In this thread, I'll post some sims of a Bang and Olufsen lens. These were licenses from Sausalito Audio Works, who continues to sell speakers that use a variation of the B&O lens.
Here's my original thread on the lens (it's closed now.) https://www.diyaudio.com/community/threads/cloning-a-3200-speaker-for-400.163015/
Here's a pic of the best measuring one that I've ever built. My version looks quite a bit different than the B&O original:
Details are buried in this thread: https://www.diyaudio.com/community/threads/improved-saw-lens.350889/
Here's my original thread on the lens (it's closed now.) https://www.diyaudio.com/community/threads/cloning-a-3200-speaker-for-400.163015/
Here's a pic of the best measuring one that I've ever built. My version looks quite a bit different than the B&O original:
Details are buried in this thread: https://www.diyaudio.com/community/threads/improved-saw-lens.350889/
Attached are four pics:
1) The first two pics are the lens that I made in Autodesk 123D. The reason you only see half is because I have symmetry turned on in abec.
2) The second two pics are the model in ABEC, after being meshed.
1) The first two pics are the lens that I made in Autodesk 123D. The reason you only see half is because I have symmetry turned on in abec.
2) The second two pics are the model in ABEC, after being meshed.
Attachments
Attached are three pics. They show the wavefronts produced the the SAW lens at 2000, 4000, and 8000Hz.
What I notice, is that the wavefront produced by the lens is mostly spherical (which is good) at 2000Hz and 4000Hz, but by 8000Hz, the wavefront is really random. That's because the aperture is too large. For instance, 8khz is just 1.6875" in diameter. (43mm.) Because the aperture is larger than the wavefront (because the "hat" on the lens is almost two inches away from the exit of the compression driver.)
What I notice, is that the wavefront produced by the lens is mostly spherical (which is good) at 2000Hz and 4000Hz, but by 8000Hz, the wavefront is really random. That's because the aperture is too large. For instance, 8khz is just 1.6875" in diameter. (43mm.) Because the aperture is larger than the wavefront (because the "hat" on the lens is almost two inches away from the exit of the compression driver.)
Attachments
OK, I warned you about the wavefront issues - here are the predicted polars. I believe I can improve on these.
One thing to add here: it's possible the vertical and horizontal are swapped. Because the simulated response shows that the beamwidth is vertically wide and horizontally narrow, when the shape of the lens should be producing the opposite.
I'll post my model; I may have swapped a set of values here. Or it could be something else causing it. Stay tuned.
One thing to add here: it's possible the vertical and horizontal are swapped. Because the simulated response shows that the beamwidth is vertically wide and horizontally narrow, when the shape of the lens should be producing the opposite.
I'll post my model; I may have swapped a set of values here. Or it could be something else causing it. Stay tuned.
Attachments
![2025-06-19 22_25_33-VacsViewer - (new) - [Converted from PM_SPL_Horizontal].png](/community/data/attachments/1382/1382552-7f1d7a25a999b9d0a71d2c69e730bd6e.jpg?hash=uy_QD9rrH-)
![2025-06-19 22_30_00-VacsViewer - (new) - [Converted from PM_SPL_Vertical].png](/community/data/attachments/1382/1382553-41bacbaabb6d1d203609fbb62b32eb70.jpg?hash=U36t61COfF)
Attached are all the files you need to sim this in ABEC. You only need the five files highlighted in yellow. (see pic.)
In addition, I've included two versions of the model from 123D (The one you want is the one with a 907pm timestamp), as well as a step file from gmsh.
Instructions on how to make your own models and sim them in ABEC are here:
In addition, I've included two versions of the model from 123D (The one you want is the one with a 907pm timestamp), as well as a step file from gmsh.
Instructions on how to make your own models and sim them in ABEC are here: