4.1.1. Troubleshooting Motherboards, RAM, CPUs, and Power
💡 First Principle: To diagnose boot failures, you must interpret audible (beep codes) and visual (POST errors, BSOD) cues to form a primary theory, then test that theory by reseating components before replacing them.
Scenario: A user reports their 3-year-old PC randomly shuts down, but only when they are exporting large video files. They mention it doesn't happen during normal web browsing. The case feels warm to the touch. You check Event Viewer and see "Kernel-Power" critical errors but no other specific cause.
Technician's Diagnostic Path:
- Initial Theory: The symptoms (shutdown under heavy load, heat) point strongly to either an overheating issue or an insufficient/failing Power Supply Unit (PSU).
- Test for Overheating:
- Plan: Install a temperature monitoring utility (e.g., HWMonitor).
- Action: Ask the user to start a video export and monitor the CPU and GPU temperatures.
- Expected Result: If temperatures climb above 90-95°C before shutdown, overheating is the confirmed cause.
Visual: Thermal compound on CPU - Use pea-sized amount in center. Spreads when heatsink is mounted. Dried or missing paste is a common cause of overheating.
- Test for PSU Failure:
- Plan: If temperatures are normal, the PSU may not be providing stable power under load.
- Action: Use a PSU tester to check basic voltage rails. For a more definitive test, swap the PSU with a known-good unit of appropriate wattage.
- Expected Result: If the system is stable with the new PSU, the original PSU is faulty.
⚠️ Common Pitfall: Forgetting to check for swollen or leaking capacitors on the motherboard. This is a clear sign of a failing motherboard that needs to be replaced.
Visual: Failed motherboard capacitors - Note the bulging tops (should be flat). This causes system instability and random shutdowns.
Symptom Quick Reference (Power, POST, Boot, and Heat):
| Symptom | Most likely cause | First action |
|---|---|---|
| No fans, no lights, nothing | Dead outlet or cable, PSU rear switch off, failed PSU, or front-panel power switch/header | (1) Verify the wall outlet and the PSU rear switch; (2) inspect the 24-pin and 4/8-pin connectors at the motherboard; (3) test the PSU with a PSU tester or the paperclip test; (4) if it passes, suspect the front-panel power switch, then the motherboard |
| Fans spin and LEDs light, but no video and no beeps | RAM not seated or incompatible (most common), then CPU or motherboard | (1) Confirm the monitor is on and cabled to the correct output (integrated vs. discrete GPU); (2) power down, ground yourself, and reseat the RAM; (3) test with one known-good stick in the first slot; (4) reseat the CPU, check for bent pins, and clear CMOS |
| CPU fan spins for a second, stops, and repeats (new build) | 4/8-pin CPU power connector not plugged into the motherboard | Connect the CPU power connector |
| Beep codes at power-on | A hardware fault found during POST. Patterns vary by BIOS vendor (read the motherboard manual); e.g., three short beeps usually points to memory and one long plus two short to the video card | Reseat or replace the part the code names, usually RAM or the video card |
| Wrong date/time after power-off, BIOS settings reset | Dead CMOS battery (see 2.2.1) | Replace the CR2032 cell |
| Random shutdowns or reboots, fans at full speed, throttling and slowdowns under load | Overheating (dust in the heat sink or fans, failed fan, dried thermal paste on an older system) or a failing PSU | Confirm with temperature-monitoring software under load; power off and blow dust out with compressed air, replace failed fans, reapply thermal paste; if temperatures are normal, test the PSU |
| Blue screen (BSOD) | Failing RAM, a failing drive, or a bad driver; a MEMORY_MANAGEMENT stop code points to RAM | Note the STOP code and check Event Viewer system logs; run a memory test (MemTest86 or Windows Memory Diagnostic), first for memory-related codes; update or roll back recent drivers |
| Slow performance | Insufficient RAM, malware, or a failing hard drive (high network latency affects only network tasks) | Check Task Manager, scan for malware, then test the drive and memory |
| Restarts over and over before the Windows logo appears | Firmware/boot settings or a hardware fault (RAM, PSU, heat) | Least destructive first: enter BIOS/UEFI to check boot settings and temperatures, reseat components; reinstalling Windows or replacing parts comes last |
| Reboot loop after Windows starts loading, before the desktop | Corrupted driver or OS update | Boot to Safe Mode or the recovery environment and roll back the update or driver |
| "Operating system not found", BOOTMGR missing, invalid BCD | OS/boot-configuration problem (wrong boot order, damaged boot files) or a failed drive; see 4.1.2 | Check boot order and drive detection, then repair boot files |
| INACCESSIBLE_BOOT_DEVICE stop error | Storage controller or boot configuration (BCD) problem | Check the controller mode and BCD |
| Tingling or shocks from the metal case | Ground fault: improper grounding of the outlet, cord, or PSU | Unplug the system and have the outlet and grounding checked |
Rule of thumb for classifying a startup failure: problems that occur before the operating system starts loading (beep codes, no video while fans spin, a POST memory error) are hardware/POST failures; messages that name boot files or Windows internals (BOOTMGR missing, invalid BCD, corrupt registry) are OS/boot-configuration failures.
Tools to know:
- Multimeter: measures the actual voltages on the PSU connectors (3.3 V, 5 V, 12 V). A PSU tester gives a quick pass/fail on the rails. A POST card displays POST codes when there is no video. A loopback plug tests a network or serial port, not a PSU.
- Hardware diagnostics outside the OS: the UEFI/BIOS built-in diagnostics (or a bootable memory tester such as MemTest86) test hardware without booting Windows.
- ESD protection: wear an ESD wrist strap connected to a ground point (and use an antistatic mat). Raising humidity reduces static but does not replace a strap.
Reflection Question: Justify why you would test for overheating before replacing the PSU. Design a 3-step plan to resolve the issue if you confirm the cause is CPU overheating.
