A couple more videos with standing waves in a sphere. The spherical standing waves are fascinating! 😍Read the video description and look for another links on the same problem in comments under the video
The point source on the wall
The point source in the centre
I'm building a 6" diameter sphere. Can Akabak3 model damping in the enclosure? What density of damping would smooth the response from 300Hz and up in a 6" diameter sphere? Simply packing the sphere tightly with fiberglass would avoid making tapered cylindrical enclosures for the back of each of 20 drivers and packing them with damping. That site isn't secure and Windows throws warnings about installing it. I could go build a virtual machine I guess.I'm not sure what you want to hear 🙂. I think that you will figure out how to build a spherical box without my help, especialy if you understand some theoretical basics. Nevertheless, I don't build spherical boxes because the spherical shape is not very pleasing to the eye (at least for me).
I was not too lazy to spend half an hour yesterday to simulate a spherical box SPL in a fixed microphone point, excited by a point source in different places (in the center and on the wall). As the theory predicts, certain modes are supressed when the point source is placed in the node (this statement is true in general for all boxes, not only for spherical).
View attachment 1199057
For those who want to play with placement of the microphone and the point source in 1m sphere, i prepared Akabak3 project file, you can run the project in Akabak3, which is free to download. Akabak3 automatically draws SPL graphs, you just need to download and install VACS viewer.
Can I suggest a Plan B?
50 or so years ago, the now-defunct Design Acoustics company was famous for its D-12 dodecahedral (12-sided polygon) loudspeaker. To me, it seems easier to build, especially in terms of mounting the drivers, than a sphere.
FWIW & YMMV
john
50 or so years ago, the now-defunct Design Acoustics company was famous for its D-12 dodecahedral (12-sided polygon) loudspeaker. To me, it seems easier to build, especially in terms of mounting the drivers, than a sphere.
FWIW & YMMV
john
Last edited by a moderator:
I'm building a 6" diameter sphere. Can Akabak3 model damping in the enclosure? What density of damping would smooth the response from 300Hz and up in a 6" diameter sphere? Simply packing the sphere tightly with fiberglass would avoid making tapered cylindrical enclosures for the back of each of 20 drivers and packing them with damping. That site isn't secure and Windows throws warnings about installing it. I could go build a virtual machine I guess.
Akabak3 allows you to deal with a wall damping, but generally speaking, damping of lowest modes by covering the walls with porous material is not a good idea, because the porous material works most efficiently when it is placed in the node of the standing wave, where the speed of the air particles is greatest (whereas at the wall it is almost equal 0). Usually in acoustics, wool/fiberglass with a density of 15-30 kg /m3 is used to absorb standing waves.
Here is an interesting experiment with standing wave absorption in a sphere. Probably I don't need to translate the video
I believe that Dave of Planet10 mentioned long ago the icosahedron. How about icosahedron segments of a spherical shell with a driver in each segment.
What you have illustrated is exactly what i imagined.
I would print as many triangles already joined together as the size of the printer allows. I have 35mm Foster copies of the patented JBL (patent expired).
Thanx for the file and the pictures.
dave
Totally agree with Dave. I thought about how difficult it would be to cleanly glue all those segments together and have nice looking seams. So I went back and made some drawings of the whole sphere. It may be nice to print two pieces, or just a single piece with an access hole at the back or bottom. It looks like a I can fit it on my printer. Here's a drawing with some of the driver mounting holes shown. The laptop won't render with all the drivers holes in place. Looking at it here the driver holes look too big. A bit more work to do.
Sacrifices were made. I wasn't able to find a buyer for the Ikea bowls. So the drivers came out of the bowls and went into the newly completed sphere hot off the printer. The trial fit is OK. I will have to file off a few corners on the driver flanges where they overlap. It would be great to get them closer together, but they are colliding on the inside as it is. Now to decide how many channels of DSP and amplifier to use. Maybe four or five for each sphere so I can go between omni, cardioid, dipole or point source settings on the remote control. Lots of work to do before there will be any sound. 20 gaskets to print, 80 screws to mount the drivers and lots of wires.
Impressive.
I guess the omni and the point source setting would be the same, with all drivers playing evenly at the same time since you have the shape for it.between omni, cardioid, dipole or point source
I'm working to get a pair of these done for the Seattle DIY show in Sept. One down, one to go. https://www.diyaudio.com/community/threads/seattle-diy-show-sept-9-2023.402223/
Two spheres in orbit, ready for the DIY show on Saturday Sept 9th, Noon till 4 PM at the Congregational Church on Mercer Island
4545 Island Crest Way, Mercer Island, Washington 98040, United States
https://audiosociety.org/meetings
Time to tune the crossover to the 12" woofer and EQ for the different radiation pattern modes: Point source, Omni, Dipole and Cardioid.
The microphone stands seem to work pretty well.
4545 Island Crest Way, Mercer Island, Washington 98040, United States
https://audiosociety.org/meetings
Time to tune the crossover to the 12" woofer and EQ for the different radiation pattern modes: Point source, Omni, Dipole and Cardioid.
The microphone stands seem to work pretty well.
I hope they're geostationary or you may need to duck from time to time.Two spheres in orbit,
That has got to be one of the coolest, most original speakers we have seen on this site in many months... maybe all year.
You know. these things would be awesome in a HT.
Half a sphere for the front center channel, 12 inches for the front L/R and 8" for the four surrounds.
