Positive air pressure means your case takes in more air than it pushes out, so almost every gap in the chassis leaks air outward instead of sucking dust inward. It is the single most effective way to keep dust off your GPU, radiator fins, and motherboard over a build’s lifetime. The catch: getting it wrong in either direction costs you. Too much intake with no exhaust creates dead hot zones around the GPU and VRMs; too much exhaust quietly flips you to negative pressure and turns your PC into a dust filter in reverse. This guide walks through the exact fan configurations, BIOS curve settings, and verification steps to lock in positive pressure without sacrificing thermals, plus how to tune it for your hardware tier.
The priority order is simple: first, get the intake-to-exhaust fan ratio right; second, tune fan curves so intake fans stay slightly ahead of exhaust under load; third, filter every intake and leave exhausts unfiltered; fourth, verify with real measurements instead of guessing. Thermals still matter more than dust theory — a clean PC that thermal throttles is worse than a slightly dusty one that runs fast.
As an Amazon Associate we earn from qualifying purchases at no extra cost to you.
Our Top Picks
CORSAIR 7000D Airflow Full-Tower ATX PC Case7000 Series21.65 x 9.76 x 23.62 inches41.23 poundsCheck on Amazon
Corsair 5000D Airflow Mid-Tower Case20.5" x 9.6" x 20.5"20.47 x 9.65 x 20.47 inches27.34 poundsCheck on Amazon
Top 3 picks at a glance
Quick Recommended Settings
| Setting | Recommended value | Why |
|---|---|---|
| Fan layout ratio | 3 intake : 2 exhaust (or 2:1 on smaller cases) | Guarantees more cubic feet per minute (CFM) enters than leaves, even after filter drag |
| BIOS fan mode | PWM (4-pin DC off) for all chassis fans | Allows precise curve control per header; DC mode runs hotter and noisier at low RPM |
| Intake fan curve | 40% at 40°C, 70% at 60°C, 100% at 75°C (source: CPU package temp) | Keeps intake slightly ahead of exhaust as heat rises under load |
| Exhaust fan curve | Same source, but 10–15% lower duty at every point (e.g., 55% where intake is 70%) | Preserves the pressure differential at every load level, not just idle |
| CPU fan header source | CPU package temperature (not socket/external sensor) | Fastest, most relevant response for tower coolers and AIO pumps |
| Hysteresis / response time | 3–5 seconds ramp-up, 5–10 seconds ramp-down | Prevents fan “sawtooth” oscillation when temps flicker |
| Intake filters | Dust filters on all intakes; none on exhausts | Filters drop effective intake CFM 10–30%; compensate with a higher intake ratio |
| AIO radiator position | Front-mounted as intake, fans in pull or push-pull | Feeds the loop cooler air; rear/top exhausts (1–2 fans) remove the remainder |
| Verification method | Incense/smoke test at all seams, or a differential pressure gauge at a side vent | Confirms airflow direction empirically; CFM math alone lies |
| Zero-RPM / fan stop | Enable for intake fans only; disable for exhaust | At idle, exhausts running alone would flip you negative |
Setting-by-Setting
Fan layout: 3 intake, 2 exhaust
Set front and bottom mounts as intake, rear and top-rear as exhaust. Most mid-towers (e.g., a standard ATX mid-tower with 3 front, 1 rear, 2 top mounts) support exactly this. Use front: 3 × 120mm intake, rear: 1 × 120mm exhaust, top-rear: 1 × 120mm exhaust, leaving the top-front mount empty — a top-front exhaust steals air before it reaches the GPU.
The trade-off: top exhaust slots are useful heat escapes, but every top fan pulls against intake. If your case has three top mounts, run at most one, positioned rear-most. If your GPU exceeds ~300W, consider bottom intakes instead of adding exhaust.
