Why does your gigabit NIC only run at 100Mbps? Why does your new PC with 650W PSU keep restarting? Why is WiFi signal full but speed slow? This is the finale of the Computer Fundamentals Deep Dive Series. We'll comprehensively cover NIC types and configuration, PSU power calculation and selection, cooling system optimization, complete PC building process, and 30+ common fault diagnosis solutions. This is a practice-oriented hardcore guide covering everything from hardware connections to system optimization, making you a true PC maintenance expert.
Series Navigation
π Computer Fundamentals Deep Dive Series (5 Parts): 1. CPU & Computing Core 2. Memory & High-Speed Cache 3. Storage Systems 4. Motherboard, Graphics & Expansion 5. β Network, Power & Troubleshooting (NICs, PSU, cooling, diagnostics) β Series Finale
Part 1: Network Cards
Wired NICs
| Standard | Speed | Actual | Application |
|---|---|---|---|
| Fast Ethernet | 100 Mbps | 12 MB/s | Obsolete |
| Gigabit | 1 Gbps | 125 MB/s | Mainstream β |
| 2.5G | 2.5 Gbps | 312 MB/s | High-end |
| 10GbE | 10 Gbps | 1250 MB/s | Enterprise |
Wireless Standards
| Standard | Frequency | Speed | Year |
|---|---|---|---|
| WiFi 4 (n) | 2.4/5 GHz | 600 Mbps | 2009 |
| WiFi 5 (ac) | 5 GHz | 3.5 Gbps | 2013 |
| WiFi 6 (ax) | 2.4/5 GHz | 9.6 Gbps | 2019 |
| WiFi 7 (be) | 2.4/5/6 GHz | 46 Gbps | 2024 |
2024 Recommendation: WiFi 6 or 6E
Network Addressing & Communication Principles
Special IP Addresses Explained
π Intuitive Understanding: Why Special Addresses?
In the networking world, some IP addresses have special meanings, like concepts of "local" and "everywhere" in real life. Understanding these special addresses is key to mastering network communication.
Loopback Address (127.0.0.1)
What is it: 127.0.0.1 is a special IP address called the "loopback address" or "localhost".
Intuitive Analogy: Imagine calling yourself on the phone at home β you dial the number, but the call never leaves your house and doesn't go through outside phone lines. 127.0.0.1 is like this "call to yourself" address.
Why do we need it: - Local testing: Developers can test network applications locally without a real network connection - Quick verification: Check if the network protocol stack is working properly - Security isolation: Test traffic never leaves the machine, won't affect external networks
Practical Applications: 1
2
3
4
5# Test local web server
curl http://127.0.0.1:8080
# Test database connection
mysql -h 127.0.0.1 -u root -p
Technical Details: - Packets don't go through physical NICs, but complete loopback within the OS network protocol stack - The entire 127.0.0.0/8 subnet (127.0.0.1 to 127.255.255.255) are loopback addresses - Extremely low latency (microseconds), unaffected by network conditions
Special Binding Address (0.0.0.0)
What is it: 0.0.0.0 has different meanings in different scenarios, most commonly as a server binding address.
Intuitive Analogy: Imagine a hotel reception saying "I serve guests from any entrance"β whether you come through the main entrance, side entrance, or back entrance, the reception will serve you. 0.0.0.0 is like "accepting connections from any network interface".
Why do we need it: - Multi-NIC support: Servers typically have multiple NICs (wired, wireless, VPN, etc.) - Flexible deployment: No need to configure each NIC separately - Simplified management: Bind once, listen on all NICs
Comparison: | Bind Address | Access Method | Use
Case | |--------------|---------------|----------| |
127.0.0.1 | Only local access | Local testing | |
192.168.1.100 | Only through this NIC | Single NIC service
| | 0.0.0.0 | Through any NIC | Production
β
|
Practical Applications: 1
2
3
4
5# Flask bind to all NICs
app.run(host='0.0.0.0', port=5000)
# Nginx listen on all NICs
listen 0.0.0.0:80;
Common Misconceptions: - β Misunderstanding: 0.0.0.0 is an accessible address - β Correct: 0.0.0.0 can only be used as a binding address, not as an access target
Ports: Door Numbers for Services
π Intuitive Understanding: Why Ports?
Imagine a large office building β the building's address is the IP, but the building has hundreds of companies, each with their own door number β this is what ports do. A single IP address can run multiple services, and ports are the numbers that distinguish these services.
Common Ports and Their Services
| Port | Service | Protocol | Purpose |
|---|---|---|---|
| 22 | SSH | TCP | Remote login (encrypted) |
| 80 | HTTP | TCP | Website access |
| 443 | HTTPS | TCP | Encrypted website access |
| 3306 | MySQL | TCP | Database |
| 6379 | Redis | TCP | Cache |
| 27017 | MongoDB | TCP | NoSQL database |
| 53 | DNS | UDP | Domain resolution |
Troubleshooting Example:
Problem: Ping works but website doesn't
Reason: - ping uses ICMP protocol,
doesn't depend on ports - Website access requires TCP ports 80/443,
service may not be running
Troubleshooting Steps: 1
2
3
4
5
6
7
8# 1. Check if port is listening
netstat -tnlp | grep 80
# 2. Test port connectivity
telnet example.com 80
# 3. Check service status
systemctl status nginx
NAT and Network Communication
π Intuitive Understanding: Why NAT?
Imagine a large company where internal employees have their own employee IDs (private IPs), but all external communications use the company's unified phone number (public IP). The receptionist (router) records "which employee called whom", and when calls return, they're transferred back to the corresponding employee. This is how NAT (Network Address Translation) works.
How NAT Works
Background Problem: - IPv4 addresses are only about 4.2 billion, far fewer than global devices - Private IP addresses (192.168.x.x, 10.x.x.x) cannot be used on the public internet - Multiple devices need to share one public IP
NAT Translation Process:
1 | LAN device accessing internet: |
NAT Mapping Table Example: | Internal Address | Internal Port | Public Port | External Address | |------------------|---------------|-------------|------------------| | 192.168.1.100 | 5000 | 30000 | 8.8.8.8:53 | | 192.168.1.101 | 8080 | 30001 | 1.1.1.1:443 |
Domain Name Resolution (DNS)
π Intuitive Understanding: Domains are the "Address Book" of URLs
You remember your friend's name but not necessarily their phone
number β the address book translates "names" into "numbers". The Domain
Name System (DNS) does exactly this: translates
www.example.com into 1.2.3.4.
DNS Resolution Process
1 | User input: www.example.com |
Why does the same domain have different IPs in different locations:
Reason: CDN (Content Delivery Network)
Analogy: McDonald's has many branches nationwide. When you search "McDonald's" in Beijing vs. Shanghai, the navigation will point to different branches (proximity principle).
Technical Implementation: - DNS server determines geographic location based on user's IP address - Returns IP of the nearest CDN node to the user - User accesses the nearest server, accelerating content loading
Network Connection Modes Explained
Why Understand Network Modes?
When using virtual machines (VMware, VirtualBox) or containers (Docker), you often encounter issues like "VM can't access internet" or "host can't ping VM". Understanding network modes is key to solving these problems.
Bridged Mode
π Intuitive Understanding
Imagine you have a router at home connected to your computer, phone, and TV. Now you buy a new computer (VM) and plug a network cable directly into the router β this is bridged mode. The VM is like "another real computer" in your home, with equal status to the host.
