V	Generated by ISL9239 from PPVBUS. Powers SMC always-on domain. Should appear within 2 seconds of connecting external power.
`PPBUS_G3H`	Sleep	10.5–12.5 V (single outlet power), 20.0–22.0 V (Thunderbolt 3 cable at full spec)	Primary 12 V bus from CD3217. Most failure-prone rail on this board after liquid spill. Check continuity from USB-C connector to ISL9239 input before powering on.
`PP5V_S5`	S5 (Firmware Recovery)	4.95–5.25 V	5 V rail for secondary power domain. Gated by ISL9239. Appears after successful PPBUS and PP3V42 stabilization.
`PP3V3_S5`	S5	3.25–3.35 V	3.3 V secondary domain. Supplies flash memory, SoC reference clocks. Instability here often indicates failing ISL9239 or capacitor plague near NAND.
`PP1V8_S3`	S3 (Standby)	1.75–1.85 V	1.8 V reference for I/O and analog blocks. Low output often precedes display timeout faults. Verify with scope for ripple under >50 mV.
`PP3V3_S3`	S3	3.25–3.35 V	Standby 3.3 V rail. Powers always-on I/O, trackpad, and keyboard controllers. Must be present before any wake event.
`PPVCORE_S0`	S0 (Active)	0.65–0.95 V (dynamic, idles ~0.75 V)	M1 SoC core rail. Tightly regulated by ISL9239 buck converter. Excessive noise (>30 mV pk-pk) or inability to reach 0.65 V under load causes kernel panics. Use 20 MHz BW-limited scope.
`PP1V2_S0`	S0	1.18–1.22 V	SoC I/O rail during active mode. Failure to reach this under load indicates throttling or thermal shutdown imminent.
`PPVDDIO_SYS`	S0	1.78–1.82 V	System I/O voltage reference. Weak output correlates with NAND communication errors and spontaneous resets under I/O load.
`PP5V_USB_SYS`	S0	4.95–5.25 V (per USB port)	USB host power. CD3217 distributes this to both USB-C ports independently. Single port failure indicates CD3217 output FET dead; both indicate source issue.
`PP20V_SYS`	S0 (if attached to 20 V Thunderbolt supply)	19.8–20.5 V	Detected through ISL9239 input sense. Not always present; depends on external power source capability. Check CD3217 gate drive if not present with proper USB-C cable.

