Acoustic Horn Design – The Easy Way (Ath4)

[augerpro]

close, 0.137 mm away. Overall diaphragm is 30.1mm and throat is 30.374mm. This is a Bliesma T25B BTW.

 

[mabat]

Is the surround convex or concave? It's not clear from the drawing.

 

[mabat]

- Does this resemble what you have?

STL file attached.

 

[augerpro]

Yes pretty much that. Except I can't see if you have small 0.5mm lip. Also the surround is concave.

 

[mabat]

I don't believe these sub-milimeter details are important for the analysis - this is all I can offer. The template of cfg attached.

 

[augerpro]

You have it in the appendix as part of the dome source: P5 to WG0. Functionally the same as my lip seal. I agree at 0.5mm it won't do much, but I want to experiment with it.

 

[mabat]

OK, go ahead

Source.Contours = {
zoff -0.5
point p1 3.5 0 2.5
point p2 0 13.05 0.5
point p3 -1 14.05 0.5
point p4 0 15.05 1.0

cpoint c1 -22.58 0
cpoint c2 0 14.05

arc p1 c1 p2 1.00
arc p2 c2 p3 0.75
arc p3 c2 p4 0.25
line p4 WG0 0
}

 

[augerpro]

Awesome, thank you! One more small request if I may LOL, could I get a version with OS contours so I can start comparing OS vs constant radius and can be confident I didn't screw something up copy/pasting?

 

[mabat]

In the meantime I ran the analysis - is this it?

- Ouch, same mistake again, I forgot to switch the velocity to axial - so this is wrong... OK, once again.

 

[mabat]

So this is the final CFG, with the 0.5 mm lip and corrected for the axial velocity -

 

[mabat]

Simulations for the axial dome motion - it's not much better

Raw horizontal polars (constant acceleration):

Vertical:

Attached also the normalized responses to 0 deg.
I'm really interested to see the measured data - how much are we off with this.

 

[augerpro]

I'll have to look at the results I posted but that looks about right.

 

[augerpro]

Here are my measured results, 0-60 degree in 10 degree steps, horizontally, followed my manufacturer's specs. What are the steps in ABEC?

 

[mabat]

You can set any step you need. This is for 0 - 60 / 10 for the raw polars ("SPL_H/V" charts). Polar maps are per 5 deg by default.

ABEC.SimType = 1 ; infinite baffle
ABEC.f1 = 1000 ; [Hz]
ABEC.f2 = 20000 ; [Hz]
ABEC.NumFrequencies = 20
ABEC.MeshFrequency = 1000 ; [Hz]
ABEC.Abscissa = 2 ; linear

ABEC.Polars.Dist = 1.5 ; mic distance - adjust this to your value [m]
ABEC.Polars.Step = 10 ; angular step [deg]
ABEC.Polars.Points = 7 ; # of polars
ABEC.Polars.Horizontal = 1
ABEC.Polars.Vertical = 1
ABEC.Polars.Diagonal = 0
ABEC.Polars.DiagonalInclination = 0.0
ABEC.Polars.PMapNorm = 10 ; polar map normalizing angle [deg]

I think the match is pretty good. There may still be differences due to the different radiating conditions. I work on finite baffles - that will be even closer.
BTW, this was calculated for 20 points only. Solved in 10 minutes.

 

[mabat]

Looking more closely on your geometry, it seems that the vertical profile arc touches the baffle at slightly higher angle. When I checked my radiuses they were 74.6 and 25.5 mm, horizontally and vertically.

Maybe this will be a bit closer to what you have:

CircArc.TermAngle = 0.8 + 7*sin(p)^2

Now the radiuses are 73.9 and 30.5 mm. Updated STL attached.

 

[augerpro]

Looking more closely on your geometry, it seems that the vertical profile arc touches the baffle at slightly higher angle. When I checked my radiuses they were 74.6 and 25.5 mm, horizontally and vertically.

Maybe this will be a bit closer to what you have:

CircArc.TermAngle = 0.8 + 7*sin(p)^2

Now the radiuses are 73.9 and 30.5 mm. Updated STL attached.

I'm not quite following, are the wall profile radiuses not 73.66 and 30.48 mm as I specified? Looking at the .cfg I see they are not even there, so some function with certain assumptions (tangent at baffle for example) must be calculating a best fit when given mouth size, throat size, depth, etc? Is there a way to direct code the radius?

 

[mabat]

If you think about it, it's not that easy - what's the radius on the diagonals and everywhere between? If you wanted to set the radii directly you would have to know this. I chose a different approach - you define the throat, the mouth outline (whatever that is) and the angle at which the arc should meet the baffle (=CircArc.TermAngle). The radii around the waveguide are simply calculated from that.

 

[augerpro]

Good point, I wonder how Fusion360 does it? I only draw the four "guide" arcs you see in my pic a few posts ago and the Loft function interpolates the rest. Must be some standard formula?

 

[mabat]

It can be anything, there even don't have to be circular arcs anywhere else. I don't know the loft algorithm.

In my approach there are simply circular arcs everywhere.

 

[jcarothers]

Forgive the naive question: what's the difference between constant velocity and constant acceleration simulation? When should one be used over the other?