Week 4: Power, Cooling, Expansion, and Displays

A computer can have perfectly compatible parts and still fail because the system cannot deliver power, remove heat, or move a video signal from the graphics source to the display.

This week is about those paths.

Power is not one yes-or-no state. A fan can spin while the CPU power connector is missing. A graphics card can sit correctly in a PCIe slot while its supplemental power is disconnected. A system can boot normally at idle and shut down only when heat or electrical load rises.

The display path works the same way. A black screen can begin at the GPU, the output port, the cable, the selected monitor input, the panel electronics, or even earlier when the machine never completes POST.

So we are going to stop treating power, cooling, expansion, and display as isolated topics and trace them as dependent systems.

The system question

The question for Week 4 is:

When a machine powers, overheats, shuts down, or loses video, which path failed first and what evidence proves it?

Use three models:

POWER
wall / UPS
   ↓
PSU
   ↓
motherboard / CPU / GPU / drives

HEAT
CPU / GPU / power circuitry
   ↓
thermal interface
   ↓
heatsink / radiator
   ↓
airflow
   ↓
room

DISPLAY
GPU / graphics source
   ↓
output interface
   ↓
cable or adapter
   ↓
selected display input
   ↓
panel / projector

A symptom only becomes useful when you can place it somewhere on one of those paths.

How the chapter moves

Lesson 1 traces AC input, PSU conversion, DC rails, connectors, wattage, efficiency, modular cabling, and redundant-power concepts as one distribution system.

Lesson 2 follows heat from the component into the room. Fans, heatsinks, thermal paste or pads, liquid cooling, airflow, throttling, and shutdown behavior only make sense when you see the whole thermal path.

Lesson 3 connects expansion cards to display technology. You will compare graphics, sound, capture, and network cards while separating IPS, TN, VA, OLED, Mini-LED, touchscreen/digitizer, refresh rate, resolution, pixel density, color gamut, and older inverter-based display behavior.

Lesson 4 is the planning lab. You will prove that a media or esports workstation has enough power, cooling, expansion capacity, and display-path capability before it is assembled.

Lesson 5 turns those models into fault isolation across dead systems, load-related shutdowns, GPU failures, monitor faults, and projector symptoms.

What you should be able to defend

By the end of the week, you should be able to explain:

• the difference between AC input and the 3.3 V, 5 V, and 12 V DC rails used inside a PC;
• why a 20+4-pin motherboard connection, CPU power, GPU power, and storage power are separate failure boundaries;
• what wattage, energy efficiency, modular cabling, and redundant PSUs actually mean;
• how heat moves through thermal interface material, heatsinks, fans, liquid loops, radiators, and case airflow;
• why sound, video, capture, and network expansion cards create different slot, lane, power, and clearance requirements;
• how panel technology and display attributes change the workload fit; and
• how to distinguish power, thermal, video, monitor, and projector failures from observable evidence.

The safety line

Never open a PSU enclosure. Remove external power before internal service unless an approved procedure specifically requires powered observation.

Stop normal testing for burning smell, smoke, swollen components, melted connectors, exposed wiring, liquid damage, or other unsafe electrical evidence.

You do not need a second failure to prove the first one was dangerous.