Diode Mode Reference

Test Point / Net	Red Probe (−) to GND	Expected Range	Common Bad Reading	Notes
`ISL9239_VBUS_IN`	Input sense to ISL9239 pin 1	0.48–0.52 V	0.35 V or <0.25 V	First-stage PMIC input. Corroded sense resistor divider or failed CD3217 rectifier causes low reading. Compare both USB-C ports; asymmetry points to CD3217 fault.
`CD3217_Q1_DRAIN` (USB-C Port 1)	High-side FET drain	0.50–0.58 V	0.15–0.25 V (stuck on) or >0.85 V (open)	USB-C port 1 power switch. Stuck low = short in downstream buck; stuck high = gate drive dead. Reflow gate drive IC if voltage asymmetric between ports.
`CD3217_Q2_DRAIN` (USB-C Port 2)	High-side FET drain	0.50–0.58 V	0.15–0.25 V (stuck on) or >0.85 V (open)	USB-C port 2 power switch. Most common failure point: asymmetric charging on single port. Often fails after thermal cycling or liquid exposure to USB-C connector area.
`ISL9239_VDD18_OUT`	ISL9239 pin 28 to GND	0.52–0.58 V	<0.30 V or floating	PMIC 1.8 V internal reference. Low or absent = ISL9239 dead or not boot-strapped. Check power sequencing before replacing PMIC.
`PPVCORE_DROOP`	M1 Vcore output point (mid-inductor)	0.60–0.72 V (varies with load state)	0.35 V or >1.0 V	Buck converter output. Excessive ripple (>50 mV) or inability to stay in range under synthetic load indicates failing buck inductor or output capacitors. Measure with 20 MHz BW limit.
`PPBUS_SCHOTTKY` (rectifier cathode toward ISL9239)	Schottky diode cathode	0.28–0.35 V	<0.10 V (open diode) or >0.50 V (shorted or paralleled with FET)	Primary rectifier in buck input stage. Liquid damage near USB-C port frequently shorts this node. Test with multimeter in diode mode from pin to GND; forward bias is normal.
`ISL9239_COMP` (feedback network at comp pin)	ISL9239 pin 14 to GND	0.42–0.52 V	<0.20 V or >0.70 V	Feedback compensation node. Erratic readings indicate failed output capacitors near Vcore buck or broken feedback divider resistor. Scope this for oscillation if board boot-loops.
`NTC_THERM_ADC`	Thermistor pad to GND	0.35–0.55 V (cold state, ~20°C ambient)	>0.75 V or <0.15 V	Thermal sensor divider. High reading = thermistor open or disconnected, causing SMC to throttle or shut down immediately after boot. Reflow thermistor or check for liquid under heat sink.
`SMC_NRESET`	SMC nRESET pin to GND	0.52–0.62 V	<0.10 V (reset asserted) or floating	SMC reset control. Stuck low = SMC firmware fault or watchdog timeout. Stuck floating = watchdog IC failed. Measure with hi-Z DMM; scope will show 0 Hz frequency.
`EEPROM_SDA` (SMC config line)	I2C SDA pull-up to 3.3 V	0.60–0.72 V (idle, pulled high)	<0.20 V (stuck low) or >1.0 V	SMC communication bus. Stuck low = short on data line, often from corrosion. Stuck high = pull-up resistor dead. Disconnect SMC IC at test pad to isolate short source.
`PPBUS_SENSE_DIV`	Vbus divider tap (between resistors toward CD3217)	0.45–0.62 V (depends on actual Vbus)	<0.15 V (open divider) or >1.0 V (shorted to rail)	Voltage sense for input regulation. Liquid or corrosion here prevents ISL9239 from detecting valid input power. Clean divider area thoroughly after spill.
`DRAM_VDD` (unified DRAM direct read)	DRAM supply pad near M1	1.12–1.20 V	<0.80 V or >1.40 V	M1 integrated DRAM rail. Excessive sag under boot load indicates capacitor plague or poor trace routing. RAM overheating faults correlate with low voltage persistence under thermal stress.

Common Faults & Symptoms

No Power After Liquid Spill → Corrosion on PPBUS_G3H distribution rail → Disconnect external power immediately. Measure PPBUS_G3H at ISL9239 input (pin 1) with multimeter in voltage mode; expect 0.48–0.52 V in diode mode with no power applied. If reading is <0.20 V or negative, a short exists between USB-C input and ground. Inspect the area under the keyboard deck and around the trackpad flex connector zone with magnification; corrosion appears as green or white crystalline deposits. Clean with isopropyl alcohol and soft brush, or replace the trackpad interconnect cable if continuity is broken. After cleaning, test PPBUS_G3H again. If voltage is now present, boot test under external power. If still no power, measure PP3V42_G3H; if absent, ISL9239 failed to initialize and likely needs replacement. Do not power on repeatedly; multiple startup attempts can worsen corrosion damage.

USB-C Charging on One Port Only → CD3217 output FET fault (Q1 or Q2) → Test both USB-C ports individually with a multimeter set to voltage mode while connected to a known-good 20 V power supply. Measure PP5V_USB_SYS at the output of each port's downstream buck; one should read 5.0–5.2 V, the other 0 V or floating. Immediately disconnect to prevent backfeed. Switch multimeter to diode mode and measure CD3217_Q1_DRAIN and CD3217_Q2_DRAIN individually to ground. The faulty port will read >0.80 V (open-circuit FET) or <0.15 V (stuck-on FET). If open, the FET gate drive has failed; reflow the ISL9239 PMIC and gate drive capacitors (C2091, C2090) around the CD3217 IC. If stuck on, the FET itself is shorted; remove and replace CD3217 (BGA reball required). This fault often appears after thermal cycling or moisture intrusion near USB-C connectors.

