FINALLY, it's done! Believe it or not, this was the most exacting task in the whole endeavour - throat adapters for segmented horns (octagonal in this example).
It makes a smooth R-OSSE contour in each of the segments. The petals are just bended flat sheets.
It makes a smooth R-OSSE contour in each of the segments. The petals are just bended flat sheets.
Last edited:
Just to be clear, the DI knee that I was referring to is that in the system DI, not just the waveguide. Getting this knee right requires the right woofer, among other things.
They seem to be a natural way to get the horn design right. Maybe it's already been done.
Do they? I know the Fibonacci spiral (which is not really a fractal), which is a series of circular arcs connected together with sharp curvature discontinuities. I would expect it to behave not as nicely overall, but I may be wrong. Do you know of any other examples how could "Fractal and Fibonacci" be applied to a horn design?
Thanks! Wow, I didn't expect the wool could make such a big difference. That could mean that a lot of the non-flatness commonly observed is actually caused by the throat wavefront mismatch (which is almost always the case). But what baffles me is the difference below ~1.5 kHz. That's 17 dB attenuation at 1 kHz, compared to 8 dB at 10 kHz. How is that possible?
It's only a pity you haven't tried the rockwool with the DE111 also.
Last edited:
I think the measurements are fine otherwise, you can make them indoors... It doesn't require a much longer time window at this point, I think. Also, such small devices can be safely measured from 1m or maybe even less. You can gain several more milliseconds of reflection-free time by going closer. You certainly don't need to be at 2 m in this case.
Anyway, I think this is one of the most interesting findings.
Anyway, I think this is one of the most interesting findings.
Last edited:
Something to keep in mind… unless the rockwool is “teased out” (pulled apart to reduce its density), it will typically have quite a high acoustic flow resistivity which means sound is more likely to bounce off of it than get damped or absorbed. I’m not sure if this is relevant for the frequencies being studied though - I’d have to refresh my memory on the various charts available online.
No I don't. I wonder if when Webster wrote about horns 100yrs ago he thought of Fractals?
Having tried many densities of foam, I found that the higher the density the smoother the response, but the more loss was incurred. It becomes a tradeoff. I would use a less dense material since a 17 dB loss is a lot. Unworkable with a passive Xover, but maybe OK with active where the gain can be higher for the tweeter.
In a passive system, the tweeters high end becomes the limiting factor in system efficiency (due to the falloff at HFs from the Constant directivity waveguide.) Hence lowering the tweeters output causes the whole system response to be lower, requiring a pad on the woofer, which is not desirable for many reasons.
The fact is that some sound gets through the wool, and this sound seems to be free of the resonances seen when there's no wool.
- Remember that this was (intentionally) an extreme test with a very high loss, to really see the effect as clearly as possible. No one, I assume, would use such amount in a final product. But it shows really well, IMHO, that there are resonances caused by the throat, that are maybe more common than usually anticipated. Another argument for having a look into this.
Last edited:
I believe that this was a mistake and that it has another explanation, as there's "only" 8 dB loss at 10 kHz. At least I can't see how that would be possible...
Now for the petals... What sheet metal material and thickness would you choose to make them from, for a CLD structure?
I can easily let laser-cut stainless steel, aluminum or brass. Bending them to a template even with a DIY jig should not be a problem if thin enough. I would probably glue them together with an epoxy but I have no idea what sheet thickness would be the most appropriate. As the damping layer I would use some elastic glue, haven't really looked yet at what's available.
Last edited:
Tried playing around with ath and I seem to get stuck at "reading bem_mesh.geo" no matter which script (demos or snippets from here) I try. gmsh starts running in the background but doesn't seem to complete anything. Any ideas?
Info : Running 'C:\Users\asdasdasd\Desktop\ath\gmsh\gmsh.exe bem_mesh.geo -' [Gmsh 4.11.1, 1 node, max. 1 thread]
Info : Started on Wed Jun 14 20:27:02 2023
Info : Reading 'bem_mesh.geo'...
Last edited: