On Fri, Jan 3, 2014 at 10:00 AM, Fons Adriaensen <fons@linuxaudio.org> wrote:

On Fri, Jan 03, 2014 at 11:03:59PM +1100, Patrick Shirkey wrote:

> Theoretically, how big can this scale?
>
> http://www.youtube.com/watch?feature=player_embedded&v=odJxJRAxdFU

Not so much before it becomes really dangerous and unpractical.

What supports those objects is the pressure differences generated
by the standing waves. Now 120 dB SPL is 20 Pa, or 2 kg / m^2, or
2 milligrams per square millimeter. Which suggests they are already
using more than 120 dB SPL.

These are nice back-of-the-envelope calculations, but I think there's a little more explanation needed (not sure, but bear with me).

Why do we see this technique applied with high frequencies? At the node locations, there is the highest peak spatial derivative of the pressure while at the same time, the volume velocity of air goes to 0.

Vice-versa at the anti-nodes.

It may not require so much power as the frequency becomes very high (although the physical scale between nodes also diminishes making it impractical for "large" objects), if the variable of interest is not the pressure difference itself, but the magnitude of its spatial derivative.

Other thoughts: the apparent acoustic power only needs to be large in the region of standing waves. You may not be exposed to dangerous SPL's when you get far enough away from the focus.

I have been contemplating some designs on acoustic levitation, but I haven't risen to the task of figuring how much power is involved. This discussion is helpful to me (and fun!)

Chuck