3d Printed Speakerenclosure 2 Way

[Terplover]

I thought it was about time to start a thread on my second 3D-printed speaker build! I'm not an expert in speaker design or acoustics, so I’d be genuinely grateful for any tips, feedback, or advice you’re willing to share.

After being really impressed by the results of the D-Note 3D-printed speakers,

https://www.reddit.com/r/diyaudio/comments/101fwos/d_notes_dayton_c_notes_with_3d_printed_cabinets

I decided to go for another printed project. The plan is to build a 2-way speaker using:

• SB Acoustics SB17CAC35-4 (woofer)
• SB Acoustics SB26CDC-C000-4 (tweeter, with horn or waveguide)
• SB Acoustics SB15SFCR-00 (passive radiator)

I picked these drivers because I like SB Acoustics’ design and, based on what I’ve read, they offer excellent value – great performance for the price.



A few days ago, I came across the Ascilab C6b and instantly fell in love with the clean, all-white look.

ascilab c6b

ascilab c6b passive speaker

ascilab.com

I fired up Fusion 360 and designed a ~15L enclosure with an integrated waveguide. Because of print size limitations, I’ll split the box in half and align the pieces using metal pins for proper assembly.

Now, here’s where I could really use your input:



Tweeter Horn/Waveguide Shape
Found a Oval Waveguide for the SB26CDC, but I don’t like the look. Can I simply "cut" the oval guide in half and rotate the profile to make a round horn? Or how would this change its behavior? I’m not experienced enough to design a proper horn from scratch.

Wall Thickness or Plaster of Paris?
I considered printing the enclosure and then filling the walls with plaster of Parisor just print with enough infill. Costs almost the same. Price for Filament is around 6€/KG. With 17mm walls i need 4kg per Speaker. With 20mm for Front an Back and 10mm for the rest, It,s around 2,9kg. With Plaster of Paris its still 1,6kg + The Plaster and additives ~10-15€
I found a lot of info online about infill percentages and number of walls, but not much on actual wall thickness.
The D-Note (with a smaller 5” woofer) had ~8 mm side walls and a thicker baffle, and it sounded great — no resonance or leakage I could detect.
Would something like 20 mm for the baffle/front and rear and 10 mm for the sides be a good compromise between performance and material use? In my first design, I added inner support beams to improve rigidity.

Internal Diffusors Worth It?
Is it worth printing some internal surface texture like small waves or multiple beams to break up reflections?


Tweeter Chamber Separation?
I read somewhere that some designs isolate the tweeter from the main volume. Is that worth doing? I haven’t seen it implemented in many builds.

Thanks for reading through all of that! I’m really excited to continue this project and would love to hear what you think or what you would do differently.

 

[DVDdoug]

The main things are the cabinet volume and how it's "tuned" with the passive radiator. If it's not a tried-and-tested design, look for some speaker design software to get it optimized. (I have WinISD but it only models sealed & ported designs.)

The waveguide is also important but I assume it's an "ideal compromise".

And the cabinet should be stiff-enough that it doesn't resonate.

Is it worth printing some internal surface texture like small waves or multiple beams to break up reflections?

No... But ribs to stiffen it might be worthwhile. It shouldn't resonate when you knock on it... Lining the box with fiberglass (or similar) or "stuffing" it sometimes helps and "doesn't hurt".

The smaller wavelengths (higher frequencies) that would be affected by something like that aren't coming out of the back of the tweeter so they don't exist inside the box. The internal box resonances (ignoring the passive radiator) are probably somewhere in the 1kHz range. Wavelength Calculator

I read somewhere that some designs isolate the tweeter from the main volume. Is that worth doing? I haven’t seen it implemented in many builds.

Virtually all tweeters are sealed on the back side (they have their own tiny "cabinet") so that's not necessary. Some midranges, or some speakers used as midranges, have open backs and need their own "cabinet".

 

[Digital_Thor]

Check out Augerpro, his designs work well, I tried and were very satisfied.
Haven't tried his round ones, but his measurements are right there to prove they work too.

