How We Test Our Components
Our complete, transparent process for ensuring every used part is reliable, stable, and ready for years of use.
Buying used PC components should feel just as safe as buying new ones. At PC Power Up, every item we sell goes through a strict, multi‑stage testing pipeline designed to catch faults, verify performance, and guarantee long‑term stability. This page explains exactly how we test each type of component, the tools we use, and the standards a part must meet before it reaches our store.
1. Intake, Cleaning & Visual Inspection
Every component begins with a careful intake process:
- Visual inspection for bent pins, damaged traces, corrosion, cracked solder joints, or missing components.
- Connector checks for worn PCIe fingers, damaged DIMM contacts, or loose capacitors.
- Cleaning using compressed air, soft brushes, and isopropyl alcohol where needed.
- Thermal interface removal on CPUs and GPUs, followed by fresh, high‑quality thermal paste during testing.
This stage filters out physically compromised items before any stress testing begins.
2. RAM Testing (MemTest86 + TestMem5)
Memory faults can be subtle, so we use two independent testing suites to eliminate false positives and catch intermittent errors.
What we run
- MemTest86 (full pass) — industry‑standard low‑level memory testing that checks every addressable cell.
- TestMem5 (custom configs) — more aggressive patterns that expose timing‑related instability.
What we look for
- Zero errors across all passes
- Stable operation at rated XMP/JEDEC speeds
- No thermal throttling on high‑density modules
When RAM fails
Any stick showing even one error is rejected. Memory instability can cause crashes, corrupted files, and unpredictable behaviour — so we only accept flawless modules.
3. GPU Testing (FurMark, Unigine Heaven, 3DMark, MATS/MODS)
GPUs are tested more thoroughly than any other component because they’re pushed hardest in real‑world use.
Stress & stability tests
- FurMark — extreme thermal load to verify cooling performance and stability under maximum heat.
- Unigine Heaven — real‑world 3D rendering to check for stuttering, artifacting, or VRAM issues.
- 3DMark benchmarks — performance validation to ensure the card scores within expected ranges for its model.
VRAM integrity tests
- MATS/MODS — low‑level NVIDIA memory diagnostics that detect VRAM faults, bit errors, and unstable memory channels.
What we look for
- No artifacting, flickering, or driver crashes
- Stable clocks without power limit oscillation
- Temperatures within safe operating ranges
- Benchmark scores consistent with healthy cards
When GPUs fail
Cards with VRAM errors, unstable boost clocks, or thermal runaway are rejected. If a card passes but runs hotter than expected, we re‑paste and re‑test.
4. CPU Testing (Prime95)
CPUs are extremely reliable, but we still run them through a full stability and thermal validation process.
What we run
- Prime95 (Small FFTs) — maximum heat and power load to test thermal behaviour.
- Prime95 (Blend) — memory + CPU stress to catch IMC‑related instability.
What we look for
- Stable operation with no rounding errors
- Temperatures within safe limits under sustained load
- No throttling on supported motherboards
- Correct boost behaviour for the CPU generation
When CPUs fail
Any CPU that errors, overheats, or shows inconsistent boost behaviour is rejected.
5. Performance Verification & Grading
Once a component passes all stress tests, we verify:
- Performance vs. expected baseline
- Thermal behaviour under load
- Noise levels (for GPUs)
Only components that meet our highest standards are listed for sale.
🛡️ Why Our Testing Matters
Used PC parts can be incredibly reliable — if they’re properly validated. Our testing pipeline ensures:
- Zero‑error RAM
- Stable, artifact‑free GPUs
- Cool, consistent CPUs
- Accurate performance matching new‑part expectations
This is how we confidently offer reliable used components at a fraction of the cost of new hardware.