Home/ Symptoms/ random-shutdown
Symptom Hub · All Boards

Random Shutdown Diagnosis

2 boards 7 diagnostic checks Updated March 15, 2026

Overview

Random shutdowns fall into three distinct fault categories that require different diagnostic approaches: (1) thermal protection — the board shuts down when a component exceeds its thermal limit; (2) power delivery instability — voltage droop or ripple on a critical rail causes a brown-out; (3) PMU/SMC fault — the management controller incorrectly triggers a shutdown. The key diagnostic tool is a USB-C power meter or current clamp on the PCIe slot to monitor power draw at the moment of shutdown, combined with temperature logging.

Diagnostic Methodology

Follow these steps in sequence. Each step eliminates an entire fault zone — do not skip ahead.

1. Log temperature before shutdown
Apple: iStatMenus or powermetrics. GPU: GPU-Z. The CPU/GPU temperature at shutdown time identifies thermal faults instantly.
2. Thermal paste inspection
Repaste CPU and GPU if system age > 3 years. Dried paste causes 30–50°C temperature increase. MacBook hinge area and T2 chip also require paste.
3. Check VRM ripple under load
Probe PPVCORE or GPU VCORE with oscilloscope under stress test. Ripple > 50 mV peak-to-peak on a 1V rail indicates failing output capacitors.
4. Measure capacitor ESR
Output capacitors of CPU/GPU VRM fail with increased ESR over time. Use ESR meter on bulk caps in-circuit. Values >0.1Ω on VCore caps = replacement required.
5. Inspect SMC/T2 thermal table
Apple boards: SMC thermal management can trigger false shutdowns if thermal sensor reads incorrectly. Shorted or open thermistors (e.g., near CPU) cause premature shutdown.
6. Check power delivery headroom
Apple: measure PPBUS_G3H under load. Voltage should not sag below 11.0V under load. Sag = failing battery or charger IC.

Per-Board Fault Trees

Board-specific checks ordered by failure likelihood. Most common root cause listed first.

MacBook Air A2337 (M1) — Random Shutdown Start: PPBUS_G3H under load
⊞ Repair guide ◈ Schematic hub
  • Battery voltage sag
    M1 MacBook draws up to 30W under full load. Aging battery sags below 10V under load → M1 triggers shutdown. Replace battery if capacity < 80% or voltage dips.
  • M1 VCore ripple
    M1 integrated VRM can develop output ripple if output caps fail. Probe VDD_CPU_G0D under load. Excessive ripple = capacitor replacement on M1 VCore outputs.
  • Thermal paste degradation
    M1 SoC uses a graphene thermal pad not standard paste. If replaced with incorrect paste, thermal resistance increases and throttling/shutdown occurs at 90°C+.
GTX 1060 — Random Shutdown / Crash Start: PCIe 12V under load
⊞ Repair guide ◈ Schematic hub
  • PSU 12V sag under load
    PCIe slot + 6-pin combined deliver up to 150W. Underpowered PSU causes 12V sag → GPU throttles or crashes. Use PSU with ≥500W and ≥18A on 12V rail.
  • GPU VRM output ripple
    NCP81174 output caps near GPU die. Under full load, measure GPU VCORE oscilloscope — ripple > 60 mV = failing decoupling caps on VRM output.
  • Thermal paste on GP106
    GP106 die is large-area (200mm²). Dried thermal paste causes GPU-Z Temp > 85°C under load → driver crash or BSOD. Repaste with Kryonaut or similar.
  • GDDR5 overheating
    GDDR5 modules have no thermal pads on some GTX 1060 SKUs. Memory overheating causes VRAM artifacts. Add 1.0mm thermal pads between GDDR5 and backplate.

Frequently Asked Questions

How do I tell if it's a thermal vs. power fault?
Thermal: CPU/GPU temperature at shutdown is ≥ throttle threshold (typically 100°C for Intel, 90°C for Apple Silicon, 88°C for GPU die). Power: sudden shutdown with no thermal event, often during load spikes (spinning up a game, video export). Power faults can be confirmed by measuring rail voltage with a logger at shutdown moment.
Can failing VRM capacitors cause random shutdowns?
Yes, this is a common failure mode on boards aged 5+ years. The polymer capacitors on CPU/GPU VRM output slowly increase in ESR. Under load, this causes voltage droop and ripple that exceeds the processor's voltage tolerance window, triggering an internal protection shutdown. Replace all VCore rail output caps together — don't replace one at a time.