SB Acoustics SB26 | Mysite

www.somasonus.net

 

[Terplover]

Check out Augerpro, his designs work well, I tried and were very satisfied.
Haven't tried his round ones, but his measurements are right there to prove they work too.

SB Acoustics SB26 | Mysite

www.somasonus.net

Ah, thanks! I didn’t realize there is a circular waveguide too. I probably should’ve looked a bit closer before asking . Somehow I completely missed that one. the Oval one is already from him.

 

[Digital_Thor]

No worries, We're here to help and learn, that's the beauty of this forum

 

[kemmler3D]

Awesome thread, thanks for sharing all this! I think the thickness and density of the cabinet does matter, you'll probably get better results with the plaster of Paris. Something I've also thought about with designs like this (always thinking but never building) is that you can potentially achieve higher transmission loss through the cabinet walls by alternating high and low density layers via increased total internal reflection. So maybe 2 thinner plaster of Paris layers with relatively light infill could do something. I'm not sure but would love to see an experiment.

@DVDdoug is unfortunately correct that even though you can put fun diffuser shapes inside a 3d printed cabinet, the feature size will tend to be much smaller than any wavelength the woofer puts out, which is no fun.

However, despite what people at DIY Audio often say (spheres only have one resonant mode) that's not correct as far as I know. Spheres have a series of resonances like any other shape, and spheres happen to have an inharmonic series of resonances, unlike boxes. And although I'm not very confident in this idea, I think this is partly why spherical speakers sound a little nicer than you would expect.

A long way of saying that I don't think diffusion makes sense inside speakers but significantly non-rectangular internal volumes might be worthwhile?

 

[Salt]

You want to know what is going on within a sphere? Just revert a tweeter or insert one and measure it's output ...

 

[a4eaudio]

...I have WinISD but it only models sealed & ported designs..

@DVDdoug - WinISD can model passive radiators, see screenshot below.

BUT... @Terplover VituixCAD's encloser modelling tool is more accurate than WinISD.

 

[ppataki]

Just FYI, BassBox Pro can also simulate PR cabinets

 

[Terplover]

The main things are the cabinet volume and how it's "tuned" with the passive radiator. If it's not a tried-and-tested design, look for some speaker design software to get it optimized. (I have WinISD but it only models sealed & ported designs.)

The waveguide is also important but I assume it's an "ideal compromise".

WinISD is capabel of Passive Radiator too. got 14,2 L for optimal results with it and vitruixCAD. Added 0,8L for internal beams and Driver Volume. Found a Circulkar 6,5 inch version of it too, will use this.

And the cabinet should be stiff-enough that it doesn't resonate.

I think i will definitly try with smaller fully printed walls. Measured some of the D-Note Speaker and they are between 8-5mm and front around 13-15 if i had to guess.

No... But ribs to stiffen it might be worthwhile. It shouldn't resonate when you knock on it... Lining the box with fiberglass (or similar) or "stuffing" it sometimes helps and "doesn't hurt".

The smaller wavelengths (higher frequencies) that would be affected by something like that aren't coming out of the back of the tweeter so they don't exist inside the box. The internal box resonances (ignoring the passive radiator) are probably somewhere in the 1kHz range. Wavelength Calculator


Virtually all tweeters are sealed on the back side (they have their own tiny "cabinet") so that's not necessary. Some midranges, or some speakers used as midranges, have open backs and need their own "cabinet".

Makes sense. Thanks for the Input

 

[Terplover]

Awesome thread, thanks for sharing all this! I think the thickness and density of the cabinet does matter, you'll probably get better results with the plaster of Paris. Something I've also thought about with designs like this (always thinking but never building) is that you can potentially achieve higher transmission loss through the cabinet walls by alternating high and low density layers via increased total internal reflection. So maybe 2 thinner plaster of Paris layers with relatively light infill could do something. I'm not sure but would love to see an experiment.

@DVDdoug is unfortunately correct that even though you can put fun diffuser shapes inside a 3d printed cabinet, the feature size will tend to be much smaller than any wavelength the woofer puts out, which is no fun.

