Lesson 5: Hardware, Video, and Projector Fault Isolation

Week 4 ends with systems that all look "broken."

They are not broken in the same way.

The useful skill is being able to look at a symptom, place it on the correct path, and resist the urge to name a part before the evidence earns that conclusion.

Case A: completely dead

Evidence:

no LEDs
no fans
no display
outlet: known-good

The symptom is system-wide, so begin with the earliest shared dependency: power delivery.

Useful checks include:

• external power cable;
• PSU switch and input state when applicable;
• main motherboard power;
• front-panel power-switch path;
• approved PSU or board diagnostic procedure.

Do not open the PSU.

Do not begin by replacing the motherboard because "dead computer" sounds dramatic.

Case B: starts, then shuts down under load

Evidence:

boot: normal
idle: stable
heavy workload: temperature climbs
CPU fan behavior: abnormal
system: shuts down

Thermal failure becomes a strong hypothesis.

But keep one competing explanation, such as PSU instability under load.

The safest next check should gather evidence without creating a new failure.

Case C: burning smell or unusual electrical noise

Evidence:

user reports burning smell or buzzing/electrical noise
system was under load
source not safely identified

The correct next action is not "turn it on and see if it happens again."

Stop power and escalate.

Safety changes the troubleshooting workflow because the cost of another test is too high.

Case D: new GPU, no display

Evidence:

system powers on
GPU installed
monitor and cable work elsewhere
GPU supplemental power: not verified

Stay near the changed boundary.

Check GPU power and installation before reinstalling drivers or the operating system.

The OS may not even be part of the failure yet.

Case E: monitor says No Signal

Evidence:

PC appears to complete startup
monitor: powered
cable: DisplayPort
monitor input selected: HDMI

The strongest current explanation is simple: wrong input source.

This is exactly why technicians check the boring part first.

A working GPU cannot deliver an image to an input the monitor is not listening to.

Case F: projector image is fuzzy, distorted, or sized incorrectly

Possible boundaries include:

• projector focus;
• projector distance or zoom;
• source resolution;
• aspect ratio;
• digital scaling;
• keystone or geometry correction;
• signal quality.

Notice what is missing from the top of that list: "replace the GPU."

The symptom points more directly toward the optical and display-configuration path.

Case G: projector is dim

Possible causes depend on the projector design and include:

• brightness or eco settings;
• aging lamp or light source;
• blocked optics or filters;
• thermal conditions.

On a lamp-based projector, a burnt-out bulb can leave the unit apparently powered while producing no usable image.

That is a light-source failure, not automatically a source-computer failure.

Case H: projector shuts down intermittently

If internal temperatures rise and airflow is blocked, thermal protection becomes a strong theory.

Projectors produce heat too.

A repeatable shutdown after temperature rises is different evidence from an immediate loss of input signal.

Case I: dead pixel or burn-in?

These are different symptom classes.

A dead pixel is an individual pixel that fails to operate normally.

Burn-in or persistent image retention is a lasting image pattern associated with some display technologies and usage conditions.

One bad point and a persistent ghost image do not imply the same mechanism or service outcome.

Use the visible pattern as evidence.

Case J: flashing screen

Possible boundaries include:

• loose or damaged cable;
• unsupported resolution or refresh combination;
• GPU or driver state;
• monitor input capability;
• panel electronics.

A flashing image is not automatically a bad monitor.

Test the path.

A known-good cable is often a cheap, low-risk comparison before replacing expensive hardware.

Case K: incorrect colors

Inspect:

• cable or interface path;
• GPU or driver state;
• display color settings or profile;
• panel hardware evidence.

A wrong-color image is different from a missing image.

That distinction changes the path you inspect first.

Case L: video works, audio does not

HDMI and DisplayPort can carry audio as well as video.

If the image is present but audio is missing, inspect:

• selected audio output device;
• monitor or interface audio capability;
• mute and volume state;
• driver or operating-system configuration;
• cable or adapter limitations.

The video signal working does not automatically prove the audio path is configured correctly.

One cable can carry multiple logical functions.

Case M: sluggish behavior and random shutdown under load

Evidence:

idle: stable
heavy workload: performance falls
fans: unusually loud
temperature: high
shutdown: intermittent

A thermal explanation becomes stronger.

Now change the evidence:

idle: stable
heavy workload: sudden shutdown
temperature: normal
PSU capacity: marginal for new GPU

The power path now deserves more attention.

Same user complaint. Different evidence.

Build three service records

Choose three cases from different failure families.

For each, preserve:

• exact symptom;
• known-good evidence;
• last proven-good boundary;
• strongest current theory;
• one alternative;
• controlled next check;
• safety boundary;
• expected result if the theory is correct;
• verification after correction;
• one thing the test still does not prove.

That last line prevents a good test from becoming an oversized conclusion.

Defend one complete chain

Use one scenario to explain the system from beginning to end:

safe service
   ↓
identify hardware or display technology
   ↓
verify compatibility
   ↓
trace power / heat / video path
   ↓
locate failed boundary
   ↓
controlled correction
   ↓
verify required function
   ↓
document limits

A strong defense sounds like a technician explaining evidence, not a student reciting a checklist.

What the first four weeks should have changed

By now, these statements should feel obviously incomplete:

It fits, so it is compatible.

The fans spin, so the PSU is good.

The drive is missing, so it is dead.

The screen is black, so the monitor is broken.

Each statement skips the mechanism between the symptom and the conclusion.

Your job in this course is to stop skipping that middle.