I'd like to adapt ATH to my uses in my waveguide project, but it appears I'm limited in two ways (correct me if I am wrong): I cannot make an elliptical mouth, and I cannot have a horn profile of constant radius (section of a circle)? Would it be terribly difficult to implement these?
The order should not make a difference and I put it exactly where you had it in your example script and it worked first time. Perhaps just a gremlinFigured it out. The Source.Contours values needed to be in the Abec Project Setting area of the cfg. I assume sticking it at the end of that section is ok?
The elliptical mouth can be had by using a superformula (SF) and setting the morph to Raw.I'd like to adapt ATH to my uses in my waveguide project, but it appears I'm limited in two ways (correct me if I am wrong): I cannot make an elliptical mouth, and I cannot have a horn profile of constant radius (section of a circle)? Would it be terribly difficult to implement these?
This is a somewhat elliptical one I tried earlier
GCurve.SF = 1.09,0.6,4,4,5.2,1.87
superformula for a visual and for you to tweak if you want
The guiding curve for the throat can only be OS or Conical.
What you might be able to do is use the dome script to make a circular section but then the waveguide would be circular too.
You could also use Ath for the dome generation and use it with a mesh of your own waveguide, but that is much easier said than done.
The mouth can be virtually any shape you want, either by setting an explicit formula for the coverge angle (and/or perhaps some other parameters) or using a guiding curve. For an ellipical outline it's more straigtforward to use an ellipse (or a superellipse) as the guiding curve:I'd like to adapt ATH to my uses in my waveguide project, but it appears I'm limited in two ways (correct me if I am wrong): I cannot make an elliptical mouth, and I cannot have a horn profile of constant radius (section of a circle)? Would it be terribly difficult to implement these?
GCurve.Type = 1 ; 1=superellipse | 2=superformula
GCurve.Dist = 1.0
GCurve.SE.n = 2.0
GCurve.Width = the width you want ; [mm]
GCurve.AspectRatio = 0.7 ; adjust this to your needs
As for the radius as the waveguide profile this is not possible at the moment. It wouldn't be difficult to implement it as the third profile option, I guess, I might look at that. Is there any advantage in using a simple radius for the whole profile?
It really should make no difference and I tried the very file you attached and that gave me a dome source, so I'm little in doubt that was the cause. It is true however that if the syntax was wrong somehow, the item Source.Contours would be ignored which would lead to a default source instead - that would be the one you showed. That's why I wrote you should see a mentioning of reading the file in the output messages.Figured it out. The Source.Contours values needed to be in the Abec Project Setting area of the cfg. I assume sticking it at the end of that section is ok?
There's basically no syntax checking - what's not recognized is just ignored.
Last edited:
BTW, forget about guiding curves
It's much more flexible and convenient to use the free expressions, as introduced in 4.6 - that's maybe the biggest improvement. It probably wouldn't be possible to desing the Tritonia waveguide just by using a guiding curve.
It's much more flexible and convenient to use the free expressions, as introduced in 4.6 - that's maybe the biggest improvement. It probably wouldn't be possible to desing the Tritonia waveguide just by using a guiding curve.
Convenient if you can think in mathematical expressions I can't and I don't think I would be the only one If you have got any tips (besides what is written in the user guide) they would be helpful to better understand how this part works.BTW, forget about guiding curves
It's much more flexible and convenient to use the free expressions, as introduced in 4.6
That would be too complex to implement just for fun, I'm afraid. Maybe what I could do is to save each mesh part as a separate file (STL?) that could be edited manually outside the current processing chain. You could modify the waveguide surface mesh (add holes, etc.), save it back and run the simulation. But I would have to think this through, this is just an idea.
Definitely! Edit the whole stl/mesh and then remesh for new abec project. Maybe something along those lines? In combination with hornresp we get some crazy possibilities!
There would have to be a possibility of keeping the STL mesh intact where you don't touch it, so that the individual parts of the mesh still stick together seamlessly (i.e. without doing a complete re-meshing). I don't know if this is possible.
Well, at least the line segments on the edge would have to stay the same. Meshing of the interior of the surface could of course change. Can you edit and save an STL file and keep the edges intact? If you can, this could work.
Well, at least the line segments on the edge would have to stay the same. Meshing of the interior of the surface could of course change. Can you edit and save an STL file and keep the edges intact? If you can, this could work.
Last edited:
As for this kind of construction, you could then also add the midrange diaphragms inside their chambers manually and simulate the whole thing. I can imagine this but I'm terrible at CAD modeling so I have never actually tried to do something similar. It's definitely possible.
mabat> thanks for the ellipse script! Regarding the the profile radius, constant radius are easy to design and manufacture, which is why most of my waveguide designs have used them. I've got quite good results, so it would be nice to model something I already have measured to compare to the ABEC sim. After I'm satisfied everything is working correctly, I would move on to OS or conical contours to see if there is improvement, but with dome tweeters being such non-ideal sources the simple constant radius might be better for some tweeters.
OTOH if, as you talked about above, we could pull mesh surface in, maybe that would work. I experimented lasted night with saving just the horn surface as an .stl, then converting to mesh in Gmsh. Worked fine. I could pull it into ABEC from there, but I have no idea how to setup ABEC, especially adding a driver element, so that is what i really like about ATH, or at least the demos files where it is already setup correctly, I just tweak the values I need changed.
OTOH if, as you talked about above, we could pull mesh surface in, maybe that would work. I experimented lasted night with saving just the horn surface as an .stl, then converting to mesh in Gmsh. Worked fine. I could pull it into ABEC from there, but I have no idea how to setup ABEC, especially adding a driver element, so that is what i really like about ATH, or at least the demos files where it is already setup correctly, I just tweak the values I need changed.
Last edited:
The spherical profile won't be difficult to implement.
- Today I experimented a bit with such small shallow waveguides for dome tweeters and this ATH profile seems quite promising: profile parameters
It translates simply as follows:
Geometry.Definition = 1
Throat.Profile = 2
Throat.Diameter = 36 ; [mm]
Throat.Angle = 0 ; [deg]
Coverage.Angle = 0
Length = 30 ; [mm]
Term.s = 2.8
Term.n = 1.33
Term.q = 0.996
- Today I experimented a bit with such small shallow waveguides for dome tweeters and this ATH profile seems quite promising: profile parameters
It translates simply as follows:
Geometry.Definition = 1
Throat.Profile = 2
Throat.Diameter = 36 ; [mm]
Throat.Angle = 0 ; [deg]
Coverage.Angle = 0
Length = 30 ; [mm]
Term.s = 2.8
Term.n = 1.33
Term.q = 0.996
Attachments
In these cases it might make more sense to step the frequency points linearly:
ABEC.Abscissa = 2 ; 1=log | 2=linear
Could you explain that?
- Home
- Loudspeakers
- Multi-Way
- Acoustic Horn Design – The Easy Way (Ath4)