Acoustic Horn Design – The Easy Way (Ath4)

[mabat]

Not quite the same but similar, this is "only" 810 x 567 x 340 mm:

I see a big hole on-axis around 7 kHz on this one, not sure it's real (may be a mesh size issue):

Otherwise it's pretty smooth and controlled, acoustic impedance of this throat arrangement being a bit untypical.

 

[lbstyling]

Not quite the same but similar, this is "only" 810 x 567 x 340 mm:

View attachment 1075368
View attachment 1075369 View attachment 1075371

I see a big hole on-axis around 7 kHz on this one, not sure it's real (may be a mesh size issue):
View attachment 1075372

Otherwise it's pretty smooth and controlled, acoustic impedance of this throat arrangement being a bit untypical.

What's the probability that the hole on axis is in simulation only?

This is getting close to what I'm looking for. I would be aiming for wider pattern (+/-30deg at -1db) otherwise it's excellent for my use (assuming the 7khz issue is an artifact.)

 

[mabat]

To verify that means going on the edge of possible using ABEC, the number of BEM elements rises quickly with such a large device. I don't have a very powerful machine...

- One can further reduce the number of mesh elements by making the rear surface flat (mathematical), i.e. smaller:
Mesh.RearShape = 2

 

[mabat]

The on-axis hole looks quite real (simulated on a little bit smaller version of the horn, 720 x 508 x 300mm) -

Normalized at 30deg:

My guess is that it would disappear with a smoother throat adapter.
Haven't tried the rings...

ABEC project attached. Someone might try to show a field observation.

 

[mabat]

This sounds like it would work with my bends as well.

Here's my new Fusion360 script. It uses a new input file format (*.asc), commented example attached.
It just imports sketches defined in the input file:

Then you can manually use the loft feature(s) or whatever:

ASC file format (ordinary text file):
The first number is always a command, the rest are its parameters (some of them may be optional).
Comment lines start with '#'.

 

[kipman725]

Thanks Mabat this looks great!

 

[mabat]

No problem.
Some random experimenting with the 2-profile geometry: 2" throat, 480 x 482 x 320 mm. ATH project attached.

 

[mabat]

Horizontals normalized at 10deg:

Vertically it's not so good

 

[mabat]

This would be the throat adapter -

 

[aragorus]

Does this throat make it a diffraction horn?

 

[mabat]

You mean where's the distinction what is a diffraction horn and what is not? I have no clue
It definitely reminds one but there's no sharp discontinuity.

 

[mabat]

Anyway, I think this could have a potential to make for a nice device overall. Worth pursuing further, IMO. Unfortunately an optimization of a full 3D mesh is very slow and it will take some time to get the most out of it. Perhaps a task for a collaborative effort?

 

[mabat]

What about laminating using this?

It's a paper 500 g/m^2, 0.8 x 1.2 m, 0.8 mm thick, 2 €/kg. Could be cut with scissors with no dust

 

[aragorus]

Anyway, I think this could have a potential to make for a nice device overall. Worth pursuing further, IMO. Unfortunately an optimization of a full 3D mesh is very slow and it will take some time to get the most out of it. Perhaps a task for a collaborative effort?

I'm going to give it a go

What about laminating using this?

View attachment 1075862

It's a paper 500 g/m^2, 800 x 1200 mm, 0.8 mm thick, 2 €/kg. Could be cut with scissors

That looks good!
In fusion you can do sheet metal folding. It can be used to calculate the "projected" flat piece needed.
I was thinking about kerf bending plywood, but this looks good too.

I tried a couple of simulations today with different mouth aspect ratios, but it quickly deteriorates.
Could be nice to use an ESP plug before the throat adapter. Unless the horn is only to be used until 10k

 

[mabat]

Maybe some horizontal vanes could be inserted inside the adapter (made a part of it) to improve the vertical dispersion on HF. Horizontally I don't think that's needed - in the above example it makes a 50mm throat radiate as if it was 30mm or even better. The same can be done with a 36mm throat and get even better horizontal polars than with a 25mm driver (at 14 kHz it would look as 10 kHz above).

 

[airvoid]

Carbon fiber prepreg using a polished alu mold would be the ultimate. 🔝

 

[mabat]

Yea but that mold would cost as a new car.

 

[airvoid]

Not quite.. I would say a couple of thousand €:s though, the machining is pretty simple. Plastic molds would be cheaper but not give the same surface quality and would only last for a a hundred or so cycles.

 

[mabat]

The point is that for this shape we don't need a mold, only some templates or jigs for bending the boards.

 

[weltersys]

Maybe some horizontal vanes could be inserted inside the adapter (made a part of it) to improve the vertical dispersion on HF. Horizontally I don't think that's needed - in the above example it makes a 50mm throat radiate as if it was 30mm or even better. The same can be done with a 36mm throat and get even better horizontal polars than with a 25mm driver (at 14 kHz it would look as 10 kHz above).

The Electro Voice "TransPlanar™" horns used horizontal vanes in their diffraction throats, dispersion was quite good in both planes.