[JohninCR]

Below is in response to JoshK's question about greater velocity at the edges of OBs.


Josh,

I don't recall the location, but the idea first presented itself when I was researching line arrays, so possibly through JBL's technical stuff regarding line arrays.  Someone put a bunch of woofers in an OB line array, and the wings were constructed with an edge shape similar to an airplane wing to address the issue.  I didn't recall that info until fairly recently, when I was trying to optimize a pair of small baffles for my FE108, which are useful down to about 200hz.  I wasn't happy with the imaging until I put 2" radius roundovers on the edge of the 10" wide baffle, and the imaging really came into focus, far better than any test baffle I've ever used.  Note that my experience with boxed speakers has been only subtle improvements with large roundovers, and this was far more significant.

I'm the type who needs to understand results in order to progress and improve designs.  Then I recalled reading that near the edges of an OB is the area of highest velocity, which I didn't really understand at the time.  Now I think I've got a handle on it.  As sound propagates, air molecules move back and forth (they don't flow like a river).  Olsen demonstrated the effects of edge diffraction, which showed a baffle edge shape of a 90 degree angle to be the worst shape.  I visualize it as the air molecules moving back and forth as the wave bends around a sharp corner causing turbulence, which disturbs the wave form.

Now let's throw into the mix than an area of higher pressure rushes in to fill a rarefaction.  This is what is happening at the edges of an OB.  As the higher pressure part of a wave spreads around the end of the baffle, it's counterpart (a rarefaction since it's directly out of phase) is bending around the front, causing the molecules to move faster.  Then as the front wave goes into rarefaction, they rush back the other direction.  Obviously going from +1 to -1 and back the other way will exaggerate the effect compared to just from +1 to 0 and back, like with a boxed speaker.

I hope this explanation makes sense.  If I could display my visualization with an animated graphic, it would make things much easier than trying to explain using words.

[Russell Dawkins]

If this is true, and logic suggests it is, then I can project that at some point in the evolution of the ideal baffle shape we will try something that:
1. looks like a multipointed star (or starfish)
2. is trimmed in fur on the edges.

One way of looking at acoustic diffraction is to see it as turbulence in negotiating a tight corner. Using a star shape would distribute the frequencies affected.

Using fur would almost totally eliminate turbulence around the edge. Why do you think they use fur on those zeppelin windshields for outdoor micing? For the same reason that fur is used around the edge of parka hoods - to cut down on wind turbulence or, at least to randomly distribute its effect.

Not sure what the WAF would be - guess it depends on the decor in the room!

It has always seemed to me that the worst baffle shapes (from the point of view of diffraction effect) are those where the edges are roughly equidistant from the driver center as in, for example, the Bastanis Apollo
http://www.baulsaudio.com/apollogermany.html

[JohninCR]

Russell,

I'll have to put the fur idea to the test, but with some polyfill batting instead of fur, although I'll be eyeing the kids' stuffed animals closely for those with long hair.   I'm just not sure if I can pull off tasteful for adults, so I'll probably stick to large roundovers.

Regarding baffle shape, the best proposal I've seen is a nautilus shell shape.  Then an exact starting frequency occurs at only 1 specific point on the baffle.  A star shape will have the same frequencies occur at many different points.  I'll just stick to elongated rectangles or trapezoids.  They're easier to construct and make look good, and the duplication of frequency points is at most 4 locations on the baffle. 

Also, typically we refer only to the specific frequency where the wave starts to wrap around the baffle, ie the baffle step, however, I think that's only the point at which the issue begins.  As far as actual diffraction caused by the shape of the edge, it should continue from that point on down although I'm sure there's some frequency dependence with the higher frequencies being affected more audibly.