Acoustic Horn Design – The Easy Way (Ath4)

Magnitude + phase:
1722445768284.png


Group delay:
1722445789953.png


Burst decay:
1722445909254.png
 
The effect of filling the long adapter with open-cell foam:

1722447645645.png


With foam:
foam.png

Without foam:
nofoam.png


Except the overall attenuation, there's almost no difference at high frequencies, which IMHO means that it must all happen inside the driver, even before it enters the throat of the horn. Then it just passes once through the waveguide.

Group delay:
1722449017276.png
 
Last edited:
Group delays zoomed (with and without the foam). There's actually something around 17 kHz that the foam seems to improve, but this is so high and basically out of band that this is of little interest.

1722449550730.png


No foam, excess group delay in ARTA (the black curve), i.e. the difference between the measured one and the one corresponding to the minimum phase calculated from the magnitude frequency response:

1722449997396.png


Actually, this could be a HOM.
 
Last edited:
Group delays zoomed (with and without the foam). There's actually something around 17 kHz that the foam seems to improve, but this is so high and basically out of band that this is of little interest.

View attachment 1339947

No foam, excess group delay in ARTA (the black curve), i.e. the difference between the measured one and the one corresponding to the minimum phase calculated from the magnitude frequency response:

View attachment 1339950

Actually, this could be a HOM.
My guess up where none of us can hear, is a serious resonance. It's a short enough wavelength that it could be a reflection. Did you stuff the entire throat with the reticulated foam?
 
It is just around these frequencies where the higher order modes of the source start to propagate: https://www.diyaudio.com/community/...-design-the-easy-way-ath4.338806/post-6677700

It could be one of those. Resonances within the fundamental mode apparently don't cause an excess group delay. The value of the delay in this case seems rather high to me, hard to say if it makes sense, I can't imagine the propagation anyway and I can't spend too much time analyzing this at the moment, but would like to get back to this some day, in a more systematic way, including an effect on directivity.

In this particular case (and probably in most cases with a 1" throat), it can be safely ignored, IMO. Put a piece of foam inside the waveguide and forget about it...
 
Last edited:
I'm glad it's clear now. Let us know how it goes. It shouldn't be too hard but it's not always easy either 🙂
Printed the missing parts and finished rough sanding the first horn. Here is a dry fit test. The jig is a lifesaver! Thanks Mabat.
 

Attachments

  • IMG_5046.jpg
    IMG_5046.jpg
    273.2 KB · Views: 182
  • IMG_5044.JPG
    IMG_5044.JPG
    758.6 KB · Views: 183
  • IMG_5043.jpg
    IMG_5043.jpg
    610.9 KB · Views: 184
It is just around these frequencies where the higher order modes of the source start to propagate: https://www.diyaudio.com/community/...-design-the-easy-way-ath4.338806/post-6677700

It could be one of those. Resonances within the fundamental mode apparently don't cause an excess group delay. The value of the delay in this case seems rather high to me, hard to say if it makes sense, I can't imagine the propagation anyway and I can't spend too much time analyzing this at the moment, but would like to get back to this some day, in a more systematic way, including an effect on directivity.

In this particular case (and probably in most cases with a 1" throat), it can be safely ignored, IMO. Put a piece of foam inside the waveguide and forget about it...
So if were a HOM, and you measured the on-axis response at different distances from the horn, you should be able to observe the peak "coming and going" ?
....or at least appearing at a distance, while being absent close to the horn...
 
Excuse my ignorance, I have a question. If one has access to big 3d printers, is it still not recommended to print these horn flares in one piece. Does it not print out properly? If it can be printed properly and the limitation is the print bed size hence the petal approach, would it be possible to get stl files for whole horn flares in cults?
 
Excuse my ignorance, I have a question. If one has access to big 3d printers, is it still not recommended to print these horn flares in one piece. Does it not print out properly? If it can be printed properly and the limitation is the print bed size hence the petal approach, would it be possible to get stl files for whole horn flares in cults?
I'd argue not worth it. All those side flanges that allows gluing together also raise the stiffness and therefore the resonance of the horn. That's worth a lot in terms of how a high frequency horn sounds in the real world.
 
  • Like
Reactions: aragorus and kodomo
Excuse my ignorance, I have a question. If one has access to big 3d printers, is it still not recommended to print these horn flares in one piece. Does it not print out properly? If it can be printed properly and the limitation is the print bed size hence the petal approach, would it be possible to get stl files for whole horn flares in cults?
I'm doing that currently with some larger horns. I think its a good idea to if you have the means to do it right. Rigidity is the primary concern, so you'll want to modify the meshes to include some degree of ribbing & gussets exterior to the horn walls. There's a lot of literature in the plastic injection molding world on how this is best accomplished. Worth looking at how other OEMs have done it for inspiration i.e. the ME464 below.

1722867380931.png