However, despite what people at DIY Audio often say (spheres only have one resonant mode) that's not correct as far as I know. Spheres have a series of resonances like any other shape, and spheres happen to have an inharmonic series of resonances, unlike boxes. And although I'm not very confident in this idea, I think this is partly why spherical speakers sound a little nicer than you would expect.

A long way of saying that I don't think diffusion makes sense inside speakers but significantly non-rectangular internal volumes might be worthwhile?

Funny thing is, there always seem to be two camps:

Party 1 – “You need massive, thick walls for good sound.”
Party 2 – “I’ve already tested thinner walls with low infill and noticed zero quality loss.”

I’m leaning more toward the second group, but I guess it always depends on the design and driver combo.


I know a sphere might have some acoustic benefits, but I’ll stick with this current boxy design — I just really like how it looks. As for filling the walls with plaster: yeah, I’m still unsure. I thought about using something like gyroid infill at 2–5% to allow it to flow through, but I have no clue if that would actually work or if the plaster would just clog everything up and create airpockets. Or if i should just construct the Infill myself. But would be sht ton of work with my small knowledge of CAD.

I might just give it a shot with thinner walls and focus on good engineering for stiffness. If it works out well, it could save me a lot of effort on future projects without making things overly complicated. And if I end up disliking the first enclosure, I can still go the more solid route with thick plaster walls later on.

 

[Terplover]

You want to know what is going on within a sphere? Just revert a tweeter or insert one and measure it's output ...

Good point. I think I’ll just follow the advice and skip the separate chamber for now. Probably not worth overcomplicating things at this stage. But I’ll definitely add it to my “boring but maybe useful” list for later on

 

[kemmler3D]

You want to know what is going on within a sphere? Just revert a tweeter or insert one and measure it's output ...

So, yeah

Funny thing is, there always seem to be two camps:

Party 1 – “You need massive, thick walls for good sound.”
Party 2 – “I’ve already tested thinner walls with low infill and noticed zero quality loss.”

I’m leaning more toward the second group, but I guess it always depends on the design and driver combo.


I know a sphere might have some acoustic benefits, but I’ll stick with this current boxy design — I just really like how it looks. As for filling the walls with plaster: yeah, I’m still unsure. I thought about using something like gyroid infill at 2–5% to allow it to flow through, but I have no clue if that would actually work or if the plaster would just clog everything up and create airpockets. Or if i should just construct the Infill myself. But would be sht ton of work with my small knowledge of CAD.

I might just give it a shot with thinner walls and focus on good engineering for stiffness. If it works out well, it could save me a lot of effort on future projects without making things overly complicated. And if I end up disliking the first enclosure, I can still go the more solid route with thick plaster walls later on.


View attachment 460620View attachment 460621

If I were going to do a plaster layer I would design the void into the case from the beginning (no infill) and then seal it after the fact, either with a two part print or a plug of some kind...

I'm not disputing the mdf vs. print result, but did you look at decay times / resonances also? I'd wonder if the results hold up in that regard as well as at lower frequency.

I guess you'd also want to look at transmission through the walls / vibration which could be audible but not show up in a fr plot very well. Basically I'm just trying to think of reasons I've seen people try to improve on standard MDF in DIY Audio threads.

All that said I think you can print a lot better bracing than you can build with MDF so maybe thicker and heavier walls really are moot. You could do a full internal polyhedral grid type of thing and only need to scale up volume a little.

 

[somebodyelse]

However, despite what people at DIY Audio often say (spheres only have one resonant mode) that's not correct as far as I know. Spheres have a series of resonances like any other shape, and spheres happen to have an inharmonic series of resonances, unlike boxes. And although I'm not very confident in this idea, I think this is partly why spherical speakers sound a little nicer than you would expect.

