The name mabat gave to the throat plug with channels. External Shaping Plug (ESP).
He linked the application note in the name but it might not have been obvious.
http://www.at-horns.eu/release/ath-AP1.pdf
I've just printed a new one, where the rings have equal area, and played a bit with the exponent on each. Kind of trcky to make them have the values I want without the channels crossing over each other.
I haven't measured it yet, just simulated. The ESP sits on a horn that is about 45cm in diameter and doesn't have a foldback, hence the wavy response.
The newest one I have has the following parameters:
; --------------------------- ESP related code ----------------------------
; Equal Area Rings
; 4 rings: Pos0 = 0.446,0.632,0.774,0.894
; 3 rings: Pos0 = 0.500,0.707,0.866
SP50 = {
Dt = 50
At = 6.3
Ae = 30
L = 80
Pos0 = 0.446,0.632,0.774,0.894
ExpRate = 3,3,3,4,5
Sk = 1
CP1 = 0.2
CP2 = 0.7
;WT = 1
;EndAngle = 1
}
Source.Contours = ::esp SP50 ; source definition
; -------------------------------------------------------------------------
I haven't measured it yet, just simulated. The ESP sits on a horn that is about 45cm in diameter and doesn't have a foldback, hence the wavy response.
The newest one I have has the following parameters:
; --------------------------- ESP related code ----------------------------
; Equal Area Rings
; 4 rings: Pos0 = 0.446,0.632,0.774,0.894
; 3 rings: Pos0 = 0.500,0.707,0.866
SP50 = {
Dt = 50
At = 6.3
Ae = 30
L = 80
Pos0 = 0.446,0.632,0.774,0.894
ExpRate = 3,3,3,4,5
Sk = 1
CP1 = 0.2
CP2 = 0.7
;WT = 1
;EndAngle = 1
}
Source.Contours = ::esp SP50 ; source definition
; -------------------------------------------------------------------------
It has a weird rising spread after 15k that I can't seem to get rid of.
At the same time I'm thinking it's not the worst that can happen to a horn
Yes, it turns out it's not so straightforward to design a perfect conical horn for a spherical source. It still requires some experimentation. I've been quite busy lately so I can't pursue this further myself. In the meantime, I at least implemented variables into the script "language" so I'll be able to perform some pretty complex optimizations, including the ESP parameters. There's still a lot to be achieved, I think.
You GOGO Mabat
I've been considering migrating over to Comsol for that reason. It's faster since it meshes for each frequency, and you can do parameter optimization.
It has a built in CAD, so you can basically design the waveguide (solid and axisymmetic) and let the optimisation tweak the shape within your constraints.
It's quite a bit more difficult to set up, since it's a general physics solver, but it could be an interesting challenge.
I believe @kyleneuron has already gotten the grip of it by now
It is, but occasionally one has a friend in university that has access to it
I think it's not necessary. A strong argument would be the speed but as one iteration takes around a minute in ABEC to solve, I personally don't feel the need. Otherwise everything mentioned is possible with Ath/ABEC as well, if one is willing to create some automated framework around it (still for free). I'm only a bit hesitant to release what I have (using AutoIt and some yet undocumented Ath features), as it's not very user-friendly and I don't have the time for support.
Last edited:
Two examples of speakers using "this horn":
https://www.diyaudio.com/community/threads/hf10ak-12p80nd-custom-wg.360973/https://www.diyaudio.com/community/...-design-the-practical-way.354772/post-6214298
https://www.diyaudio.com/community/threads/hf10ak-12p80nd-custom-wg.360973/https://www.diyaudio.com/community/...-design-the-practical-way.354772/post-6214298