Petaluma is not a long drive from Seattle.... about 800 miles. 5 to 505 to 80 to the bridge in Vallejo (580?) and then 101... or if you got time, 5 to Grants Pass, cut out to the Coast at Eureka and down the 101 to Petaluma. THAT has to be one of the most scenic drives in the entire USA.
Half a sphere for the front center channel, 12 inches for the front L/R and 8" for the four surrounds.
Petaluma is not a long drive from Seattle.... about 800 miles. 5 to 505 to 80 to the bridge in Vallejo (580?) and then 101... or if you got time, 5 to Grants Pass, cut out to the Coast at Eureka and down the 101 to Petaluma. THAT has to be one of the most scenic drives in the entire USA.
I have done it quite a few times. and other routes. Vancouver to San Fransisco.
There is 21 miles from 101 to 1 (?) that winds around a couple mountains. Just amazing. We drove our bikes all the way across from the Sierra Nevadas on our way back so we could do it again going the other way.
dave
There is 21 miles from 101 to 1 (?) that winds around a couple mountains. Just amazing. We drove our bikes all the way across from the Sierra Nevadas on our way back so we could do it again going the other way.
dave
I was asked to provide a more complete writeup of the sphere speakers at the show, so here it is.
The sphere drivers are wired into three groups. Front cap of six, middle ring of eight, rear cap of six.
This produced a narrow dispersion pattern to the front.
Cardiod was produced by inverting the phase of the rear cap and delaying the signal to the rear so it time aligned with the wave from the front as it came around the sphere. This worked well until I applied equalization to the front cap. With some EQ of the rear cap signal the cancelation to the rear should be nearly perfect.
The 6" size of the sphere results in the predicted high frequency roll off that was compensated using the PEQ filters in the miniDSP. During demonstrations no one commented on the roll off above 15 kHz. A small moon with tiny tweeters could be made to orbit this planet to fill in above 15 kHz.
All the comments were very positive from the 15 or so people that listen to the speaker at the PNWAS DIY show yesterday. \The shaded omni and the omni settings seemed to be the favorite as they produce a huge image at the end of the room. I found that getting the playback volume was critical to producing a realistic effect. If the vocals were at a level that could actually be produced by a human standing in the room it was amazingly realistic. Too loud and it sounded like you were standing too close to the singer and very weird. The balance of the singer to the instrumentation worked better or worse with different recordings. Pure instrumentals were more forgiving of level but is is always tempting to turn it up too much and then get blasted later in a song.
The sphere drivers are wired into three groups. Front cap of six, middle ring of eight, rear cap of six.
- The 12" ScanSpeak woofer in a half cube box on the floor is the fourth driver.
- A four channel miniDSP 2x4hd and four channels of amplification is needed for each speaker. I used four Dayton APA150 amplifiers with approximately 70 Watts/channel at the show. So eight amplifier channels and eight dsp channels in all for the stereo pair.
- The crossover to the sphere was at 350 Hz. The individual Aurasound drivers are rated at 5 Watts RMS continuous, so these amplifiers can drive the sphere to its limit and were able to reproduce realistic drum kit and organ SPL levels in the room.
- Point Source delay
- Cardioid
- Shaded Omni
- Omni directional
This produced a narrow dispersion pattern to the front.
Cardiod was produced by inverting the phase of the rear cap and delaying the signal to the rear so it time aligned with the wave from the front as it came around the sphere. This worked well until I applied equalization to the front cap. With some EQ of the rear cap signal the cancelation to the rear should be nearly perfect.
- Omni is just sending the same signal to all three channels with no delay
- Shaded omni, I applied a low pass filter to the center ring signal to reduce the cancelation effects when it combines with the front cap. Sort of treating the center ring as a midrange rather than a full range
The 6" size of the sphere results in the predicted high frequency roll off that was compensated using the PEQ filters in the miniDSP. During demonstrations no one commented on the roll off above 15 kHz. A small moon with tiny tweeters could be made to orbit this planet to fill in above 15 kHz.
All the comments were very positive from the 15 or so people that listen to the speaker at the PNWAS DIY show yesterday. \The shaded omni and the omni settings seemed to be the favorite as they produce a huge image at the end of the room. I found that getting the playback volume was critical to producing a realistic effect. If the vocals were at a level that could actually be produced by a human standing in the room it was amazingly realistic. Too loud and it sounded like you were standing too close to the singer and very weird. The balance of the singer to the instrumentation worked better or worse with different recordings. Pure instrumentals were more forgiving of level but is is always tempting to turn it up too much and then get blasted later in a song.
Last edited by a moderator:
It depends on what you are trying to build. Do you have a picture or drawing that you can post here? I drilled a pilot hole and then used a hole saw on a drill for the 1 1/2 inch holes. For larger holes you can make a template board to guide a router bit that has a bearing on it. That requires mounting the bowl to a table and then securing the template board above the bowl at what ever angle is required. The Blanda bowls are segments of bamboo So they are a bit fragile to drill.
Look at this post: https://www.diyaudio.com/community/threads/how-to-build-a-spherical-speaker.374278/post-7142865
and imagine an Ikea Gronsaker on the back instead of the Blanda.
and imagine an Ikea Gronsaker on the back instead of the Blanda.
I cut a circle of card the diameter of the hole I needed, put it inside the bowl and drew round it. I then got a handsaw and made small cuts progressively further in until I got close to the circle, then cut all the way round. Taped some sandpaper to the floor and rubbed it down to the circle. Bit long-winded, but worked
Brian
Brian
- Home
- Loudspeakers
- Multi-Way
- How to build a spherical speaker?