No Image on Internal Display (External Display Works) → eDP flex damage or disconnection → Boot into macOS with external display connected and verify system recognizes the M1 SoC and GPU. Open System Report and confirm GPU memory is accessible. If GPU memory is missing or video output is completely absent, measure PP1V8_S3 at the display power domain; expect 1.75–1.85 V. Low voltage here causes display controller reset. Check the eDP flex ribbon routing under the keyboard assembly; this board routes the eDP flex directly beneath the keyboard deck and is prone to crease damage during reassembly. Disconnect the flex at J2700 (display connector) and inspect the pads with magnification for lifted traces or corrosion. If flex looks intact, test continuity on the eDP signal lines (D0+, D0−, D1+, D1−, etc.) using an ohmmeter on the 2Ω–20Ω scale; expect <1Ω per differential pair. If continuity is broken, replace the eDP flex. If flex tests good but display remains black, the M1 integrated display controller may have failed; this requires PPVCORE_S0 to be stable and PP3V3_S3 to be present. Measure both rails and verify they meet spec before ordering a replacement SoC.

Kernel Panics Under Load (Especially Boot or Heavy GPU Load) → RAM overheating and voltage sag under thermal stress → Install Activity Monitor and run Geekbench 5 Multi-Core while monitoring CPU/GPU package temperature with a tool like Macs Fan Control. Panic should occur between 80–95°C. This indicates throttling and eventual watchdog timeout. First, verify PPVCORE_S0 under load with an oscilloscope (20 MHz BW limit); the core voltage should hold 0.70–0.80 V at max utilization, but if it droops to <0.65 V, the ISL9239 buck converter is failing. Check for burned or cracked capacitors in the Vcore output filter (C series near the M1 SoC). If capacitors look intact, scope the feedback node at ISL9239_COMP for unstable oscillation (>20 kHz ringing), which indicates a failing output capacitor. Second, verify DRAM_VDD does not sag below 1.10 V under boot stress. Low DRAM voltage causes timing errors and parity failures. Third, inspect the thermal interface material (TIM) between M1 and the heatsink; lifted corners or dried-out TIM patches are visible with the heatsink removed. Replace TIM with fresh thermal paste (Bergquist or Fujipoly 17.5 W/m·K recommended) and ensure heatsink mounting pressure is even. If voltage rails are stable and TIM is fresh but panics persist, the M1 SoC core itself has degraded; RMA or SoC replacement is required.

Diagnostic Workflow (No Power)