How It Works
1 | Physical network topology: |
Key Features: - VM gets an independent IP in the same subnet as the host - VM can be directly accessed by other devices in the LAN - VM accesses internet through router's NAT
Data Flow: 1
2VM(192.168.1.101) ββ VMware Virtual Switch ββ Host NIC
ββ Router(192.168.1.1) ββ Internet
Use Cases: - β VM needs to be accessed by other LAN devices (e.g., test server) - β VM needs completely independent network identity from host - β Insufficient LAN IP addresses (each VM occupies one IP)
Common Issues: - Problem: VM cannot obtain IP - Cause: Router DHCP address pool is full - Solution: Manually configure static IP or expand DHCP range
NAT Mode (Network Address Translation)
π Intuitive Understanding
Imagine the host is a "small router" that creates a "virtual LAN" for the VM. The VM has its own IP in this virtual network (e.g., 192.168.182.128), but this IP is only valid within the virtual network. When the VM accesses the internet, it needs to go through the host's "double NAT translation".
How It Works
1 | Network hierarchy: |
Detailed Data Flow:
VM accessing internet: 1
2
3
4
51. VM(192.168.182.128:5000) sends request to VMware gateway(192.168.182.2)
2. Host NAT: 192.168.182.128 β 192.168.1.100 (changed to host IP)
3. Router NAT: 192.168.1.100 β Public IP (changed to public IP)
4. Request reaches internet server
5. Return reverse translation: Public IP β 192.168.1.100 β 192.168.182.128

Key Features: - VM uses separate subnet IP (typically 192.168.x.x) - External network sees host IP, can't see VM - VMs can access each other (within same virtual network) - External cannot proactively access VM (unless port forwarding is configured)
Comparison with Bridged Mode: | Feature | Bridged Mode | NAT Mode | |---------|--------------|----------| | VM IP | Same subnet as host | Separate virtual subnet | | Occupies LAN IP | Yes | No | | External can access VM | Yes | No (needs port forwarding) | | Security | Low | High | | Use Case | Test server | Daily development |
Use Cases: - β Daily development and testing (most common) β - β Insufficient LAN IP addresses - β Need to isolate VM network - β Need external direct access to VM
Port Forwarding Configuration (allow external access
to VM): 1
2VMware β Edit β Virtual Network Editor β NAT Settings β Port Forwarding
Host port 8080 β VM 192.168.182.128:80
Host-Only Mode
π Intuitive Understanding
Imagine a "dedicated network cable" between the VM and host, with only these two connected to this cable, completely isolated from the outside world β the VM cannot access the internet, external networks cannot access the VM, it can only communicate with the host.
How It Works
1 | Isolated network: |
Key Features: - VM can only communicate with host - Cannot access internet - Cannot access other LAN devices - Other LAN devices also cannot access VM
Use Cases: - β Test environment needs complete isolation - β Sensitive data processing (no internet desired) - β Malware analysis (isolated environment)
Mode Selection Decision Tree
1 | What do you need the VM for? |
Network Troubleshooting in Practice
Systematic Troubleshooting Method
Troubleshooting Pyramid (bottom to top):
1 | 7. Application Layer β Application configuration errors |
Troubleshooting Order: Bottom to top, confirm layer by layer.
Classic Fault Cases
Fault 1: Ping Works But Website Doesn't
Symptoms: 1
2
3$ping example.com
64 bytes from example.com: time=20ms β Network works$curl http://example.com
curl: (7) Failed to connect β Website doesn't work
Root Cause Analysis: - ping uses
ICMP protocol, doesn't depend on ports - HTTP uses
TCP port 80, service may not be running
Troubleshooting Steps: 1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16# 1. Confirm port is listening
netstat -tnlp | grep 80
# or
ss -tnlp | grep 80
# 2. Test port connectivity
telnet example.com 80
# or
nc -zv example.com 80
# 3. Check service status
systemctl status nginx
systemctl status apache2
# 4. View service logs
tail -f /var/log/nginx/error.log
Common Causes: 1. Web server not running 2. Port occupied 3. Firewall blocking port 80 4. Server listening on 127.0.0.1 instead of 0.0.0.0
Fault 2: Domain Cannot Be Resolved
Symptoms: 1
2$ping www.example.com
ping: cannot resolve www.example.com: Unknown host
Troubleshooting Steps: 1
2
3
4
5
6
7
8
9
10
11
12# 1. Test DNS server
ping 8.8.8.8 β If works, network is fine
# 2. Manual DNS query
nslookup www.example.com
dig www.example.com
# 3. Check DNS configuration
cat /etc/resolv.conf
# 4. Try changing DNS
echo "nameserver 8.8.8.8" > /etc/resolv.conf
Common Causes: 1. DNS server failure 2. DNS hijacking/poisoning 3. Domain not registered or expired
Fault 3: VM Cannot Access Internet
Scenario: VMware VM configured in NAT mode but cannot access internet
Troubleshooting Steps:
1 | # 1. Check IP configuration in VM |
Common Causes and Solutions: | Symptom | Cause |
Solution | |---------|-------|----------| | No IP address | DHCP not
started | dhclient or restart network service | | Has IP
but no gateway | Route configuration error | Add default route | | Can
ping gateway but not internet | VMware NAT service not running | Restart
VMware NAT Service | | Can ping IP but not domain | DNS configuration
error | Modify /etc/resolv.conf |
hosts File Deep Dive
π Intuitive Understanding: The Local "Address Book"
The hosts file is like your phone's contact list. When you call "Mom", the phone checks the contact list first and directly dials the corresponding number, instead of looking through the phone book. Similarly, when accessing a domain, the system checks the hosts file first before querying DNS servers.
How It Works
Domain Resolution Priority: 1
2
3
4
51. Browser cache
2. OS cache
3. hosts file β High priority
4. Local DNS server
5. Recursive DNS query
hosts File Location
| Operating System | Path |
|---|---|
| Windows | C:\Windows\System32\drivers\etc\hosts |
| macOS | /etc/hosts |
| Linux | /etc/hosts |
Editing hosts File
Windows: 1
2# Run Notepad as administrator
notepad C:\Windows\System32\drivers\etc\hosts
macOS/Linux: 1
2
3sudo nano /etc/hosts
# or
sudo vim /etc/hosts
Practical Applications
Application 1: Local Development Testing
Scenario: You're developing a website and need to
access local server using domain dev.example.com
Configuration: 1
2
3
4# /etc/hosts
127.0.0.1 dev.example.com
127.0.0.1 api.example.com
127.0.0.1 admin.example.com
Access: 1
2http://dev.example.com:3000 β Access local port 3000
http://api.example.com:8080 β Access local port 8080
Note: hosts file doesn't handle ports, you need to specify them in the URL.
Application 2: Multiple Domains Pointing to Same Server
Scenario: Your server has multiple domains, all pointing to the same IP
Configuration: 1
2
3192.168.1.100 www.example.com
192.168.1.100 blog.example.com
192.168.1.100 shop.example.com
The server distinguishes domains through the HTTP Header's
Host field and returns different content.
Application 3: Block Ads/Malicious Sites
Principle: Point ad domains to invalid IPs
Configuration: 1
2
30.0.0.0 ad.example.com
0.0.0.0 tracker.example.com
127.0.0.1 malware.com
When the browser tries to load these domains, it will connect to
0.0.0.0 (invalid address) or 127.0.0.1 (local
machine), thus blocking access.
Common Issues
Issue 1: hosts Modification Not Taking Effect
Cause: DNS cache
Solution: 1
2
3
4
5
6
7
8
9
10
11# Windows
ipconfig /flushdns
# macOS
sudo dscacheutil -flushcache
sudo killall -HUP mDNSResponder
# Linux
sudo systemd-resolve --flush-caches
# or
sudo /etc/init.d/nscd restart
Issue 2: Chrome Still Accesses Old IP
Cause: Browser cache
Solution: 1. Clear browser cache 2. Enter
chrome://net-internals/#dns in Chrome address bar, click
"Clear host cache" 3. Test using incognito mode
Complete Process of Browser Accessing a Webpage

Detailed Step-by-Step Analysis
1. User Enters URL
1 | User input: https://www.example.com/page.html |
Browser parses URL: - Protocol: https (uses port 443) -
Domain: www.example.com - Path: /page.html
2. DNS Resolution
1 | www.example.com β 93.184.216.34 |
Resolution Steps: 1. Check browser cache 2. Check OS cache (hosts file) 3. Query local DNS server (usually router or ISP's DNS) 4. Recursive query: Root DNS β .com TLD DNS β example.com authoritative DNS 5. Return IP address
3. Establish TCP Connection (Three-Way Handshake)
1 | Client β SYN β Server |
4. TLS Handshake (HTTPS)
1 | Client β ClientHello β Server |
5. Send HTTP Request
1 | GET /page.html HTTP/1.1 |
6. Server Processes Request
Server (Nginx/Apache): 1. Parse HTTP request 2. Find
/page.html file 3. Generate HTTP response
7. Return HTTP Response
1 | HTTP/1.1 200 OK |
8. Browser Renders Page
- Parse HTML: Build DOM tree
- Parse CSS: Build CSSOM tree
- Execute JavaScript: Modify DOM/CSSOM
- Render: Merge DOM and CSSOM, draw page
9. Load Resources
Resources in the page (images, CSS, JS) trigger new HTTP requests, repeating steps 2-7.
10. Close Connection (Four-Way Handshake)
1 | Client β FIN β Server |
β Network Knowledge Summary
Core Concepts Review
- Special IP Addresses:
127.0.0.1: Loopback (localhost)0.0.0.0: Bind to all NICs
- NAT Principle:
- Internal private IP ββ Public IP mapping
- Router/host acts as "translator"
- Network Modes:
- Bridged: VM = Real computer (same subnet)
- NAT: VM in host's virtual network (recommended)
- Host-Only: Complete isolation, only communicates with host
- Troubleshooting:
- Ping works but can't access β Check ports and services
- Domain resolution fails β Check DNS configuration
- VM can't access internet β Check gateway and VMware services
- hosts File:
- Higher priority than DNS
- Used for local testing, blocking ads
- Need to clear cache after modification
Common Misconceptions
| β Wrong Understanding | β Correct Understanding |
|---|---|
| 0.0.0.0 can be accessed | 0.0.0.0 can only be used as binding address |
| NAT mode VM can be accessed externally | Needs port forwarding configuration |
| hosts file can specify ports | hosts only handles IP mapping |
| Ping works = website normal | ping uses ICMP, website uses TCP |
Part 2: Power Supply Calculation
Formula:
Example:
- CPU (i5-13600K): 125W
- GPU (RTX 4070): 200W
- Other: 85W
- Total: 410W Γ 1.3 = 533W β Choose 650W PSU
| Config Level | Recommended | Example |
|---|---|---|
| Office | 300-400W | No dGPU |
| Light Gaming | 450-550W | RTX 3050 |
| Mainstream | 650-750W | RTX 4060/4070 |
| High-end | 850-1000W | RTX 4080/4090 |
Part 3: Common Faults (30+ Cases)
Boot Issues
Fault 1: No response when pressing power button
- Check power cable
- Check PSU switch
- Check 24pin + 8pin cables
- Test with different PSU
Fault 2: Fans spin but no display
- Reseat RAM (80% success rate)
- Check monitor cable
- Clear CMOS
- Minimal test (CPU+RAM+GPU only)
Series Summary Cheat
Computer 1024 disk 1000, broadband bits speed bytes;
CPU brain RAM desk, Intel single-core AMD multi-core;
DDR gens up freq down voltage, dual-channel doubles bandwidth;
HDD slow but large SSD fast but pricey, TLC mainstream thousand enough;
Motherboard interfaces each role, GPU parallel CPU serial;
NIC gigabit check cable, PSU power multiply 1.3;
Black screen wipe fingers, blue screen check RAM!
π Series Complete! Thank you for reading!
Part 2: Power Supply (PSU) - The Stable Power Source
PSU Specifications Explained
80 PLUS Certification
| Certification | Load Efficiency | Price | Application |
|---|---|---|---|
| 80 PLUS | β₯ 80% | $30+ | Entry |
| Bronze | β₯ 85% | $50+ | Mainstream |
| Silver | β₯ 88% | $60+ | Mid-range |
| Gold | β₯ 90% | $70+ | Recommended β |
| Platinum | β₯ 92% | $100+ | High-end |
| Titanium | β₯ 94% | $150+ | Enthusiast |
Efficiency meaning:
Assuming PC power consumption 400W:
- 80% efficiency: Draw from wall = 400W Γ· 0.8 = 500W (waste 100W)
- 90% efficiency (Gold): Draw from wall = 400W Γ· 0.9 = 444W (waste 44W)
Annual electricity savings (8 hours/day):
Modular vs Non-Modular PSU
| Type | Definition | Pros | Cons | Price Diff |
|---|---|---|---|---|
| Non-modular | All cables fixed | Cheap | Hard cable management, extra cables | Baseline |
| Semi-modular | Main cables fixed, others removable | Balanced | Main cables still fixed | +$10 |
| Fully modular | All cables removable | Easy management, aesthetic | Expensive | +$20 |
Recommendation:
- Beginners β Non-modular (cheaper)
- Aesthetics priority β Fully modular
PSU Wattage Selection
Real-world cases:
Config 1: Office PC 1
2
3
4
5
6
7CPU: i3-12100 (60W)
Motherboard: H610 (30W)
RAM: 8GB DDR4 (5W)
SSD: 512GB (5W)
Total: 100W
Recommended PSU: 100W Γ 1.5 = 150W β Choose 300W PSU
Config 2: Mainstream Gaming 1
2
3
4
5
6
7
8
9CPU: i5-13600K (125W)
GPU: RTX 4060 (115W)
Motherboard: B760 (50W)
RAM: 16GB DDR5 (10W)
SSD: 1TB NVMe (5W)
Fans: 20W
Total: 325W
Recommended PSU: 325W Γ 1.5 = 487.5W β Choose 650W PSU β
Config 3: High-end Workstation 1
2
3
4
5
6
7
8
9CPU: i9-13900K (253W)
GPU: RTX 4090 (450W)
Motherboard: Z790 (80W)
RAM: 64GB DDR5 (20W)
SSD: 2TB NVMe Γ 2 (10W)
Fans/AIO/RGB: 50W
Total: 863W
Recommended PSU: 863W Γ 1.3 = 1122W β Choose 1200W PSU β
PSU Q&A
Q1: Does higher wattage PSU consume more electricity?
A: No! Actual PC power consumption determines electricity usage!
Example:
- You buy 1000W PSU
- PC actual consumption 300W
- Draw from wall = 300W Γ· efficiency (e.g., 90%) = 333W
Conclusion: PSU wattage is just upper limit, won't waste power without reason!
Note:
- β οΈ PSU most efficient at 50-80% load
- Too small (>90% load): Efficiency drops, easy to overheat
- Too large (<20% load): Slightly lower efficiency, but impact minor
Q2: How often to replace PSU?
A: Normal use 5-10 years!
Replacement signals:
- PC frequently restarts unexpectedly
- Boot difficult (need multiple presses)
- Strange smell (capacitor burn smell)
- Fan abnormal noise or not spinning
Extend lifespan:
- β Buy brand-name PSU (Seasonic, Delta, FSP)
- β Buy Gold+ certification
- β Avoid full-load operation (leave 20% headroom)
- β Regular dust cleaning (every 6 months)
Part 3: Cooling Systems
CPU Cooler Types
Air Cooling
| Type | Cooling Capacity | Price | Suitable TDP | Representative |
|---|---|---|---|---|
| Stock cooler | 65W | $0 (CPU included) | β€ 65W | Intel stock |
| Single tower | 150W | $15-30 | β€ 125W | Hyper 212 |
| Dual tower | 220W | $45-70 | β€ 200W | Noctua NH-D15 |
| High-end air | 280W | $70+ | β€ 250W | Dark Rock Pro 4 |
Liquid Cooling
| Type | Cooling Capacity | Price | Suitable TDP | Noise |
|---|---|---|---|---|
| 120 AIO | 180W | $45 | β€ 150W | Medium |
| 240 AIO | 250W | $70 | β€ 200W | Low |
| 360 AIO | 350W | $120 | β€ 300W | Very low |
| Custom loop | 400W+ | $300+ | Any | Very low |
Selection recommendations:
| CPU Model | TDP | Recommended Cooler |
|---|---|---|
| i3/R3 | 65W | Stock or single tower |
| i5/R5 (non-K) | 65-95W | Single tower |
| i5K/R5X | 125-150W | Dual tower or 240 AIO |
| i7K/R7X | 150-200W | Dual tower or 280 AIO |
| i9K/R9X | 200-250W | 360 AIO |
Part 4: Complete Build Process (Detailed)
Pre-build Checklist
Tools: 1. β Phillips screwdriver (essential) 2. β Anti-static wrist strap (recommended) 3. β Cable ties (for cable management) 4. β Thermal paste (usually included with cooler)
Work environment:
- β Dry, ventilated room
- β Wooden desk (avoid static)
- β On carpet (generates static easily)
Building Steps (20 steps detailed)
Steps 1-5: Installing CPU and RAM to Motherboard
1 | 1. Open CPU socket cover |
Steps 6-10: Installing Motherboard into Case
1 | 6. Install I/O shield in case |
Steps 11-15: Installing Storage and GPU
1 | 11. Install M.2 SSD to motherboard M.2 slot |
Steps 16-20: Connect Cables and Test
1 | 16. Connect CPU 8pin power |
First Boot Checklist
1 | β 1. Does monitor display image? |
Part 5: Complete Troubleshooting Manual
Hardware Faults (15 issues)
1. Memory-related
- Black screen on boot β Reseat RAM, clean golden fingers
- Frequent BSOD β MemTest86 test, reduce frequency
- Incorrect capacity β Check 32/64-bit system, single-stick test
2. Disk-related
- SSD slow β Check free space, 4K alignment, TRIM
- Not detected β Check SATA cable, BIOS mode (AHCI)
- Bad sectors β chkdsk /f /r, backup data
3. GPU-related
- Artifacts β Check power cables, reinstall driver, cooling
- Low performance β Confirm plugged into GPU (not motherboard)
- Black screen β Check PCIe slot, clean golden fingers
4. Motherboard-related
- USB not detected β Update chipset driver, try rear ports
- M.2 not detected β Check BIOS, confirm protocol compatibility
- Dead CMOS battery β Replace CR2032 battery
5. PSU-related
- Auto restart β Check if wattage sufficient
- Won't power on β Test PSU (paperclip short test)
Software Faults (15 issues)
6. System issues
- Slow boot β Clean startup items, check SSD health
- BSOD β Check error code, update drivers
- System stuttering β Task Manager check resource usage
7. Network issues
- Disconnects β Update NIC driver, replace cable
- Slow WiFi β Switch to 5G band, adjust channel
- Cannot connect β Check IP config, DNS settings
8. Performance issues
- Low gaming FPS β Update GPU driver, check power mode
- Program crashes β Check memory usage, close background
- Overheating β Clean dust, replace thermal paste
Part 6: Practical Tool Recommendations
Hardware Detection Tools
| Tool Name | Purpose | Free? |
|---|---|---|
| CPU-Z | View CPU/RAM/motherboard details | β Yes |
| GPU-Z | View GPU info, real-time monitoring | β Yes |
| CrystalDiskInfo | Disk health detection | β Yes |
| HWiNFO64 | Comprehensive system monitoring | β Yes |
| AIDA64 | Stress test, temperature monitoring | β Paid |
Performance Testing Tools
| Tool Name | Test Item | Free? |
|---|---|---|
| 3DMark | GPU performance benchmark | β οΈ Partially free |
| Cinebench | CPU render performance | β Yes |
| CrystalDiskMark | Disk read/write speed | β Yes |
| MemTest86 | Memory stability test | β Yes |
β Q&A: Network, Power & Troubleshooting Questions
Q1: Wi-Fi 6 vs Ethernet - When to Use Each?
Short answer: Ethernet for stability and speed, Wi-Fi for convenience and mobility.
Decision tree:
1 | Is your device stationary? (Desktop PC, NAS, gaming console) |
Real-world scenarios:
| Use Case | Recommendation | Reason |
|---|---|---|
| Gaming PC | Ethernet (Gigabit) | Latency matters: 1ms Ethernet vs 20ms Wi-Fi = noticeable difference |
| 4K Streaming | Either works | Wi-Fi 6 handles 4K (25 Mbps) easily; Ethernet more stable |
| File transfers | Ethernet (2.5G/10G) | Large files benefit from consistent 300+ MB/s speeds |
| Laptop/Phone | Wi-Fi 6 | Mobility > raw speed for most users |
| Home office | Ethernet for desktop | Video calls more stable, no dropouts |
When Wi-Fi 6 beats Ethernet:
- Multiple devices (Wi-Fi 6 handles 8+ devices better than older standards)
- No cable routing possible (apartment rentals, aesthetics)
- Device mobility required
Bottom line: Desktop PCs and servers β Ethernet. Everything else β Wi-Fi 6 is fine.
Q2: Mesh Network vs Traditional Router - Pros/Cons
Traditional router: Single device broadcasting signal. Signal weakens with distance and walls.
Mesh network: Multiple nodes working together, creating seamless coverage across large areas.
Comparison table:
| Factor | Traditional Router | Mesh Network |
|---|---|---|
| Coverage | 100-150 sq ft (one floor) | 3000+ sq ft (multi-floor) |
| Setup complexity | Simple (plug and play) | Moderate (place nodes strategically) |
| Cost | $50-200 | $150-500 |
| Speed | Fast near router, drops off | Consistent across coverage |
| Dead zones | Common (corners, upstairs) | Rare (nodes fill gaps) |
| Best for | Small apartments, single floor | Large homes, multi-story buildings |
When to choose traditional router:
- β Apartment < 1000 sq ft
- β Single floor layout
- β Budget conscious ($50-100)
- β Simple needs (browsing, streaming)
When to choose mesh:
- β House > 2000 sq ft
- β Multiple floors
- β Dead zones with current router
- β Need seamless roaming (devices switch nodes automatically)
Mesh placement strategy: 1. Main node: Near modem, central location 2. Satellite nodes: 30-50 feet apart, avoid corners 3. Test: Walk around with phone, check signal strength app 4. Rule of thumb: Each node covers ~1500 sq ft
Popular choices:
- Traditional: TP-Link Archer AX50 (
130) - Mesh: Eero 6 (
300), ASUS ZenWiFi AX ($400)
Q3: PSU Wattage Calculator - How Much Do I Need?
Quick formula:
(CPU + GPU + 100W overhead) Γ 1.3 = Minimum PSU wattage
Step-by-step calculation:
Step 1: List component power draw
| Component | Power Draw (W) | Notes |
|---|---|---|
| CPU | 65-253W | Check TDP (Thermal Design Power) |
| GPU | 75-450W | Gaming GPUs: RTX 3050 (130W) to RTX 4090 (450W) |
| Motherboard | 30-80W | Basic boards ~30W, high-end ~80W |
| RAM | 5-20W | 8GB ~5W, 32GB ~15W |
| SSD (NVMe) | 3-8W | Per drive |
| HDD | 5-10W | Per drive |
| Fans | 2-5W | Per fan |
| RGB lighting | 5-20W | Optional, varies |
| AIO cooler | 10-20W | Pump + fans |
Step 2: Real-world examples
Example 1: Office PC 1
2
3
4
5
6
7CPU: i3-12100 (60W)
Motherboard: H610 (30W)
RAM: 16GB (10W)
SSD: 512GB (5W)
Fans: 2Γ (10W)
Total: 115W
Calculation: 115W Γ 1.5 = 172W β Choose 300-400W PSU β
Example 2: Mainstream Gaming 1
2
3
4
5
6
7
8
9
10CPU: i5-13600K (125W)
GPU: RTX 4060 (115W)
Motherboard: B760 (50W)
RAM: 32GB DDR5 (15W)
SSD: 1TB NVMe (5W)
Fans: 4Γ (20W)
RGB: (15W)
Total: 345W
Calculation: 345W Γ 1.3 = 448W β Choose 650W PSU β
(Headroom for future upgrades, efficiency at 50-70% load)
Example 3: High-end Workstation 1
2
3
4
5
6
7
8
9CPU: i9-13900K (253W)
GPU: RTX 4090 (450W)
Motherboard: Z790 (80W)
RAM: 64GB DDR5 (20W)
SSD: 2TB NVMe Γ 2 (10W)
Fans: 6Γ (30W)
AIO: 360mm (20W)
Total: 863W
Calculation: 863W Γ 1.3 = 1122W β Choose 1200W PSU β
Step 3: Online calculators
- OuterVision PSU Calculator: Most accurate, accounts for overclocking
- Newegg PSU Calculator: Simple, good for quick estimates
- Be Quiet! Calculator: European market focus
Common mistakes:
- β Buying exactly calculated wattage (no headroom)
- β Ignoring GPU power spikes (RTX 30/40 series can spike 20% above TDP)
- β Forgetting future upgrades (add 100-200W buffer)
Efficiency sweet spot: PSUs are most efficient at 50-80% load. If you calculate 400W, a 650W PSU runs at 62% load = optimal efficiency.
Q4: 80 PLUS Ratings - Worth the Cost?
80 PLUS certification levels:
| Rating | 20% Load | 50% Load | 100% Load | Price Premium | Worth It? |
|---|---|---|---|---|---|
| 80 PLUS | 80% | 80% | 80% | Baseline | Entry builds |
| Bronze | 82% | 85% | 82% | +$10-15 | Budget builds |
| Silver | 85% | 88% | 85% | +$15-20 | Rare, skip |
| Gold | 87% | 90% | 87% | +$20-30 | β Best value |
| Platinum | 90% | 92% | 89% | +$40-60 | High-end builds |
| Titanium | 90% | 94% | 91% | +$80-100 | Servers/workstations |
Cost-benefit analysis:
Scenario: 500W PC running 8 hours/day, electricity$0.12/kWh
80 PLUS (80% efficiency):
- Wall draw: 500W Γ· 0.80 = 625W
- Daily cost: 0.625kW Γ 8h Γ
0.60 - Annual cost:$219
Gold (90% efficiency):
- Wall draw: 500W Γ· 0.90 = 556W
- Daily cost: 0.556kW Γ 8h Γ
0.53 - Annual cost: $195
- Savings:$24/year
Break-even: If Gold PSU costs $30 more, pays for itself in 1.25 years.
When Gold is worth it:
- β PC runs 6+ hours daily
- β High power consumption (400W+)
- β Electricity costs >$0.10/kWh
- β Planning to keep PSU 5+ years
When Bronze is enough:
- β PC runs < 4 hours/day
- β Low power consumption (< 300W)
- β Budget build (< $800 total)
- β Electricity is cheap (<$0.08/kWh)
Real-world recommendation: Gold is the sweet spot for most builds. The $20-30 premium pays off within 2 years, and Gold PSUs typically have better build quality and longer warranties (7-10 years vs 3-5 years for Bronze).
Q5: Modular vs Non-Modular PSU
Types explained:
| Type | Cables | Pros | Cons | Price |
|---|---|---|---|---|
| Non-modular | All fixed | Cheapest, no loose cables | Cable clutter, unused cables | Baseline |
| Semi-modular | 24pin + CPU fixed, others removable | Balanced, cleaner than non-modular | Main cables still fixed | +$10-15 |
| Fully modular | All removable | Cleanest, custom cables possible | Most expensive | +$20-30 |
Visual comparison:
Non-modular: 24pin, CPU 8pin, PCIe cables, SATA cables, Molex cables - all permanently attached. Even if you only use 3 cables, you have 8 cables dangling.
Semi-modular: 24pin and CPU 8pin fixed (you always need these), but SATA, PCIe, Molex cables are removable. Cleaner, but main cables still there.
Fully modular: Every cable detaches. Only connect what you need. Perfect for custom builds, RGB showcases, small form factor cases.
When to choose each:
Non-modular:
- β Budget builds (<$600)
- β First-time builders (less to think about)
- β Cases with good cable management (hidden PSU shroud)
Semi-modular:
- β Most users (best balance)
- β Mid-range builds ($800-1500)
- β Want cleaner look without premium price
Fully modular:
- β High-end builds ($1500+)
- β Small form factor cases (ITX)
- β Custom water cooling loops
- β Aesthetic-focused builds (glass side panels, RGB)
Cable management impact:
Non-modular in budget case: Cables everywhere, blocks airflow, looks messy.
Fully modular in premium case: Only 4-5 cables visible, clean routing, better airflow.
Bottom line: For most users, semi-modular is the sweet spot. You save$10-15 vs fully modular, and the fixed cables (24pin, CPU) are always needed anyway. Only go fully modular if you're building a showcase PC or need custom cable lengths.
Q6: Cable Management Importance
Why it matters:
- Airflow: Tangled cables block air paths β higher temperatures β thermal throttling β performance loss
- Maintenance: Messy cables make upgrades/troubleshooting harder (can't see what's connected where)
- Aesthetics: Clean builds look professional (if you care)
- Safety: Loose cables can get caught in fans, cause shorts
Temperature impact (real test data):
| Cable Management | CPU Temp (Idle) | CPU Temp (Load) | GPU Temp (Load) |
|---|---|---|---|
| Poor (cables everywhere) | 45Β° C | 85Β° C | 82Β° C |
| Good (routed behind tray) | 38Β° C | 72Β° C | 68Β° C |
| Difference | -7Β° C | -13Β° C | -14Β° C |
13Β° C difference under load = can mean the difference between stable operation and thermal throttling.
Cable management checklist:
Before building:
- β Plan cable routes (which cables go where)
- β Measure distances (avoid cables too short/long)
- β Group cables by destination (CPU, GPU, storage, fans)
During building:
- β Route 24pin behind motherboard tray
- β Route CPU 8pin through top grommet
- β Bundle excess cable length with zip ties
- β Use case's built-in cable management features (tie-down points, channels)
After building:
- β Tuck loose ends behind PSU shroud
- β Use Velcro straps (reusable) instead of zip ties
- β Leave 10-20% slack (don't pull cables taut)
- β Test airflow (smoke test or thermal monitoring)
Common mistakes:
- β Pulling cables too tight (strain on connectors)
- β Blocking front intake fans with cables
- β Leaving unused cables dangling (tuck them away)
- β Using too many zip ties (hard to remove later)
Tools needed:
- Zip ties or Velcro straps ($5)
- Cable combs (optional, for aesthetics,$10)
- Patience (free, but essential)
Time investment: Good cable management adds 30-60 minutes to build time, but saves hours during troubleshooting and upgrades. Worth it.
Q7: Troubleshooting Boot Failures - Systematic Approach
Boot failure decision tree:
1 | PC won't turn on |
Systematic troubleshooting steps (in order):
Step 1: Power check (30 seconds)
Step 2: Visual inspection (2 minutes)
Step 3: Minimal boot test (5 minutes)
If minimal boot works: Add components one by one until failure returns β that component is the problem.
If minimal boot fails: Likely CPU, motherboard, or PSU issue.
Step 4: Component isolation (10-15 minutes)
RAM test:
GPU test:
Storage test:
Step 5: BIOS/UEFI checks (5 minutes)
Step 6: Advanced diagnostics (if above fails)
PSU test:
- Paperclip test: Unplug 24pin, short green wire (pin 16) to any black wire (ground)
- PSU fan should spin β PSU works
- Use PSU tester ($15) for more accurate results
CPU test:
- Check for bent pins (Intel) or damaged socket (AMD)
- Try different CPU if available
- Check CPU power (8pin connector)
Motherboard test:
- Check for bulging capacitors (swollen tops)
- Smell test (burned electronics smell = bad)
- Try different motherboard if available
Common boot error codes:
| Symptom | Likely Cause | Solution |
|---|---|---|
| No power at all | PSU, power cable, wall outlet | Check power chain |
| Fans spin, no display | RAM (80%), GPU, monitor | Reseat RAM first |
| Beep codes | Varies by manufacturer | Check motherboard manual |
| Boot loop | CPU overheating, RAM, PSU | Check temps, reseat RAM |
| BSOD immediately | RAM, drivers, storage | MemTest86, check drives |
Time estimate: Systematic approach takes 30-60 minutes but finds the problem 90% of the time. Random guessing can take hours.
Q8: Temperature Monitoring and Cooling Optimization
Safe temperature ranges:
| Component | Idle Temp | Load Temp | Max Safe | Throttle Point |
|---|---|---|---|---|
| CPU | 30-50Β° C | 60-80Β° C | 95-100Β° C | 90-100Β° C |
| GPU | 30-40Β° C | 70-85Β° C | 90-95Β° C | 83-90Β° C |
| SSD (NVMe) | 30-50Β° C | 50-70Β° C | 80-85Β° C | 70Β° C+ |
| RAM | 30-40Β° C | 40-50Β° C | 60Β° C+ | Rarely throttles |
Monitoring tools:
| Tool | Best For | Free? |
|---|---|---|
| HWiNFO64 | Comprehensive (all sensors) | β Yes |
| MSI Afterburner | GPU + overlay in games | β Yes |
| Core Temp | CPU only, simple | β Yes |
| GPU-Z | GPU detailed info | β Yes |
| AIDA64 | Stress testing + monitoring | β Paid |
Cooling optimization checklist:
Case airflow (most important):
Fan configuration examples:
Budget setup (2 fans):
- 1Γ front intake
- 1Γ rear exhaust
- Result: Basic airflow, works for low-power builds
Mainstream setup (4-5 fans):
- 2Γ front intake
- 1Γ rear exhaust
- 1Γ top exhaust (optional)
- Result: Good airflow, handles mid-range GPUs
High-end setup (6+ fans):
- 3Γ front intake
- 1Γ rear exhaust
- 2Γ top exhaust
- Result: Excellent airflow, handles RTX 4080/4090
CPU cooler optimization:
GPU cooling (if overheating):
Quick fixes for high temperatures:
CPU too hot: 1. Check cooler mounting (might be loose) 2. Reapply thermal paste (if > 2 years old) 3. Increase case intake fans 4. Clean dust from CPU cooler fins
GPU too hot: 1. Improve case airflow (add intake fans) 2. Remove side panel temporarily (test if case is the issue) 3. Adjust fan curve (MSI Afterburner) 4. Check GPU thermal pads (if replaced recently)
SSD too hot: 1. Add heatsink (M.2 drives often include one) 2. Improve case airflow 3. Check if drive is under GPU (gets hot air) 4. Consider thermal pad between SSD and heatsink
Temperature monitoring routine:
- Weekly: Quick check during gaming/workload (HWiNFO64)
- Monthly: Full stress test (AIDA64 or Prime95 + FurMark)
- Every 6 months: Clean dust, check thermal paste
Red flags (take action immediately):
- CPU > 90Β° C under load
- GPU > 85Β° C under load
- SSD > 75Β° C
- Idle temps > 60Β° C (something wrong)
Q9: BIOS Settings for Stability
Essential BIOS settings (in order of importance):
1. Memory settings (most common stability issues):
| Setting | Recommended Value | Why |
|---|---|---|
| XMP/DOCP | Enable (if RAM supports it) | Runs RAM at advertised speed |
| Memory Frequency | Auto or manual (match RAM spec) | Don't overclock unless stable |
| DRAM Voltage | Auto (usually 1.35V for DDR4, 1.25V for DDR5) | Too low = crashes, too high = damage |
| Memory Training | Enable | Ensures RAM works with CPU |
If system unstable with XMP enabled:
- Try XMP Profile 2 (if available, usually more conservative)
- Manually set frequency one step below rated speed
- Increase DRAM voltage by 0.05V (be careful, don't exceed 1.5V DDR4)
- Test with MemTest86 (4+ hours, no errors)
2. CPU settings:
| Setting | Recommended Value | Notes |
|---|---|---|
| CPU Core Ratio | Auto (unless overclocking) | Leave alone for stability |
| CPU Voltage | Auto | Manual only if overclocking |
| CPU Power Limits | Auto | Prevents overheating |
| C-States | Enable (for power saving) | Can disable if causing issues |
| Turbo Boost | Enable | Performance boost, safe |
3. Boot settings:
| Setting | Recommended Value | Why |
|---|---|---|
| Boot Mode | UEFI (not Legacy) | Required for Windows 11, faster boot |
| Secure Boot | Enable (if Windows 11) | Security feature |
| Fast Boot | Enable (after initial setup) | Faster startup |
| Boot Priority | SSD/Windows drive first | Boots from correct drive |
4. PCIe settings:
| Setting | Recommended Value | Notes |
|---|---|---|
| PCIe Gen | Auto (or Gen 4 if supported) | Let system detect |
| Resizable BAR | Enable (if GPU supports) | Performance boost for RTX 30/40 series |
| Above 4G Decoding | Enable (if Resizable BAR enabled) | Required for Resizable BAR |
5. Fan control (for noise/temperature balance):
| Setting | Recommended Value | Impact |
|---|---|---|
| CPU Fan | PWM mode, curve based on CPU temp | Quieter when idle |
| Case Fans | PWM mode, curve based on CPU/GPU temp | Balance noise and cooling |
| Fan Stop | Disable (fans always spin) | Prevents dust buildup |
Fan curve example (CPU fan):
- 0-50Β° C: 30% speed (quiet)
- 50-70Β° C: 50% speed (moderate)
- 70-80Β° C: 70% speed (noticeable)
- 80Β° C+: 100% speed (loud but cool)
6. Voltage settings (advanced, be careful):
| Setting | Default | Safe Range | Warning |
|---|---|---|---|
| CPU VCore | Auto | Β±0.1V from default | Too high = damage, too low = crashes |
| DRAM Voltage | 1.35V (DDR4) | 1.2-1.5V | Exceed 1.5V = risk |
| SoC Voltage | Auto | Don't touch unless needed | AMD only, affects stability |
Stability testing after BIOS changes:
Quick test (30 minutes): 1. Boot to Windows 2. Run Prime95 Small FFTs (CPU stress) 3. Run FurMark (GPU stress) 4. Run both simultaneously (15 minutes) 5. No crashes/errors = stable
Thorough test (overnight): 1. MemTest86: 4+ passes (4-8 hours) 2. Prime95 Blend: 2+ hours 3. Real-world usage: Gaming/workload for extended period
Common BIOS mistakes:
- β Enabling XMP without testing (can cause crashes)
- β Setting voltages too high (damages components)
- β Disabling all power saving features (wastes electricity)
- β Changing multiple settings at once (hard to isolate issues)
BIOS update guide:
- When: Only if experiencing issues or need new CPU support
- How: Download from motherboard manufacturer, put on USB, use BIOS flash utility
- Risk: Power loss during update = bricked motherboard (rare but possible)
- Recommendation: Update only if necessary, use UPS if possible
BIOS reset (if things go wrong): 1. Soft reset: Load optimized defaults in BIOS 2. Hard reset: Remove CMOS battery for 30 seconds (unplug PSU first) 3. Clear CMOS jumper: Short the jumper pins (check manual)
Q10: Common Building Mistakes to Avoid
Mistake #1: Forgetting I/O shield (most common)
What happens: Install motherboard, realize I/O shield missing, have to remove everything.
Prevention: Install I/O shield before motherboard (snaps into case from inside).
Mistake #2: Not testing outside case first
What happens: Build everything in case, doesn't boot, have to remove components to test.
Prevention: "Breadboard" test first:
- Place motherboard on box (non-conductive surface)
- Install CPU, RAM, GPU, connect PSU
- Power on β Does it POST?
- If yes β Install in case. If no β Problem isolated to components.
Saves: 2-3 hours of disassembly/reassembly.
Mistake #3: CPU installation force
What happens: Bent pins (AMD) or damaged socket (Intel),$200+ repair.
Prevention:
- AMD: CPU drops in with gravity (no force). If it doesn't fit, check orientation (golden triangle).
- Intel: CPU sits on pins, close cover gently. If resistance, stop and check.
Rule: If you're applying force, you're doing it wrong.
Mistake #4: RAM not fully seated
What happens: System won't boot, or boots with half RAM capacity.
Prevention:
- Push until both clips click (should hear/feel click)
- Check both sides (some boards have clips on both ends)
- RAM should be perfectly straight, not angled
Test: If RAM sticks out even 1mm, it's not seated.
Mistake #5: Forgetting standoffs
What happens: Motherboard shorts on case, system won't boot or damages components.
Prevention:
- Case usually comes with standoffs pre-installed
- Check: Standoffs should align with motherboard screw holes
- Don't install extra standoffs where there's no hole (causes shorts)
Mistake #6: PSU wattage too low
What happens: System crashes under load, random restarts, PSU failure.
Prevention: Use calculator (see Q3), add 20-30% headroom.
Mistake #7: Thermal paste mistakes
Common errors:
- β Too much paste (spills over, can short components)
- β Too little paste (poor contact, overheating)
- β Spreading manually (creates air bubbles)
- β Using paste that came with cooler (often low quality)
Correct method:
- Pea-sized drop in center
- Let cooler pressure spread it
- Or use spreader (thin, even layer)
Mistake #8: Cable management blocking airflow
What happens: Higher temperatures, thermal throttling, reduced performance.
Prevention: Route cables behind motherboard tray, use case's cable management features.
Mistake #9: Installing GPU in wrong slot
What happens: GPU runs at x4 or x8 instead of x16, performance loss.
Prevention: Install in top PCIe x16 slot (usually closest to CPU). Check manual if unsure.
Mistake #10: Not updating drivers
What happens: Poor performance, crashes, missing features.
Prevention:
- Chipset drivers: From motherboard manufacturer
- GPU drivers: From NVIDIA/AMD website (not Windows Update)
- Network drivers: From motherboard manufacturer
- BIOS: Only if needed (see Q9)
Mistake #11: Forgetting to remove protective film
What happens: Overheating (film on CPU cooler), no display (film on monitor).
Prevention: Check CPU cooler base, monitor screen, case side panels.
Mistake #12: Overtightening screws
What happens: Stripped threads, damaged components, hard to remove later.
Prevention:
- Finger tight + quarter turn with screwdriver
- If screw stops turning easily, stop (don't force)
- Use correct screwdriver size (don't strip heads)
Mistake #13: Not checking compatibility
Common issues:
- CPU not compatible with motherboard (wrong socket)
- RAM not compatible (DDR4 vs DDR5, wrong speed)
- GPU too large for case
- PSU cables not compatible (mixing modular cables from different PSUs = dangerous)
Prevention: Use PCPartPicker compatibility checker before buying.
Mistake #14: Installing OS on wrong drive
What happens: OS on slow HDD instead of fast SSD, slow boot times.
Prevention:
- Disconnect other drives during OS installation
- Install OS on fastest drive (NVMe SSD)
- Reconnect other drives after OS installed
Mistake #15: Not testing before closing case
What happens: Find issues after closing case, have to reopen.
Prevention:
- Test boot with side panel off
- Check all fans spinning
- Check temperatures normal
- Check all USB ports work
- Then close case and do cable management
Building checklist (print this):
Before building:
During building:
After building:
Time investment: Following this checklist adds 1-2 hours but prevents 10+ hours of troubleshooting. Worth it.
π Summary: PC Building & Troubleshooting Cheat Sheet
Quick Reference: Component Selection Decision Trees
PSU Selection Flowchart:
1 | Start: What's your total power draw? |
Network Selection Flowchart:
1 | Start: What's your use case? |
Cooling Selection Flowchart:
1 | Start: What's your CPU TDP? |
Troubleshooting Decision Trees
"PC Won't Turn On" Decision Tree:
1 | No response at all? |
"System Crashes/Restarts" Decision Tree:
1 | When does it crash? |
"Poor Performance" Decision Tree:
1 | What's slow? |
Temperature Monitoring Checklist
Daily/Weekly Checks:
Monthly Deep Check:
Every 6 Months:
Red Flags (take action immediately):
- π΄ CPU > 95Β° C under load
- π΄ GPU > 90Β° C under load
- π΄ SSD > 80Β° C
- π΄ Idle temps > 60Β° C (something wrong)
BIOS Settings Quick Reference
Essential Settings (check these first):
| Setting Category | Key Settings | Recommended Value |
|---|---|---|
| Memory | XMP/DOCP | Enable (if RAM supports) |
| DRAM Voltage | Auto (usually 1.35V DDR4) | |
| CPU | Turbo Boost | Enable |
| CPU Voltage | Auto (unless overclocking) | |
| Boot | Boot Mode | UEFI |
| Boot Priority | SSD first | |
| Fast Boot | Enable (after setup) | |
| PCIe | PCIe Gen | Auto |
| Resizable BAR | Enable (if GPU supports) | |
| Fans | CPU Fan | PWM, curve based on temp |
| Case Fans | PWM, curve based on temp |
Stability Testing After Changes: 1. Boot to Windows β No crashes? 2. Run Prime95 (15 min) β Stable? 3. Run MemTest86 (1 pass) β No errors? 4. If all pass β Settings stable β
Common Issues & Quick Fixes
| Symptom | Quick Fix | If That Doesn't Work |
|---|---|---|
| PC won't turn on | Check PSU switch, power cable | Test PSU (paperclip), check 24pin |
| No display | Reseat RAM (80% fix rate) | Try integrated graphics, clear CMOS |
| Boot loop | Reseat RAM, check CPU temp | Minimal boot test, check PSU |
| BSOD | Update drivers, check RAM | MemTest86, check storage health |
| High temps | Clean dust, check cooler mount | Reapply thermal paste, improve airflow |
| Slow performance | Check if OS on SSD, update drivers | Check temps, check RAM usage |
| WiFi disconnects | Update NIC driver, restart router | Change WiFi channel, check interference |
| GPU not detected | Reseat GPU, check power cables | Try different PCIe slot, update BIOS |
Building Process Checklist (Print This)
Pre-Build (30 minutes):
Build Phase 1: CPU & RAM (15 minutes):
Build Phase 2: Motherboard (20 minutes):
Build Phase 3: Storage & GPU (15 minutes):
Build Phase 4: Cables & Testing (30 minutes):
Post-Build (1-2 hours):
Total Time: 3-4 hours for first build, 2-3 hours for experienced builders.
