AKABAK 3

I did not use the mesh file but baffle component to define the baffle.

Use the tag (23 in this case) to align the location of the diaphragm to mesh define baffle.

Hope this answer questions.
 

Attachments

  • Speaker Cabinet SP38 Mesh-Files - OB 2.png
    Speaker Cabinet SP38 Mesh-Files - OB 2.png
    14.4 KB · Views: 500
  • Speaker Cabinet SP38 Mesh-Files - OB 3.png
    Speaker Cabinet SP38 Mesh-Files - OB 3.png
    43.1 KB · Views: 507
  • Speaker Cabinet SP38 Mesh-Files - OB 4.png
    Speaker Cabinet SP38 Mesh-Files - OB 4.png
    4 KB · Views: 508
Is this general rule that alignment to mesh file boxes is made with pointing mesh tag? What point of mesh is actual alignment point? All meshes had 3 points.
What if mesh file box had no mesh element in the center of speaker installation hole, if it has just empty hole?
 
Last edited:
Use the tag number, which consists of multiple points (elements), to align the driver.

We use tag numbers (names) to select which parts of the mesh file is included in a specific subdomain.

You need to add a small disc in CAD drawing solely for driver alignment.
 

Attachments

  • Speaker Cabinet SP38 Mesh-Files - OB 5.png
    Speaker Cabinet SP38 Mesh-Files - OB 5.png
    11 KB · Views: 481
Last edited:
I get some result out of AKABAK with test OB speaker.
But result is not very reasonable.
As can be seen on image OB speaker missing dipole peak and dipole minimums in level lines.
Measuring points are 1 m from speaker.
What I had possibly made wrong?
Project file contains 3182 elements and full calculation time is over 1 hour on 3.7GHz 8 core CPU, calculate at your own risk.
 

Attachments

  • AKABAKresult1.PNG
    AKABAKresult1.PNG
    51.9 KB · Views: 461
  • test2b.zip
    40 KB · Views: 65
I find something.
On Log Calc in section Boundary Meshing is information after project test2b calculation:
Free Term calculation
Subdomain ElType Min Max Mean StdDev/1000
Sub1 Ext -0.028 32.515 0.736 1592.680
What probably is not good (specially StdDev/1000 1592.680), but I not find anywhere explanation what this "Free Term calculation" actually means. With example project this StdDev/1000 figure is less than 1
 
Last edited:
I had more questions.
When I look at example project "Speaker Cabinet SP38 Mesh-Files" I see rear and front diaphragm had offset 12 mm. In my understanding this can give wrong results. What can be the reason of diaphragm offset?

This is an "error". The project is updated as the attached. The baffle thickness is 10 mm. An offset of -10 mm is used to not taking the "tunnel" in the model.

I shall compare the AKABAK result to LEAP.
 

Attachments

  • Speaker Cabinet SP38 Mesh-Files - OB_2.zip
    60.8 KB · Views: 102
Did this mean that offset give smaller error than tunnel?
But in real life there is also existing "tunnel" and sides of diaphragm had zero offset.

Had you look at my OB project, why I am missing dipole peak and dipole minimums in level lines?
I made other mesh of open baffle speaker box without cutout and attached same speaker elements to it but get nearly same result. No dipole peak and dipole minimums in level lines.
 
Last edited:
I upload also pictures to see, what kind of problem I had.
OB is U-frame with 300 mm depth, 2 elements in it. I not using symmetry for calculation. Diaphragmas are standard, with magnets and detailed measurements (this take very long time to calculate)
Result is not very symmetrical and on level lines dipole peaks and minimums are not very accurate. If I use simple diaphragms, result is better but also not what I expect.
On images front side is down and with blue line.
 

Attachments

  • AKABAKresult3.PNG
    AKABAKresult3.PNG
    87.4 KB · Views: 409
  • AKABAKresult2.PNG
    AKABAKresult2.PNG
    83.9 KB · Views: 423
Keep the BEM elements less than ~5000. Otherwise, computational time is very long. Set the mesh length in Global. Use smaller length in critical elements. This is similar to mesh optimization in FEA. We need to control mesh size (density) wisely.

Your model is symmetry. Use it to save computing resources.

Your model is being simulated in my PC.
 
Last edited:
Yes, my test model has symmetry, but my target is to make fully asymmetrical OB frame rear part, so my test is symmetrical (results can be checked with other non-BEM software) but I make and calculate it as asymmetrical to take clear how it can/must be done.
 
Last edited:
More results.
I changed mesh of diaphragms to Delaunay, get element number down from 3100 to 2000. But also very different results compared to previous image what had Bifurcation mesh. On new image lower part had rear diaphragm offset of 2 mm, all other parameters are unchanged, this cause also very different result.
 

Attachments

  • AKABAKresult4.PNG
    AKABAKresult4.PNG
    106.2 KB · Views: 365
Last edited: