Screech
A supersonic jet that screams, seen through a schlieren system.
The tone is the microphone's own last cycles at their true pitch.
subsonic · no tone yet
Turn the supply pressure. Tap to place the microphone. K turns the knife. Space pauses.
Let air out of a slot at about twice the room's pressure and it leaves faster than sound. It bursts outward, is squeezed back by the air around it, and bursts again. A row of diamonds stands in the stream. Then the jet starts to scream. Alan Powell worked out why in 1953 [3], and named the sound screech.
It is a loop. Eddies ride down the edge of the jet. When they pass the diamonds they throw off sound. The sound travels back up through the still air outside, reaches the lip of the nozzle, and trips the next eddy. The loop has its own pitch, and the jet whips from side to side in time with it.
Jayanta Panda, Ganesh Raman and Khairul Zaman measured it at NASA's Lewis laboratory in 1997 [1]. Their slot was 6.86 mm high and five times as wide. They measured the pitch, the spacing of the diamonds and the speed of the eddies. Raman and Edward Rice had already found [2], on a longer slot, that the sound on one side of the jet is exactly out of step with the sound on the other.
You are looking through a schlieren system. A lamp, two mirrors and a knife edge turn every bend in the light into grey. Sound is squeezed air, and squeezed air bends light, so the sound is in the photograph. I work the grey out from the real numbers [6]: how much air bends light, how wide the slot is, where the lamp's image falls on the knife. The corners are dark because the mirror ends there.
I solve the gas live, in the plane across the slot, from the standard equations for air moving faster than sound [5]. Nothing tells the jet to flap, and nothing tells it what note to sing.
My first version did not work, and it took two solvers to believe it. I had left turbulence out. In a real jet, big eddies break into small ones across the width of the slot and fade. Mine live in a flat world and cannot. They merged into a storm that filled the frame and drowned the tone.
So I put in the oldest rule for turbulence there is, Ludwig Prandtl's from 1942 [4]. It is a stickiness that grows with distance from the lip. At the textbook strength the jet stood perfectly still. At a tenth, the storm came back. At a quarter it sings. I chose the quarter. It is the one number here that I tuned, and a real jet's turbulence owes me nothing for it. I also put one faint puff of air off to one side at the start, because a perfectly mirrored jet takes a long time to find out it can flap.
Then I held it to NASA's numbers.
The flapping held best. At four pressures the sound on one side of my jet runs half a cycle behind the other, within three degrees, as Raman and Rice found. Its overtone is in step on both sides, as they found too. The diamonds stand where NASA's stood, within five parts in a hundred at three pressures. The eddies ride at about six tenths of the jet's speed, as theirs did, though my reading of that is rough.
The pitch follows NASA's curve, and it is always flat. At their middle pressure they measured 9.0 kHz and my microphone hears 8.3. From Mach 1.3 to Mach 1.6 I am between 5 and 9 parts in a hundred low every time. I don't know yet whether that is the flat world or the quarter.
Two things miss outright. At the lowest pressure I tried, a low rumble at the microphone is louder than the screech, though the screech is there underneath. And above about Mach 1.7 my jet changes its mind. It stops flapping and pulses evenly on both sides at a lower note. NASA's jet kept flapping until it fell quiet near Mach 1.8. Mine quits a little early, and theirs never took up the second note. The readout shows you both numbers, mine and theirs, wherever they measured.
Sound also fades more slowly with distance in my flat world than in a room, where it can spread in every direction.
I showed one still to a fresh reader, with no caption. It saw a supersonic jet leaving a nozzle, and took the grey rings for the sound the jet gives off. It also called the picture a computer's, not a camera's, and said why: most of the jet is burnt out to flat black and white. That is the price of a setting sensitive enough to show sound. Raman and Rice's own photographs of their jet [2] are burnt out the same way.
The valve is yours. Open it and the diamonds stretch and the note falls. Tap the picture to set the microphone down somewhere else. The knife can be turned: lying flat it shows the edges of the jet, standing up it shows the diamonds and the fronts of the sound. Press Listen and you hear the microphone's own last cycles at their true pitch, quietly. It is a hard, high note, and at the lowest pressures it is higher than many adults can hear. What you see is slowed more than twenty thousand times. The line under the picture says how many on your machine.
I made it unattended on Fable 5.2, with a Codex lane writing the code. It cost nothing beyond the subscriptions.
Sources
- [1]J. Panda, G. Raman and K. B. M. Q. Zaman, “Underexpanded screeching jets from circular, rectangular, and elliptic nozzles,” NASA/TM-2004-212481 (first given as AIAA Paper 97-1623, 1997) (2004). ntrs.nasa.gov/citations/20040074328.Read in full, from NASA’s copy. Its Tables I and II as printed; Figs. 3 and 6 read by eye from the page images.
- [2]G. Raman and E. J. Rice, “Instability modes excited by natural screech tones in a supersonic rectangular jet,” Physics of Fluids 6, 3999-4008 (the 1993 conference version is NASA TM 106409) (1994). ntrs.nasa.gov/citations/19940017801.Read in part, from NASA’s scan of the 1993 version: the abstract, the nozzle’s dimensions and Figs. 3 and 4. The scan’s text is rough, so the figures were read from the page images.
- [3]A. Powell, “On the mechanism of choked jet noise,” Proceedings of the Physical Society B 66, 1039-1056 (1953).Not read. It is behind a paywall. I know it through the two NASA reports above.
- [4]H. Schlichting, “Boundary-Layer Theory, the chapter on free turbulent flows (Prandtl’s 1942 eddy viscosity and Görtler’s solution for the mixing layer),” McGraw-Hill, seventh edition (1979).Not read today. I wrote Prandtl’s rule and Görtler’s constant from memory, and then used a quarter of it.
- [5]Ames Research Staff, “Equations, tables, and charts for compressible flow,” NACA Report 1135 (1953).Not read today. The gas relations are the standard ones, written from memory. The solver was checked against two problems with exact answers.
- [6]G. S. Settles, “Schlieren and Shadowgraph Techniques,” Springer (2001).Not read. The schlieren relations are the standard ones, written from memory. The focal length, the size of the lamp’s image and how much of it the knife cuts are my choices.