A good technician often names the faulty part before opening a single screw — not by magic, but by reading symptoms the way a doctor reads a fever chart. This guide gives you that map.
Every computer problem announces itself before it becomes a breakdown. A click from the hard disk, a smell from the power supply, lines on the screen, a shutdown that always happens at the same time of day — each symptom is the machine telling you where it hurts. Beginners replace parts by guessing and pay for it twice. Technicians do something different: they read the symptom, narrow the suspects to two or three, and run one confirming test. This core guide is the complete symptom-to-cause map — sounds, smells, heat, screen faults, shutdown patterns, crash patterns, network scope, printer signs and boot messages. Learn the map here, then open the case guides below for each symptom family in depth.
👆 Six symptoms, six different first suspects. The symptom narrows the list — a test proves the cause.
Diagnosis starts with four questions, and none of them needs a screwdriver:
Half of all wrong repairs happen because someone skipped these four questions and replaced the loudest suspect. The symptom tells you where to look; it never excuses you from testing what you find there.
A healthy PC makes two sounds: fans and, on old machines, a gentle hard-disk murmur. Everything else is a message. Repeated clicking or grinding from a mechanical hard disk is the famous click of death — the drive is failing mechanically, so back up first and test later. Beeps at power-on are the motherboard speaking in BIOS code: the pattern points at RAM, graphics or CPU, but the code table belongs to your board maker, so read the manual before believing any chart on the internet. A roaring fan means heat or a dying fan bearing — and Task Manager will tell you if it is just a runaway program working the CPU at 100 percent. The full sound dictionary is in the sounds guide in this set.
A burning or sharp chemical smell means stop: switch off, unplug, and do not keep testing a power supply that is cooking itself. Heat has quieter signs too — a PC that slows down after twenty minutes, fans that scream under light work, a laptop base too hot to touch. Overheating makes the processor throttle itself, so heat often disguises itself as slowness or random shutdowns. Dust, dried thermal paste and blocked vents are the usual causes, and they are cheap to fix — if you catch them before the heat damages something expensive.
Lines, colours and flicker look like one problem and are actually three. The technician split is one screenshot: press Print Screen and view the captured image. If the fault appears in the screenshot, the graphics card produced it — GPU or its driver. If the screenshot is clean, the GPU is innocent and the fault lives in the cable, the monitor panel or its connector. Artefacts that change when you wiggle the cable blame the cable; a line that stays in the same place on every screen blames the panel. One free test, and you never replace a graphics card for a cable fault again.
Timing is evidence. A PC that is slow only at fixed times is confessing to a schedule: antivirus scans, backups, sync tools and Windows Update all run on clocks you can read in Task Scheduler. A PC that shuts down under load points at heat or a weak power supply; one that dies at random, even idle, points at power or RAM. Crashes follow the same grammar: one app crashing is usually that app, its data or one driver; everything crashing with different blue-screen codes is RAM or power until a test proves otherwise. Write the pattern down — the shutdown and schedule guides in this set show you how.
The fastest question in networking is: who else has it? Internet down on one PC means the fault is in that PC, its cable or its port. Down on every PC means router, switch or the internet line itself — and no amount of driver reinstalling on one PC will fix it. The same scope logic works everywhere: one user cannot print but others can — the problem is that user or that PC; nobody can print — the printer or the network. Scope first, tools second, and you stop repairing healthy machines.
Some of the most honest symptoms appear before Windows even starts. S.M.A.R.T. Failure Predicted means the disk itself is reporting its own death — believe it and back up today. CPU Fan Error means the fan is not spinning or not detected. A clock that resets and a CMOS checksum message mean the small CR2032 battery on the board is finished. No bootable device splits into a dead drive, a wrong boot order, or a boot mode mismatch. These messages are not riddles; they are the machine naming the part. The boot message guide translates each one.
Random faults defeat memory, so professionals write things down. For every incident, note five things:
Three diary entries often reveal the pattern that three weeks of guessing could not: the shutdown is always at lunch (heat plus a full office), or always ten minutes after the backup starts. A symptom narrows the suspects — a diary plus one test proves the cause.
All fifteen symptom patterns from the video, shown live on real faulty machines — sounds, screens, shutdowns and scopes explained step-by-step.
▶️ Watch Full VideoOften, yes — far enough to know which part to test first. Sounds, smells, timing, scope and boot messages narrow most faults to one or two suspects. Opening the case is for confirming with a test, not for the first guess.
The screenshot test for screen faults, and Event Viewer with Reliability Monitor (perfmon /rel) inside Windows for crashes. Both are built in, both separate whole families of causes in minutes.
At a burning smell, a clicking hard disk with data you cannot lose, or any fault inside a power supply. Those three can turn a cheap repair into lost data or a safety risk.