Acoustic Horn Design – The Easy Way (Ath4)

Like this? Not good.
 

Attachments

  • cone-abec.png
    cone-abec.png
    86.2 KB · Views: 503
  • cone-pmap.png
    cone-pmap.png
    33.9 KB · Views: 509
  • cone-spl.png
    cone-spl.png
    22.6 KB · Views: 496
  • cone-z.png
    cone-z.png
    14.8 KB · Views: 500
I've attached the file - thanks.

mabat, this is a new 64bit machine.

It works on my system, but here's some possibilities:

1) do you have gmsh-4.3.dll in your path? I have mine in the same directory as ath4.exe. When I deleted the dll I didn't get the same error as you, but close.

2) In your config file, please comment out the line "Out.DestDir"

The reason that you want to comment out that line is twofold. First, if you have a "D" drive, your user may not have permissions to write to the folder. Second, if you omit the folder, ath4 will create the folder for you.

Having said all that, I was able to run it fine on my side by commenting out that line.
 
It might still be some unhandled exception in the progam. You can also try this: http://at-horns.eu/release/ath4_4.3.5.zip

This still isn't working. Just to be clear that I'm doing this right I took a screen shot of my ath dir. For the shortcut right click desktop-create shortcut-search "cmd.exe"-right click properties-start in "D:\ath_release_4.3.4"

Open the new shortcut " ath4.exe examples\small.cfg"
 

Attachments

  • Capture.PNG
    Capture.PNG
    42.8 KB · Views: 437
Previously you attached a file "small.txt" - so you changed the file extension or what else is happening? It may be that the program can't open the file (i.e. the filename doesn't match) but I would excpect different message then. Hmm...this is strange. Hasn't the file been renamed by the text editor to "small.cfg.txt" or something?
 
So I finally had an hour to play around with ATH4 today, trying to design a waveguide to mate with a 6.5-7" woofer around 1.6-2kHz somewhere.

I ended up turning the Depth.ConicSectionPart almost all the way to the lowest setting, leading to a very gradual change in slope almost all the way from the throat to the exit.

This lead to a profile that looks quite beamy/narrow when looking at the profile.
10-proto.PNG

but according to the BEM, it is not:

10-proto-polar.png
10-proto-polar-curves.PNG


Will be interesting to compare a 3D printed version to the BEM once I settle on a design.

Anyone who has some favorite combo of parameters they want to share?

Again, thanks to mabat for this excellent piece of software.

Here are the settings used to obtain the above polars:


Code:
ThroatDiameter = 16                 ; [mm]
  ThroatAngle = 0                       ; [deg]
  Coverage_Horizontal = 90             ; [deg]
  Coverage_Vertical = 90                ; [deg]
  Depth = 94                          ; [mm]
  Depth.ConicSectionPart = 0.12        ; 0.1-0.9
  Shape = raw                      ; raw | raw2rect
  Shape.FixedPart = 0.2                 ; 0.2
  Shape.CornerRadius = 38.0             ; [mm]
  SEExp = 2                     ; superellipse exponent
 
Fredrik, there will be a substantial mismatch of curvatures when joined such close to the throat. That's because I forgot to take a curvature of the hyperbola into account - now it's allways connected to a clothoid that starts from zero curvature. It's OK for junctions more into the waveguide but in this case it gets much more evident. I would recommend to wait until I fix this. Or, you can try this... It will "only" create a little more diffraction in the throat than necessary I guess. (And who knows - it may be better after all :D )

Or you could just prolong the conic section part a bit..
 
Last edited:
Fredrik, there will be a substantial mismatch of curvatures when joined such close to the throat. That's because I forgot to take a curvature of the hyperbola into account - now it's allways connected to a clothoid that starts from zero curvature. It's OK for junctions more into the waveguide but in this case it gets much more evident. I would recommend to wait until I fix this. Or, you can try this... It will "only" create a little more diffraction in the throat than necessary I guess. (And who knows - it may be better after all :D )

Okey, but how substantial can it be? Looks pretty smooth both in CAD and BEM :D
 
Also don't forget that this is still only for an infinite baffle. In reality the baffle with all its edges might make a big difference (for the worse). But I would say you have quite a good starting point.

- Haven't you tried making it a bit shallower? For such a small woofer it seems pretty deep...
 
Last edited:
For sure, but first I will be interested in mounting the WG to a large baffle to mimic the infinite baffle.

Then if the correlation is good, it would be tempting to measure them in a more normal narrow baffle and try to recreate this in BEM. When you have something that matches there, the system could be optimized together.. baffle, edges and WG as a system.