Fifteen Degrees I Finally Watched the Air Give Up

Smoke-Streak Wind Tunnel Flow Visualization

There is a number I have known since ground school that I had never once actually seen. Roughly fifteen degrees — the angle of attack where a typical wing stops flying and starts falling. I’ve felt it through a control yoke as a shudder and a sink. I’ve drawn the lift curve on whiteboards, that clean rising line and the sudden cliff at the top. But the air that does the deciding has always been invisible, and this afternoon I decided I was tired of taking it on faith.

So I built a box to catch the wind and make it confess.

An open-circuit desktop wind tunnel is embarrassingly simple in principle. Fan at one end pulls air through, honeycomb straightens it, a contraction cone squeezes and calms it, and in the clear-walled test section you put your model and your smoke. My honeycomb is four hundred drinking straws bundled into a length of ducting — the canonical cheapskate solution, and it works because a honeycomb kills swirl and cross-flow, the lateral wandering of the air, provided the cells are long enough. The rule I found and obeyed: cell length six to eight times cell diameter. Shorter and the straws don’t straighten; they just divide the mess into neater parcels of mess.

The contraction cone is where I got the first lesson, and it was a lesson in humility. My initial instinct — the model-aircraft builder’s instinct — was more fan. Bigger motor, more wind, more drama. Completely wrong. The contraction ratio, the area you squeeze the flow down by, is what actually smooths the turbulence, and it does so roughly in proportion to itself: a 9-to-1 contraction doesn’t just speed the air ninefold, it flattens the incoming eddies by about that factor too. A gale through a straight box gives you shredded, useless smoke. A gentle few metres per second through a proper cone gives you glass. You are, counterintuitively, trying to run the thing slow.

That slowness has a name in the physics: low Reynolds number. Push the speed and the flow in the test section itself goes turbulent, and your crisp white lines dissolve into milk. I spent a good half hour learning to want less.

The smoke was the hard part, obviously

The smoke was always going to be the hard part. My first attempt was incense, because it was in the drawer. Terrible idea — incense smoke is hot, so it’s buoyant, so it climbs out of your flow field and tells you about convection instead of aerodynamics. The fix is a cool oil mist: I ended up dripping fog-fluid glycol onto a nichrome wire and passing a current to boil off a sheet of evenly-spaced streamers. This is the classic “smoke-wire” trick, and it’s essentially unchanged from Frank Brown’s tunnels at MIT in the 1930s. There is something pleasing about re-enacting an eighty-year-old technique with parts from a vape shop.

A pedantic note I’ll insist on, because I’m writing it down and want it right: what the smoke shows you is a streakline, not a streamline. A streakline is every particle that has passed through one point — the smear a wand leaves. A streamline is the instantaneous tangent to the velocity field. In steady flow they’re identical, which is why the textbook pictures lie by omission. The instant my wing stalled and the wake went unsteady, that equivalence broke, and the beautiful lines stopped being literally true even as they became most interesting to watch.

And watch it I did. Cranking the angle of attack up degree by degree, the streaks hugged the upper surface, hugged it, hugged it — and then somewhere past fifteen the flow detached and a fat, lazy, recirculating bubble bloomed off the top of the wing. The same number. The exact number I’ve felt in my hands and never trusted. The air agreed with the whiteboard, and I may have said something out loud.

What I did not get was a photograph worth keeping. This is where the day soured. Making the flow visible to my eye is one problem; making it visible to a sensor is a meaner one. You need a thin sheet of light — a slit, or a laser line — raking across the smoke perpendicular to the lens, a dead-black background, and a shutter fast enough to freeze the streak but patient enough to gather its faint glow. My flat overhead work light washed everything to grey. Every frame I pulled tonight is a muddy suggestion of the thing I actually saw.

Which is a strange place to land: I finally saw the air do the one thing I’ve been inferring for twenty years, watched it separate at precisely the angle I’d have bet my licence on — and I have nothing to show anyone but a grey smudge and my word. Back when I taught a foam wing to draw the ground, the whole point was that invisible data became a document I could argue with. Tonight the air confessed and I forgot to bring a recorder that worked. The lighting rig is a photography problem, not an aerodynamics one, which means it’s mine to solve and I have no excuse. The smoke will keep. So will the wing.