Wind tunnel
A 41 cm square test section as long as the Wrights' box, a fan that draws air through a honeycomb, and two balances you calibrate yourself: measure a wing before you build it.
Build
Tier 4 · Industrial revolution
Transport
Moderate
About two weeks to build; allow a month to get the flow even, as the Wrights did
Crew 2
Home: In the air, Level 4 · Engine, level 3 of 6 in its track. Serves 1 step.
Every build on this site has hazards that injure or kill — quicklime, carbon monoxide, molten metal, collapsing ground. Read the safety notice before starting, and read this build's own failure modes below.
What it is and why this shape
The Wrights' tunnel data "far exceeded anything that had ever been available ... for at least another decade" (National Park Service), and it took a box, a fan and two home-made balances. This version keeps their size but puts the fan at the outlet: the Wrights put theirs at the entrance, which swept swirl from the fan blades down the tunnel and forced them to spend nearly a month on flow straighteners (NASA). A fan drawing air through a smooth bell mouth and a honeycomb gives an even flow from the start.
Open-return wind tunnel, fan downstream
Fig 1
TimberLoose fill / grainGlass
Lift and drag balances
Fig 2
TimberGravel / aggregate
Keyed parts
| 1 | Bell mouth | A rounded inlet, about one and a half times the section's width, so air enters without separating. |
| 2 | Honeycomb and screen | Tubes 1 cm × 8 cm packed across the inlet, then a fine screen: they straighten the flow and even out its speed. |
| 3 | Test section | 1.83 m long, 41 cm square inside, with a window and slots for the balance rod and pendulum. |
| 4 | Diffuser | Widens gently behind the test section, slowing the air before the fan. |
| 5 | Fan | At the outlet, drawing air through; guarded. |
| 6 | Balances | A steelyard under the floor for lift; a pendulum from the roof for drag. |
Specifications
| Test section | 41 × 41 cm inside, 1.83 m long | 16 in square and 6 ft long, as the Wrights' box. |
| Air speed | about 11 m/s (40 km/h) | The Wrights ran at about 25 mph. |
| Fan power | about 1 kW | The air carries about 130 W at 11 m/s; the rest goes in fan and duct losses. |
| Honeycomb | tubes about 1 cm across, 8 cm long | Packed across the whole inlet. |
| Model size | frontal area under about 5 % of the section | Bigger, and the walls change the result. |
| Speed check | flat square plate face-on, drag coefficient 1.28 | NASA's figure. |
Bill of materials
| Material | Quantity | Notes and substitutes |
|---|---|---|
| Boards or plywood | for a 1.83 m box, 41 cm square inside | Smooth inside; joints sealed. |
| Window | 1, about 50 × 30 cm | Glass or clear plastic, flush with the inside wall. |
| Sheet metal or thin board | for the bell mouth and diffuser | The bell mouth rounded, with no sharp edge. |
| Honeycomb | enough tubes to fill the inlet | Drinking straws or thin tubes, about 1 cm × 8 cm, packed tight. |
| Fine wire screen | 1 | Just downstream of the honeycomb. |
| Fan and drive | about 0.6 m fan, about 1 kW | Electric motor, engine or line shaft, with a guard. |
| Balance parts | — | Steel rod, a knife-edge pivot, a beam with a scale and a sliding weight; a pendulum arm, a pivot and a protractor. |
| Calibration weights | a set | Known masses from grams to about 1 kg; thread and a pulley. |
| Flat square plate | 1, 10 × 10 cm | For the speed check. |
| Wool tufts | a few | To see the flow. |
Tools: Saw, plane and drill · Tin snips · Square and level · Protractor · Scales for the calibration weights
Construction
Build the test section
Make a box 1.83 m long and 41 cm square inside, smooth inside with every joint sealed. Set a window flush in one side at the middle. Cut a narrow slot in the floor under the model position for the lift rod, and one in the roof for the drag pendulum.
Fit the bell mouth, honeycomb and screen
At the inlet, fit a rounded bell mouth flaring to about one and a half times the section's width: a sharp edge throws the flow off the walls. Just inside, pack tubes 1 cm × 8 cm across the whole opening, then a fine wire screen.
Fit the diffuser and fan
Behind the test section, a diffuser that widens gently — a few degrees each side — to the fan. Mount the fan at the outlet, drawing air through, with a guard over it, and drive it from a motor, engine or line shaft of about 1 kW.
Make the flow even
Tape short wool tufts across the test section and run the fan. Every tuft must stream straight back and steady. If some flutter or point off-line, look for leaks, a sharp edge at the inlet, or gaps in the honeycomb. The Wrights spent nearly a month on this (NASA).
Build and calibrate the lift balance
Under the floor, a beam on a knife edge; a vertical rod from the beam rises through the slot to hold the model; a sliding weight on the beam's other arm balances it. With the fan off, hang known weights on the rod and mark on the beam where the slider balances each: that scale reads lift directly.
Build and calibrate the drag balance
From a pivot above the roof, hang a pendulum arm through the slot to hold the model; drag swings it back. Calibrate it with the fan off: pull the model backwards with a thread over a pulley carrying known weights, and mark each on the protractor.
Measure the wind speed
Mount a 10 × 10 cm flat plate face-on to the wind and read its drag. Its drag coefficient is 1.28 (NASA), so the wind's pressure is drag ÷ (1.28 × 0.01 m²), and speed = √(2 × pressure ÷ 1.225). Repeat until readings agree.
Test wings
Mount a model wing — the Wrights used strips of thin steel — at angles from 0° to 20°, and read lift and drag at each. Lift coefficient = lift ÷ (wind pressure × wing area). Keep the model's frontal area under about 5 % of the section. Change one thing at a time — aspect ratio, camber, edge shape — and record everything.
Tolerances that matter
| Dimension | Target | Why it matters |
|---|---|---|
| Flow | every tuft straight and steady | Swirl or unevenness makes every reading wrong. |
| Balance calibration | rechecked each day | Pivots wear and stick; a calibration is only as old as its last check. |
| Repeat readings | within a few per cent | If they differ more, something is loose or leaking. |
How this fails
Fan at the inlet: swirl from the blades runs down the tunnel, as the Wrights found.
A sharp-edged inlet or leaking joints: uneven flow.
Balances read by formula instead of calibrated: friction and the rod's own drag go uncounted.
Models too large for the section: the walls change the flow round them.
An unguarded fan: it takes fingers and hair.
How you know it worked
VerificationTufts stream straight in the test section; both balances are calibrated with known weights; the plate test gives the same wind speed on repeated runs; and a standard wing's lift and drag repeat within a few per cent.
Scaling and variants
The power rises with the section's area times the cube of the speed: twice the speed takes about eight times the power. For the design of tunnels in more detail, see NASA's Beginner's Guide to Wind Tunnels.