Dual Fan Cooling Box: How Airflow and Heat Load Really Work
This page explains the physics behind an open-source dual fan cooling box that can be downloaded and 3D printed by yourself. It is written for engineers and makers who want to know whether a printed enclosure will actually move heat, and where the design stops working. After reading, you can size fans, plan ducting and judge when a printed box is the wrong tool.

In this article
- 1
- 2
- 3
- 4
- 5
- 6
- 7
- 8
What a dual fan cooling box actually does
A dual fan cooling box is a printed enclosure that holds two fans and channels their output through a defined path. The fans do not cool anything by themselves. They move air, and the moving air carries heat away from a surface. If the air short-circuits back to the intake, you get noise and almost no cooling.
The open-source version of this box is popular because the geometry is simple. A rectangular shell, two fan cutouts, a louver or grille, and a lid. Anyone with a 260 × 260 mm printer bed can produce it. The hard part is not printing it. The hard part is predicting whether the finished box will remove the watts you need removed.
Air is a weak coolant by volume. It carries roughly 1.2 kg per cubic meter at room temperature, and its specific heat is about 1,005 J/kg·K. That means a small temperature rise in the exhaust is the only lever you have. To dump 100 W, you need either a lot of air or a large temperature delta.
So the design question becomes concrete. How many cubic feet per minute can the fans push through this restriction, and how much temperature rise will the exhaust show? Everything else, the color, the louver shape, the screw bosses, follows from those two numbers.
- 1Fans move airHeat removal happens at the heat source surface, not inside the box.
- 2Restriction sets flowGrilles, filters and bends cut the free-air rating, often by half.
- 3Exhaust temperature is your gaugeA 10 °C rise over ambient is a useful first-order check.
Series or parallel: the one rule that decides your layout
Two identical fans can be stacked in series or mounted side by side in parallel. The choice is not cosmetic, and it is the most common mistake in printed cooling boxes. Series doubles the static pressure at the same flow. Parallel doubles the flow at the same pressure.
If your box is open, with a short path and a large exit, pressure is not the limiting factor. Parallel wins. You get roughly twice the volumetric flow, so roughly twice the heat removal for the same temperature rise.
If your box pulls through a filter, a narrow duct, or a dense louver stack, the system curve is steep. Series wins. Each fan adds pressure, which pushes more air through a restriction that would otherwise choke a single fan.
Two fans in series with a wide-open inlet are a waste. You pay for a second fan, add noise, and gain almost nothing. Two fans in parallel behind a clogged filter are also a waste, because the restriction limits both of them. Match the layout to the impedance, not to the picture in your head.
- 1Open path, low impedanceRun the fans in parallel.
- 2Filter, duct or tight louversStack the fans in series.
- 3Mixed caseAdd pressure first, then add flow if the exhaust is still hot.
Estimating the heat load before you print
Start with the watts. A small server board may dissipate 25 W to 60 W. A bench power supply under load can reach 150 W. A 3D printer mainboard with stepper drivers is usually under 20 W. Write the number down before you open the slicer.
Then apply the airflow equation. Heat removed equals mass flow times specific heat times temperature rise. In practical units, roughly 1 CFM of air removes about 1.8 W for every 1 °C of temperature rise. So 30 CFM with a 10 °C rise handles about 540 W. That sounds generous until you subtract real losses.
Real losses are large. A printed grille with 40 percent open area can cut flow by a third. A 90-degree bend costs more. A dust filter can halve the flow within a week of use. Derate the free-air fan rating by 40 to 60 percent for a typical printed box.
Now compare. A 60 W load with 15 CFM of real flow needs only a 2.2 °C rise. That is easy. A 300 W load with the same 15 CFM needs an 11 °C rise, which pushes exhaust temperature well above what most electronics tolerate in a closed room.
If the numbers say you need a bigger fan than the box can hold, the box is the wrong answer. Move to a metal chassis with a larger fan, or split the load across two enclosures.
- 1Write the watts firstNo airflow estimate is meaningful without a heat load number.
- 2Derate by 40-60 percentPrinted grilles, bends and filters all cost flow.
- 3Check the riseMore than 15 °C over ambient usually means undersized airflow.
Material and wall choices for a printed enclosure
The enclosure itself sits in the exhaust stream, so it sees warm air, not the heat source temperature. PLA softens near 60 °C. PETG holds to about 75 °C. ABS and ASA hold to about 95 °C. PC and PA are higher still.
For most electronics cooling, PETG is the sensible default. It prints easily, tolerates a warm exhaust, and resists the small knocks a bench enclosure takes. Use ABS or ASA if the exhaust regularly exceeds 60 °C.
Wall thickness matters less than stiffness. A 2.4 mm wall with three perimeters at a 0.4 mm nozzle is usually enough. Add ribs or a slight curvature to the large flat panels. Flat panels on a 200 mm span will drum and radiate noise.
Fan mounting is where prints fail. Fan screws bite into plastic threads that strip after two or three removals. Model through-holes and use brass heat-set inserts, or design a clamp ring with M4 screws and nuts.
Vibration isolation is worth the effort. A thin TPU gasket between the fan and the printed face cuts structure-borne noise noticeably. Hard-mounting a 3,000 RPM fan to a thin printed wall turns the whole box into a speaker.
- 1PETG defaultGood to about 75 °C and easy to print.
- 2ABS or ASA for hot exhaustUse when exhaust regularly passes 60 °C.
- 3Heat-set insertsPrinted threads strip after a few fan swaps.
Where a printed dual fan box stops working
Printed cooling boxes work well for low to moderate heat loads, clean air, and a stationary bench setting. A 20 W to 80 W board in a room at 25 °C is a comfortable fit. So is a small power supply that only runs intermittently.
They work less well when the ambient is already hot. If the room sits at 40 °C, an exhaust at 52 °C gives you only 12 °C of headroom, and the components inside are running near their limits. Air cooling cannot beat ambient, no matter how many fans you add.
They fail when the air is dirty. Filters clog, and a clogged filter turns a parallel setup into a stalled one. If the environment has oil mist, metal dust or fibers, plan a cleanable metal filter and inspect it on a schedule.
They also fail when the part needs tight temperature control. A printed box with two fans gives you air movement, not a setpoint. If you need the surface held to ±2 °C, you need a controller, a sensor and a fan that accepts PWM or voltage control.
Finally, consider service access. A box that needs eight screws removed to reach the filter will not get serviced. Design a hinged or sliding lid, or at least a tool-less filter tray.
- 1Air cooling cannot beat ambientA 40 °C room caps what any fan box can do.
- 2Dirty air kills parallel layoutsOne clogged filter stalls both fans.
- 3No setpointAir movement is not temperature control.
Ducting, louvers and the recirculation trap
The single biggest performance loss in a printed cooling box is recirculation. Warm exhaust leaves the box, curls around, and gets pulled back into the intake. The fans run, the noise is there, and the intake air is already warm.
Separate the intake and exhaust by at least the fan diameter, and ideally more. If the box must exhaust and intake on the same face, add a divider plate that extends outward at least 30 mm past the face.
Louvers look tidy but cost flow. A stack of angled slats with 40 percent open area behaves like a partial blockage. Straight-through honeycomb or a simple wire grille with 70 percent open area is better if you care about CFM more than appearance.
Duct length adds resistance roughly with the square of velocity. Keep runs short and keep the cross-section larger than the fan hub area. A duct that necks down to 60 percent of fan area will choke the flow and raise noise.
If you need directed cooling, put the nozzle at the heat source, not at the box exit. Blowing a narrow jet at a hot component is far more effective than blowing into the general volume.
- 1Separate intake and exhaustAt least one fan diameter apart.
- 2Prefer open grilles70 percent open area beats decorative louvers.
- 3Aim the jetDirect air at the hot part, not at the room.
Parallel vs series: pick by system impedance
Read the impedance of your box first. The fan layout follows from it.
| Condition | Fan layout | What you gain | Watch out for |
|---|---|---|---|
| Open inlet, short exit | Parallel | About double the CFM | Intake recirculation |
| Filter or mesh screen | Series | Higher static pressure | Filter clogs, flow drops |
| Long bend or narrow duct | Series | Pressure to overcome loss | Noise at high RPM |
| Large plenum, low restriction | Parallel | More air over the part | Uneven flow across face |
| Sealed box, single vent | Series | Air actually exits | Back pressure stalls fans |
| Two different fan models | Neither | Nothing predictable | Beat frequencies, noise |
Printed enclosure materials compared
Pick the cheapest material that survives your exhaust temperature.
| Material | Useful to | Print difficulty | Best for |
|---|---|---|---|
| PLA | About 60 °C | Easy | Bench boxes, cool exhaust |
| PETG | About 75 °C | Easy | General electronics cooling |
| ABS / ASA | About 95 °C | Moderate | Warm exhaust, enclosure needed |
| PC | About 120 °C | Hard | Hot exhaust, stiff panels |
| PA (nylon) | About 120 °C | Hard | Tough parts, vibration |
| TPU | About 80 °C | Moderate | Gaskets and feet only |
The short verdict
If your heat load is under about 80 W in clean room air, print the box in PETG with two fans in parallel and an open grille. If you are pulling through a filter or a long duct, stack the fans in series instead. If the load is above roughly 300 W, or the room runs hot, skip the printed box and move to a metal chassis with a larger fan.
Questions engineers ask next
Can I use two different fan models in one box?
You can, but the result is hard to predict. Two fans with different pressure-flow curves will find an operating point somewhere between their individual curves, and the stronger fan may back-drive the weaker one.
If you must mix them, put the higher-pressure fan downstream and keep the models close in rated speed. Otherwise run a matched pair.
How do I measure whether the box is working?
Put a thermocouple in the exhaust stream about 50 mm from the outlet and another in the room air. Run the load until temperatures stabilize, usually 20 to 30 minutes.
A rise of 5 to 10 °C over ambient means the airflow is matched to the load. A rise above 20 °C means you are short on flow or recirculating warm air.
Does a bigger fan always move more air?
At the same RPM, a larger fan moves more air at lower pressure. In a restrictive box, that lower pressure can actually reduce flow compared with a smaller, faster fan.
Match the fan to the system curve. A 120 mm fan at 1,200 RPM and a 92 mm fan at 2,200 RPM can end up close in a tight enclosure.
Should the fans push or pull?
Pulling air out of the box, with a filtered intake, keeps dust off the fan blades and spreads intake air across the components. Pushing in works too, but it concentrates flow on whatever sits directly in the path.
For a box with one hot component, push air toward that component and let it exit the far side.
Will a printed box hold up to continuous running?
PETG at 50 to 60 °C exhaust will hold for years in a dry room. PLA will creep and sag if the exhaust stays near 60 °C.
Add heat-set inserts at every screw joint. Creep shows up at fasteners first, and a stripped fan mount is the usual end of a printed box.
Can I control fan speed to reduce noise?
Yes. A PWM fan with a 25 kHz control signal lets you set speed from about 20 to 100 percent. Below 20 percent many fans stall or click.
Dropping from 100 percent to 60 percent typically cuts perceived noise by half while keeping most of the airflow, as long as the system impedance is low.
Need the metal version instead?
Send us your STEP file. We return a quotation and a free DFM analysis within 12 hours, with no minimum order quantity.
12-hour quoteNo minimum order100% inspection