The deep-mechanics chapter: case pressure as a flow budget, fan curves vs restrictions, good and bad turbulence, starving bottom intakes, hot-air reuse, and why chambers are a patch on a shape problem.
We covered the story: why the box fights the air, and what happens when you shape a case around the physics instead. But that article kept a promise to stay readable, which meant several mechanisms got a paragraph where they deserved a chapter.
This is that chapter, the deep dive for the people who read "bottom intakes starve" and immediately asked by how much, and why, and what do I do about it? Four mechanisms, in order: case pressure, fan pressure, turbulence, and the two ways air betrays you inside a box: starvation and recirculation. Everything here applies to any case you own, including ours.
Fair warning: this is the nerdiest question in the book so far. That's the point.
"Positive pressure" sounds like your case is inflated. It isn't. The pressures involved are so small you could never feel them. Case pressure is really a flow budget: how much air your intake fans push in versus how much your exhaust fans pull out.
Push in more than you pull out, and the surplus has to leave somehow, so air seeps outward through every crack, slot, and unfiltered gap. That's positive pressure, and its superpower is dust control: if all the deliberate intakes are filtered, and every accidental gap is an exit rather than an entrance, dust only enters through filters.
Pull out more than you push in, and the deficit gets made up by air sneaking inward through those same unfiltered cracks. That's negative pressure, the vacuum-cleaner configuration. Slightly better raw airflow in some builds; visibly dustier internals within months, because every PCI slot gap becomes an unfiltered intake.
How much does it move temperatures? Honestly: a little. Controlled tests show pressure balance shifting component temps by small single digits at most. It's the dust story, compounding over months into clogged filters and insulated heatsinks, that makes pressure worth planning. Set it slightly positive, filter the intakes, and move on.
Scale check, because "pressure" oversells it: a case fan develops a few tens of pascals at most. Atmospheric pressure is ~101,000 Pa, so your "pressurized" case sits at roughly +0.01% or less. Nothing structural is happening; the only physics in play is that air always flows from higher to lower pressure through every available opening. Positive/negative pressure is simply choosing the direction of that seepage, and therefore the route your dust takes.
Sources (the pressure playbook, and a real test)
Two fans with identical size and RPM can behave completely differently once something stands in their way. That's the split the market labels "airflow-optimised" versus "pressure-optimised," and it's not marketing; it's blade geometry.
An airflow fan is built to move maximum volume through open space: fine in free air, but put a dust filter, a dense radiator, or a mesh panel against it and its delivery collapses. A static-pressure fan has blades shaped to keep pushing against resistance: it sacrifices a little free-air volume to hold its output when something fights back.
The placement rule follows directly:
| Position | Resistance | Fan type |
|---|---|---|
| Open rear/top exhaust | Minimal | Airflow-optimised |
| Filtered intake | Moderate | Static pressure (or a strong all-rounder) |
| Radiator / heatsink | High | Static pressure, always |
| Dense mesh front | Moderate–high | Static pressure preferred |
If you only remember one line: free air rewards airflow blades; anything you can't see through rewards pressure blades.
The engineering picture: every fan has a curve (delivered airflow falling as back-pressure rises), and every case path has an impedance curve (pressure required rising with flow). Your real airflow is where the two curves cross, the operating point. Restrictions steepen the system curve, sliding the operating point down the fan's curve; pressure-optimised blades hold more flow at that crossing. This is also why manufacturer CFM specs (measured at zero resistance) overstate what any fan delivers installed, and why two fans in series raise available pressure while two in parallel raise volume.
Sources (from the fan makers themselves)
Air can move through your case two ways: in ordered, roughly parallel paths, or as a chaotic tangle of swirls that goes everywhere and therefore nowhere. The second is turbulence, and the surprising truth is that you want some of it, in exactly one place.
Think of cooling soup. Blow gently across the surface and only the top layer cools; stir it, and hot liquid keeps arriving at the surface to dump its heat. Right at a heatsink's fins, that's what micro-scale turbulence does: it stirs the thin blanket of stagnant hot air (the boundary layer) clinging to the metal, constantly replacing it with cooler air. Fin-level mixing is a feature. Heatsink designers deliberately provoke it.
Everywhere else, turbulence is pure loss. Bulk swirls in the middle of the case are fan energy converted into air going in circles, plus the acoustic signature you know as whoosh. Air that's tumbling isn't travelling; a case full of turbulent churn can have impressive fan specs and mediocre through-flow. The design goal is boring to state and hard to build: smooth, direct bulk transport from intake to exhaust, with the chaos saved for the last millimetres at the fins.
Sharp 90° corners, abrupt obstacles, and fans blasting into flat obstructions are the classic bulk-turbulence generators, which is one more way the rectangle's geometry, not its materials, sets its thermal ceiling.
For the "well actually" row: flow regime is characterised by the Reynolds number, and virtually everything in a PC case is turbulent to some degree; true laminar flow is not on the menu at these scales and speeds. The practical distinction is between organised turbulent through-flow and disorganised recirculating churn. Heat transfer at surfaces improves with local mixing (the boundary-layer effect); system throughput suffers as bulk flow loses momentum to eddies. Both statements are true at once, at different length scales, which is why "is turbulence good or bad?" has no one-word answer.
Sources (the two faces of chaos)
The airflow article gave this one paragraph and a sink. It deserves the full plumbing inspection, because bottom intake is the most popular "good idea in theory" in current case design.
The theory is sound: the floor is unused area, GPUs love direct feed from below, and floor dust filters are easy. The practice fails on one number: the gap under the case. Those bottom fans can only inhale what fits through the slot between the case floor and your desk, typically a couple of centimetres tall, minus whatever the feet design blocks, minus everything carpet swallows.
Recall the sink: drain a full basin and the flow is steady; drain a puddle and it sputters. Bottom fans on a short-footed case are draining a puddle, and when you mount two or three of them, they're competing to drain the same puddle. Each added fan raises the demanded flow through the same fixed slot, so the air under the case has to move faster, which costs more pressure, which slides every fan down its curve (section 3). Three fans can end up delivering little more than two, while sounding like three.
Diagnose it in thirty seconds: run a load, then slide a book under one edge to widen the gap. If temps move, your floor fans were starving. Fixes, in order of dignity: taller feet or a stand, fewer bottom fans doing less work, relocating intake duty to surfaces that breathe open room air, and never, ever, carpet.
The budget arithmetic: three 120 mm fans at moderate speed demand on the order of 30–35 L of air per second, delivered horizontally through the floor gap's cross-section (perimeter length × gap height, then reduced for feet and filter mesh). Force the same litres through a smaller area and the required air velocity rises proportionally; higher velocity through a restriction costs pressure roughly with the square of velocity; the pressure bill is paid straight out of every fan's delivery. Halve the gap and you don't halve the airflow, but you can lose a third of it while adding the noise of all three fans working hard against a wall.
Starvation is air failing to arrive. The second betrayal is worse: air that arrived, collected heat, and then didn't leave.
Hot-air reuse is the vertical version. In the standard tower layout, the GPU, the biggest heat source in the box, exhausts a large share of its heat straight upward, directly into the intake of the CPU cooler sitting above it. Your CPU spends its gaming sessions breathing the GPU's used air, which is why CPU temps often track GPU load in a layout problem masquerading as a cooling problem.
Recirculation is the horizontal version: corners and dead zones where the through-flow never reaches, so the same pocket of air loops in place, picking up heat lap after lap. The rectangle's sharp corners are natural recirculation pens: flow separates at the corner, an eddy forms, and whatever heat wanders in has no bus out. You can find these zones with an incense stick and thirty seconds: where smoke lingers and curls instead of streaming, air is going in circles.
The industry's structural patch is the chamber: wall off the PSU and drives in a basement, sometimes wall the motherboard from the radiators, and manage each sub-box separately. It genuinely helps; isolation is real engineering. But notice, as we did in question one, what chambers concede: that air in one open rectangular volume cannot be adequately controlled, so it must be subdivided into smaller rectangles. The patch works precisely because the shape doesn't.
Sources (reuse and the chamber confession, from the industry itself)
Everything above compresses into a checklist that works on whatever you own today:
None of this requires a new case. It's the same physics every enclosure obeys, including ours.
By now you can predict our answer, because it's the same one from every question in this book: we chose to fight at the geometry layer, where these problems are set, rather than at the patch layer, where they're managed.
The angled intake path exists to make the bulk flow short, straight, and low-resistance: organised transport, minimal corner-eddy territory, no intake asked to drink through a floor-gap straw. Direct feeds aim room air at the heat sources so the CPU isn't queued behind the GPU's exhaust. That addresses the reuse problem by routing, not by adding a chamber wall to a shape that generates the problem. Filtered intakes with a positive budget handle the dust math like any well-planned case.
And the honest closer this book owes you every time: the Octa V1 doesn't get to skip any physics in this article. Its fans have curves, its filters have impedance, its corners, fewer and blunter, still exist. We just picked the variable everyone else had left on the table: the shape the air was asked to survive. Whether that bet pays off is a data question, and you already know how to read the data.
For dust, essentially yes (with filtered intakes). For raw temperatures, the differences are small either way. Don't chase degrees in the pressure balance; chase them in the flow path. Slightly positive is the sane default.
Slightly more intake capacity than exhaust (e.g., three in / two out), accounting for the fact that filtered intakes deliver less than their spec. Ratios are a starting point; the smoke test and your temps are the referee.
No. Fans sharing one restricted inlet compete (section 5), fans fighting the path add churn rather than flow, and every fan adds noise. Diminishing returns arrive fast after a coherent in-and-out path exists. Placement beats count.
Not meaningfully with consumer tools; it's hundredths of a percent of atmospheric. But you can observe its direction: hold a tissue strip at an unfiltered crack. Blown outward means positive, sucked inward means negative. The direction is all that matters.
You can't, and you shouldn't want to; everything at case scale is turbulent. The goal is organised through-flow in the bulk and vigorous mixing at the fins, which is exactly what a short, straight, unobstructed path gives you for free.
Pressure is a flow budget that decides which way your cracks leak: set it slightly positive and filtered, and it's solved. Fan labels are about resistance: pressure blades where you can't see through, airflow blades where you can. Turbulence is two different phenomena wearing one name: keep the chaos at the fins, keep the bulk flow organised. And the two great betrayals, starved intakes and recirculating heat, are both, at root, geometry problems: air asked to squeeze through gaps or navigate corners it was never going to manage politely.
That closes the deep-mechanics layer of this book's thermal chapters: the story (airflow), the epistemology (reading tests), the materials, the coolers, and now the fluid dynamics. Same physics for every case ever made. The only question any manufacturer gets to answer differently is what shape they ask the air to survive.
Explore the Unknown.
// Explore the Unknown
© 2026 Olicus Technology & Gaming AB. All rights reserved.
Explorers of the Unknown