Hardware Troubleshooting
Your JD asks for "Hands-on experience with hardware troubleshooting". You cannot fake this, but you can sound credible by being systematic and by knowing where to stop.
The golden rule
The commonest hardware ticket is not broken hardware. It is a cable, a seated connector, a dead power brick, or a device still selected as the default. Test, then swap, then replace.
RMA policy — the professional answer
"Within warranty, I confirm the serial number and warranty status with the vendor, then raise an RMA. Outside warranty, I quote the repair or the replacement cost and get approval before spending anything. Either way I log the asset tag and the failure, because a pattern across machines usually means an environment issue rather than a fault."
POST and no display
| Component | POST beep pattern (varies by vendor) |
|---|---|
| No beeps, no power | PSU or wall outlet |
| 1 long beep | Memory |
| 3 short beeps | Memory (varies) |
| 2 short beeps | Video card |
| Continuous short beeps | RAM or video |
| Single short beep | Normal POST |
Beep codes are not universal — AMI, Phoenix and Dell codes differ. Say "I check beep codes against the specific board documentation" rather than asserting a meaning. It is a small thing that signals real experience.
No display but the machine runs
# Is it running? Fan spinning, lights on.
Get-Process | Select-Object -First 5 Name # via remote or a borrow
# Linux
lspci | grep -i vga
dmesg | grep -i "drm\|vga\|nouveau\|nvidia"
| Cause | Check |
|---|---|
| Cable | Try a known-good one, try the other port |
| Wrong input | Cycle the monitor inputs, hold its button |
| Wrong resolution | Windows may be sending a mode the panel cannot show |
| GPU seated | Reseat, check fan spin |
| Laptop lid switch | External display often proves the machine is fine |
Storage
Get-PhysicalDisk | Select-Object FriendlyName, MediaType, HealthStatus
Get-Volume | Select-Object DriveLetter, DriveType, HealthStatus
Get-WinEvent -FilterHashtable @{LogName='System'; ProviderName='disk'; Id=7,51}
chkdsk C: /scan
| Signal | Meaning |
|---|---|
| Event ID 7 | Bad block found — replace the drive |
| Event ID 51 | I/O error on the paging file — usually a dying drive |
| Clicking from a mechanical drive | Head failure. Back up now. |
| SMART warning | Predictive failure. Replace before it fails. |
| SSD slower than its spec | Usually thermal or a worn controller |
Memory
mdsched.exe # Windows Memory Diagnostic
- Test with one stick at a time in each slot.
- Faults in one slot point at the slot; faults following a stick point at the stick.
Power
| Symptom | Check |
|---|---|
| No power at all | Wall outlet, PSU brick, PSU switch |
| Flickers under load | Failing PSU |
| Shuts off at random | Thermal, PSU, or a failing drive |
| Won't power on but fans spin | Memory, then motherboard |
Never open a PSU. Even when unplugged, capacitors retain charge and a PSU can destroy a motherboard instantly. Say this if asked.
Thermal
sensors # Linux
# No built-in PowerShell cmdlet — use a vendor tool or LibreHardwareMonitor
| Cause | Fix |
|---|---|
| Dust in heatsink | Clean it |
| Fan seized | Replace |
| Paste dried out | Repaste |
| Case airflow blocked | Reseat drives, close the side panel |
| Laptop on soft surface | Elevate it |
Laptops specifically
| Fault | Cause |
|---|---|
| Won't charge | Brick, DC jack, or the charge IC |
| Battery drains fast | Battery wear, or a high-drain peripheral |
| No display on lid open | Lid switch or hinge flex cable |
| Overheats under load | Dust, dried paste, or blocked vents |
| No network | Wi-Fi key toggle, driver, or a hardware switch |
Swapping a component — the professional framing
- Record the asset tag and back up the user if needed.
- Power down, unplug peripherals, discharge.
- Photograph the cabling before removing anything.
- Anti-static precautions.
- Swap, reassemble, test.
- Verify the fix and update the asset record.
"Anything I can swap in under fifteen minutes without opening a warranty sticker — RAM, drives, batteries, peripherals — I do myself. Anything requiring the manufacturer, I RMA rather than risk the warranty."
Interview scenario
'The user's machine has no display. What do you do?'
"First I'd ask whether it's making noise and whether the lights are on, which distinguishes a power fault from a video fault. Then I'd check the monitor itself — power light, correct input source, and a known-good cable, since those are free to rule out. If the machine is running but showing nothing, I'd try an external display, which separates a screen problem from a graphics problem. If it's a desktop with no signs of life, I'd check the outlet and power supply brick before going near components, because the PSU is the most common single point of failure and the cheapest thing to replace. Anything past that — reseating the GPU, testing RAM — is a hardware hands-on step."