BIOS fan headers: PWM mode with matched curves
Enter BIOS (usually Del or F2 during POST), navigate to Hardware Monitor / Q-Fan Control / Smart Fan (names vary: ASUS calls it Q-Fan, MSI Smart Fan Mode, Gigabyte Smart Fan 6, ASRock FAN-Tastic Tuning). Set every chassis header to PWM Mode, not DC/Auto.
Set intake headers to the curve above sourced from CPU temperature. Set exhaust headers to the same source but shift every duty-cycle point down 10–15%. This is the core mechanical trick: pressure balance must hold under load, not just at idle, and offset curves keep intake CFM ahead of exhaust CFM at every temperature point.
The trade-off: exhausts running 10–15% slower raise internal case temps by a degree or two compared to a balanced setup. That is the cost of positive pressure, and it is worth it. If CPU temps climb past your comfort threshold, raise both curves together rather than raising exhaust alone — keep the offset intact.
Fan curve source and hysteresis
In the same BIOS page, set the temperature source for chassis fans to CPU Package (AMD) or CPU (Intel) rather than “Motherboard” or a socket thermistor — package temp responds within seconds to load spikes. Set hysteresis/response to the slowest setting that still feels responsive: 3–5s up, 8–10s down on ASUS “Fan Step Up/Down Time”; equivalent fields exist on all four major board vendors.
Trade-off: slow ramp-down means fans keep spinning after a load burst ends. That is acceptable; oscillating fans that speed up and stall repeatedly are far more annoying and wear out hub bearings faster.
Dust filters: intake only
Install magnetic or slide-in filters on every intake; remove or leave off all exhaust filters (some cases ship with a top filter — remove it if the top mount is unused or exhaust). Filters cost meaningful airflow: a typical fine-mesh filter cuts effective CFM by 10–30%. This is precisely why the intake ratio must exceed 1:1 on paper — filters eat the margin.
Trade-off: filtered intakes need cleaning every 4–8 weeks in dusty environments. Set a recurring reminder; a clogged intake filter can silently flip you to negative pressure.
Zero-RPM behavior
Enable fan-stop (ASUS: “Stop” below a temperature threshold; MSI: “Fan Stop”; Gigabyte: “Fan Stop” toggle per curve) for intake fans only, typically below 40°C. Disable it on exhausts, or set exhaust minimum duty to 20–30%.
Why: at idle with everything stopped, hot air stagnates. With exhausts creeping at low RPM and intakes stopped, you have accidental negative pressure at your desk’s noise floor. A 20% exhaust minimum with stopped intakes is fine because the intake fans’ full-open position leaks, but the safest idle config is intakes at a 20% minimum too. Trade-off: a faint idle noise floor; on most quality PWM fans at 20% this is under 20 dBA.
By Hardware Tier
Low-end (entry CPU, 65–120W GPU, stock cooler)
A 2:1 intake-to-exhaust ratio is sufficient: 2 × 120mm front intake, 1 × 120mm rear exhaust. Skip the top exhaust entirely. Run a single flat-ish curve: 40% below 50°C, 100% above 75°C with a 10% exhaust offset. Stock coolers dump heat into the case, so source chassis fans from CPU package temp. Verification is easy — this configuration should show outward flow at every seam.
Mid-range (mainstream CPU with tower cooler, 200–300W GPU)
Use the full 3:2 layout above. Add a bottom intake if your case has mounts and the PSU is not shrouded-blocking them — bottom intake feeds the GPU directly and is the highest-value fan in most builds. Curves as listed in the quick table. If the GPU is the dominant heat source, source your front-intake curve from GPU temperature instead of CPU (available in most modern UEFIs under the fan header’s “temperature source” dropdown; fan-control utilities like Fan Control by Rem0o can do this if your BIOS cannot).
High-end (hot CPU, 320W+ GPU, 360mm AIO)
Mount the 360mm radiator as front intake with fans in pull configuration (pull performs nearly as well as push-pull through a radiator and is quieter). That gives 3 radiator intake fans plus optional bottom intakes against 1 rear + 1 top-rear exhaust. Warning: radiator intake pre-heats CPU coolant by roughly the GPU’s thermal contribution — if your CPU is temperature-sensitive under all-core loads, the alternative is radiator as top exhaust, but then you must add 2–3 bottom intakes to preserve positive pressure, and accept that the loop inherits case air. Between the two, front-intake radiators consistently deliver better CPU coolant temps in dust-managed builds, and positive pressure keeps the radiator fins clean — which matters more long-term than the single-digit coolant delta.
Common Mistakes
Trusting CFM spec sheets
What causes it: static pressure, mounting restrictions, and grilles make real-world airflow nothing like the boxed number; two “60 CFM” fans through different openings move very different air.
How to check: light an incense stick and hold it near every seam, PCIe slot covers, and top vents while the PC runs a sustained load. Smoke drifting inward anywhere means negative pressure at that leak.
What to do: slow your exhaust curve another 10% and re-test; if a specific vent still ingests (common on bottom-of-case slots), that intake ratio needs one more fan.
How to undo: restore your saved BIOS profile — export it first (ASUS: profile save slots in the EZ Flash area; most vendors allow exporting a profile to USB).
Filtered exhausts
What causes it: case manufacturers ship filters on top mounts, and builders leave them in.
How to check: inspect top and rear mounts for mesh or slide-in filters.
What to do: remove them; exhausts do not need filtration because air exits.
How to undo: clip them back on if you later flip a fan’s direction.
One curve for all fans
What causes it: syncing every header to a single curve accidentally equalizes intake and exhaust duty cycles, and with a filter on intake side, exhaust wins — negative pressure under load.
How to check: open the BIOS hardware monitor under a CPU/GPU stress load and compare intake vs. exhaust RPM; exhausts should read visibly lower.
What to do: apply the 10–15% exhaust offset described above.
How to undo: delete the offset points or reload the default profile.
Chasing dust with compressed air only
What causes it: treating symptoms — dust settling on components indicates pressure is wrong or filters are saturated.
How to check: if dust accumulates inside the case on board surfaces and fan hubs within a couple of months, air is entering through unfiltered paths.
What to do: re-run the smoke test, clean/replace intake filters, and re-check the exhaust offset.
How to undo: none needed — this fix is permanent once pressure is corrected.
FAQ
How do I know if my build actually has positive pressure?
Do the smoke test: run a sustained GPU or CPU load, hold incense smoke at the rear vent, top vents, and unfiltered seams. Outward drift everywhere confirms positive pressure. For a quantitative check, cheap differential pressure gauges (manometers) readable in hundredths of an inch of water column will show a small positive reading through any side-panel vent. Rough math — sum intake fan CFM minus ~20% filter losses versus exhaust CFM — gets you close but is not a substitute for the smoke test.
Is positive pressure always better than negative?
For dust control, yes, and it is not close — negative pressure pulls air (and dust) through every unfiltered gap: PCIe slot covers, expansion brackets, and cable pass-throughs. The thermal argument for negative pressure (hot air rises, exhaust it aggressively) is real but small in modern cases with well-placed intakes. A slight positive bias captures essentially all of the dust benefit at a thermal cost of roughly a degree or two.
Do I need to match fan brands or models?
No. Match airflow role, not brand. What matters is total intake CFM (after filter losses) exceeding total exhaust CFM, and curves that keep intake duty slightly higher. Mixing static-pressure-optimized fans on a radiator with high-airflow fans as case intakes is completely fine and often ideal.
What about positive pressure with an AIO — do I need special settings?
Yes, two adjustments. First, radiator fans count toward whichever side they’re mounted on: a front-intake radiator’s three fans are intakes, so size your exhausts assuming 3:2 or worse. Second, give radiator fans their own header curve sourced from CPU coolant temperature (usually labeled “CPU” when an AIO pump uses the CPU_FAN header) rather than copying your chassis-fan curve — radiator fans need higher minimum duty (40–60%) to push air through the fin stack, and pump speed should stay fixed at 80–100% rather than following a curve.