Power Consumption Reference Table
Component Power Draw (typical values):
| Component | Idle Power | Load Power | Peak Power |
|---|---|---|---|
| CPU (i5-13600K) | 15W | 125W | 180W (boost) |
| CPU (i9-13900K) | 20W | 253W | 350W (boost) |
| GPU (RTX 4060) | 10W | 115W | 140W (spike) |
| GPU (RTX 4090) | 20W | 450W | 550W (spike) |
| Motherboard | 15W | 50W | 80W (high-end) |
| RAM (16GB) | 3W | 8W | 10W |
| SSD (NVMe) | 0.5W | 5W | 8W |
| HDD | 2W | 8W | 12W |
| Fans (each) | 1W | 3W | 5W |
| AIO Pump | 5W | 15W | 20W |
System Power Examples:
| System Type | Typical Load | Peak Load | Recommended PSU |
|---|---|---|---|
| Office PC | 80W | 120W | 300-400W |
| Light Gaming | 200W | 280W | 500-550W |
| Mainstream Gaming | 350W | 480W | 650W β |
| High-end Gaming | 550W | 750W | 850W |
| Workstation | 800W | 1100W | 1200W+ |
Power Spike Notes: Modern GPUs (RTX 30/40 series) can spike 20-30% above TDP for milliseconds. Always add headroom.
Network Speed Reference
Ethernet Standards:
| Standard | Speed | Real-World | Use Case |
|---|---|---|---|
| Fast Ethernet | 100 Mbps | 12 MB/s | Obsolete |
| Gigabit | 1 Gbps | 125 MB/s | Mainstream β |
| 2.5 Gigabit | 2.5 Gbps | 312 MB/s | High-end |
| 10 Gigabit | 10 Gbps | 1250 MB/s | Enterprise |
Wi-Fi Standards:
| Standard | Speed (Theoretical) | Real-World | Year |
|---|---|---|---|
| Wi-Fi 4 (n) | 600 Mbps | 50-100 Mbps | 2009 |
| Wi-Fi 5 (ac) | 3.5 Gbps | 200-500 Mbps | 2013 |
| Wi-Fi 6 (ax) | 9.6 Gbps | 400-800 Mbps | 2019 β |
| Wi-Fi 6E | 9.6 Gbps | 500-1000 Mbps | 2021 |
| Wi-Fi 7 | 46 Gbps | 1000+ Mbps | 2024 |
Real-World Speed Factors:
- Distance from router (signal strength)
- Walls/obstacles (2.4 GHz penetrates better, 5 GHz faster)
- Interference (neighbors' Wi-Fi, microwaves)
- Device capabilities (older devices = slower)
Final Troubleshooting Flowchart
1 | Problem occurs |
Systematic Approach (always works): 1. Isolate: Remove non-essential components (minimal boot) 2. Test: Add components one by one until problem returns 3. Identify: Problem component = last thing you added 4. Fix: Replace/update/reseat that component 5. Verify: Test again to confirm fix
Time Investment: Systematic approach takes 30-60 minutes but finds the problem 90% of the time. Random guessing can take hours or days.
Quick Command Reference
Windows Diagnostic Commands:
1 | # Check disk health |
BIOS Access Keys (by manufacturer):
- ASUS: Del or F2
- MSI: Del
- Gigabyte: Del or F2
- ASRock: Del or F2
- EVGA: Del
- General: Del, F2, F10, or F12 (try during boot)
Maintenance Schedule
Weekly:
- Quick temperature check (HWiNFO64, 5 minutes)
- Check for Windows updates
- Check disk space (> 20% free?)
Monthly:
- Full temperature stress test (Prime95 + FurMark, 15 minutes)
- Check SSD health (CrystalDiskInfo)
- Clean dust filters (if case has them)
- Update drivers (GPU, chipset)
Every 6 Months:
- Deep clean (compressed air, CPU cooler, GPU)
- Check thermal paste (if > 2 years old, consider replacing)
- Check all cables (loose connections?)
- Update BIOS (only if needed, see Q9)
Yearly:
- Full system backup
- Check PSU fan (abnormal noise?)
- Consider component upgrades (if needed)
π― Master Checklist: Print this page, check off items as you build/troubleshoot. This cheat sheet covers 90% of PC building and troubleshooting scenarios. Keep it handy!
Series Finale Summary
What Have You Learned?
Part 1: CPU & Computing Core
- β Data unit conversion (Bit/Byte/KB/GB/TB)
- β CPU brand comparison (Intel vs AMD)
- β 32/64-bit system differences
- β Server CPU characteristics (Xeon/EPYC)
Part 2: Memory & Cache
- β Memory working principles
- β DDR generation evolution (DDR2-DDR5)
- β Dual-channel technology and benchmarks
- β CPU three-level cache architecture
- β Memory troubleshooting
Part 3: Storage Systems
- β HDD vs SSD deep comparison
- β SSD interfaces and protocols (SATA/NVMe)
- β NAND types (SLC/MLC/TLC/QLC)
- β SSD optimization techniques (4K align, TRIM, OP)
- β RAID array configuration
Part 4: Motherboard & GPU
- β Motherboard interface details (PCIe/USB/M.2)
- β GPU working principles (GPU parallel computing)
- β Integrated vs discrete graphics
- β VRM power modules
- β BIOS optimization settings
Part 5: Network, Power & Practice
- β NIC types and selection
- β PSU power calculation
- β Cooling system configuration
- β Complete build process
- β 30+ troubleshooting cases
Memory Cheat (Series Grand Summary)
Computer fundamentals five parts through, from CPU to practical use;
Data units 1024, manufacturer disks 1000 calculate;
Intel single-core AMD multi-core, server look at core count;
Memory dual-channel bandwidth doubles, cache three-level accelerates CPU;
HDD slow but large SSD fast, TLC NAND lifespan long;
4K align TRIM enable, OP reserve maintain performance;
Motherboard interfaces each role, GPU parallel beats serial;
PSU power multiply 1.3, Gold efficiency most economical;
Black screen wipe RAM, blue screen check drivers;
Five parts complete become expert, build maintain don't need help!
π Congratulations on completing the Computer Fundamentals Deep Dive Series all 5 parts!
You now possess:
- β Hardware purchasing ability: Know which parameters matter, won't get scammed
- β Performance optimization strategy: Dual-channel, XMP, 4K alignment etc.
- β Troubleshooting capability: From symptoms locate specific hardware
- β Build & maintenance skills: Can independently complete building and routine maintenance
Next steps: 1. Build a PC hands-on (theory + practice = true mastery) 2. Help friends troubleshoot (best learning method) 3. Follow new technology (DDR5, PCIe 5.0, WiFi 7)
Series complete, thank you for joining! π
Any questions welcome in comments!
- Post titleοΌComputer Fundamentals (5): Network, Power & Practical Troubleshooting - Ultimate Guide from Hardware to Diagnostics
- Post authorοΌChen Kai
- Create timeοΌ2023-03-18 00:00:00
- Post linkοΌhttps://www.chenk.top/en/computer-fundamentals-5-network-power/
- Copyright NoticeοΌAll articles in this blog are licensed under BY-NC-SA unless stating additionally.