Assess the damage history. Ask: liquid spill, thermal event, dropped/impact, or spontaneous failure? Liquid spill usually indicates corrosion on PPBUS_G3H or near USB-C. Impact suggests mechanical damage to power connector or inductor cracks. Open the MacBook Air and inspect the keyboard underside and trackpad area with magnification under bright light.
Isolate the power source. Do not connect the MacBook to AC/USB-C power yet. Measure the external power adapter output with a multimeter: should read 20 V ± 1 V (or 10–12 V if only 30 W USB-C provided). If the adapter outputs less than 18 V or shows erratic readings, the power adapter is faulty. Swap with a known-good USB-C PD adapter rated ≥60 W before proceeding.
Measure the coin cell battery. Locate the ML1220 coin cell (usually at the top-left of the board, near the I/O area). Measure its voltage with a multimeter set to DC voltage: expect 3.0–3.2 V. If below 2.8 V, the coin cell is depleted and should be replaced. This is not the primary cause of no-power but ensures PPVRTC_G3H can bootstrap properly.
Inspect the USB-C connector area for corrosion. Use a magnifying glass and bright LED light. Look for green/white crystalline deposits, discoloration, or dark spots around pins. If corrosion is present, clean the connector with isopropyl alcohol and a soft brush before applying power. If connector pins are visibly lifted or broken, the connector must be replaced or the board is not salvageable.
Measure PPBUS_G3H (primary power rail) in diode mode, no external power connected. Set multimeter to diode mode. Place the red probe to an unpopulated test point or the ISL9239 input pad; place the black probe to a known ground pad (chassis, screw hole, or large ground copper near the trackpad). Expected reading: 0.48–0.52 V. If reading is <0.20 V or shows negative polarity, a short circuit exists between USB-C input and ground. Do not apply external power. Inspect the area under the keyboard and trackpad connector for corrosion or moisture. If reading is 0.48–0.52 V (or reads as expected high-impedance forward bias of a diode), proceed to next step.
Connect the external power adapter and measure PPBUS_G3H (voltage mode, no load). Expected reading: 10.5–12.5 V from a single USB-C port (or up to 22 V from a dual-port high-wattage supply). If PPBUS_G3H is 0 V or <1 V, the CD3217 power controller has failed or is shorted. Measure PPVBUS_IN sense divider in diode mode (0.48–0.52 V expected); if below 0.20 V, corrosion exists on the sense resistor divider. If above 0.60 V, ISL9239 PMIC sense is stuck high and the device will not activate. If PPBUS_G3H reads correctly (11 V typical), proceed to step 7.
Measure PP3V42_G3H (always-on secondary rail) within 2–3 seconds of applying external power. Expected: 3.35–3.55 V. This rail powers the SMC and is the first indicator of ISL9239 activation. If PP3V42_G3H is 0 V, the ISL9239 PMIC failed to boot. Check ISL9239_VDD18_OUT (internal 1.8 V reference) in diode mode; expect 0.52–0.58 V. If absent, ISL9239 is dead or the boost converter (which generates the 1.8 V) has failed. ISL9239 requires replacement if this rail cannot be brought online. If PP3V42_G3H is present, proceed to step 8.
Measure PP5V_S5 and PP3V3_S5 within 5 seconds of external power application. Expected: 4.95–5.25 V and 3.25–3.35 V respectively. If both are present and stable, the ISL9239 PMIC has successfully sequenced. Visually inspect the area around the NAND flash IC (likely near bottom-right of board) for burn marks, corrosion, or lifted components. If this rail appears low (below 4.8 V), measure the 5 V buck output inductor with an oscilloscope; expect <100 mV ripple at 500 kHz switching frequency.
Measure PPVCORE_S0 (M1 core rail) and wait for SMC to signal boot attempt. Expected: 0.70–0.85 V during idle/boot. Listen for disk spin-up noise or watch for LED activity (keyboard backlight, sleep indicator). If the board has begun to boot but suddenly shuts down after 2–5 seconds, measure NTC_THERM_ADC in diode mode; expect 0.35–0.55 V when at room temperature. If above 0.70 V, the thermistor is open and SMC has imposed a thermal shutdown. Replace the thermistor (NTC, usually 100 kΩ at 25°C, marked with an "R" designator near the heatpipe).
If the board boots to the Apple logo but then hangs or reboots, check for voltage stability under load. Measure PPVCORE_S0 with an oscilloscope while pressing and holding the power button. The voltage should ramp from idle (~0.75 V) to 0.88 V under GPU load and remain stable. If voltage oscillates wildly (±100 mV or more) or cannot reach 0.88 V, the ISL9239 buck converter output capacitors have failed. Reflow or replace the PMIC. If voltage is stable but the system still reboots within 10 seconds of reaching the login screen, the M1 SoC itself may have failed; attempt an NVRAM reset by holding Command+Option+P+R during boot. If reboot persists, RMA the board.

Related Boards

The 820-02521 shares the M1 power delivery architecture with 820-02969 (MacBook Pro 13" M1 2020, A2338) and 820-02060 (Mac mini M1 2020). Key schematics and BOM entries are identical for the ISL9239 PMIC, CD3217 USB-C controller, and main buck converter topology. The primary differences are:

820-02521 vs. 820-02969: The MacBook Pro variant includes dual fans and a more aggressive thermal design, resulting in tighter PPVCORE ripple specifications (max 25 mV vs. 30 mV on Air) and higher current delivery under sustained load. The NAND and unified DRAM placement is identical, but the Pro has upgraded copper thickness on power planes. Power delivery components (capacitors, inductors) are rated for higher current, so test point readings will be the same, but bandwidth-limited scope measurements may show lower dV/dt on the Pro.
820-02521 vs. 820-02060: The Mac mini board is stationary and thermally robust, but lacks the trackpad and keyboard subsystem. The USB-C power delivery is unchanged. Notably, the Mac mini lacks an internal battery, so PPVRTC_G3H is fed only from the external power supply; the coin cell is a backup only. Troubleshooting USB-C charging faults on the mini is more straightforward because there is no secondary power path from the battery.

For schematics and detailed pin-level test points, cross-reference the M1 power delivery common reference at ISL9239 PMIC and CD3217 USB-C Controller. Corrosion repair procedures are documented in the general liquid damage remediation guide.