Room modes and cabinet internal modes are essentially the same, so you could use one of the arbitrary room shape eigenmode calculators to see what happens in non-cuboid enclosures.
https://www.audiosciencereview.com/...ds/arbitrary-shape-room-mode-simulator.20805/
https://www.audiosciencereview.com/forum/index.php?threads/room-eigenmodes-calculator.38734/ - apparently more user friendly.

For wall resonances it would be possible to use something like the FEM integration in FreeCAD - there are tutorials, but I've never got past a simple beam or tuning fork example. I expect there's something similar for those using Fusion360.

 

[kemmler3D]

Room modes and cabinet internal modes are essentially the same, so you could use one of the arbitrary room shape eigenmode calculators to see what happens in non-cuboid enclosures.
https://www.audiosciencereview.com/...ds/arbitrary-shape-room-mode-simulator.20805/
https://www.audiosciencereview.com/forum/index.php?threads/room-eigenmodes-calculator.38734/ - apparently more user friendly.

For wall resonances it would be possible to use something like the FEM integration in FreeCAD - there are tutorials, but I've never got past a simple beam or tuning fork example. I expect there's something similar for those using Fusion360.

Oh, interesting. I did look up the modes for a sphere a long time ago and they're different than rectangular prisms, at least. I mean, if you think about it, a bouncing basketball has a pretty distinct sound, it can't be the result of a regular harmonic series... Right?

 

[somebodyelse]

If I were going to do a plaster layer I would design the void into the case from the beginning (no infill) and then seal it after the fact, either with a two part print or a plug of some kind...

I've seen that method used with concrete fill for making machine tools since it doesn't flow as easily as plaster of paris. He also used an electric sander as a vibration source to make it flow more easily, which ought to work for plaster in a gyroid too. See https://www.youtube.com/@Borgedesigns/videos for the various concrete filled things so you can see the problems and the lessons learned over successive projects.

 

[Terplover]

So, yeah

If I were going to do a plaster layer I would design the void into the case from the beginning (no infill) and then seal it after the fact, either with a two part print or a plug of some kind...

I'm not disputing the mdf vs. print result, but did you look at decay times / resonances also? I'd wonder if the results hold up in that regard as well as at lower frequency.

I guess you'd also want to look at transmission through the walls / vibration which could be audible but not show up in a fr plot very well. Basically I'm just trying to think of reasons I've seen people try to improve on standard MDF in DIY Audio threads.

All that said I think you can print a lot better bracing than you can build with MDF so maybe thicker and heavier walls really are moot. You could do a full internal polyhedral grid type of thing and only need to scale up volume a little.

I'm not super deep into analyzing every graph, but it doesn't seem that bad overall. Measurement was taken on the tweeter axis (speaker is almost right in front of the wall – don’t feel like dismounting it right now ). Clarity shows a dip around 100–300 Hz, not exactly sure what that means – maybe a room mode? There's also a dip at 180 Hz visible in the waterfall plot.

There are some slight vibrations when touching the cabinet from the outside, but it’s not really more or less than what I feel when touching anything near the speaker. I think that should be manageable with proper internal structure and bracing. I’ll design the first version today and hopefully start printing later this week. And as I said, the D-Note Walls are Way thinner and with just 40% infill. consumed just 1 kg of filament per Speaker. And its doesnt have any bracing at all. just the Spherebody

 

[somebodyelse]

Oh, interesting. I did look up the modes for a sphere a long time ago and they're different than rectangular prisms, at least. I mean, if you think about it, a bouncing basketball has a pretty distinct sound, it can't be the result of a regular harmonic series... Right?

The bouncing basketball has an interaction with the wall too, which probably makes it even more interesting. The sound from tapping a rigging wire under tension is distinct too, but as far as I know that's ordinary harmonics, so I don't know. I've barely touched on spherical harmonics.

 

[Terplover]

Here is my second attempt: 15.1L without the beams and around 14,2-14,3 without each drivervolume. WINISD and VituixCAD got me 14,2 L for my configuration.

Front baffle is 18 mm, back panel 15 mm, and the sides/bottom are 10 mm thick. I've added internal beams where they won’t interfere with the drivers, to help with stiffness and rounded inner edges too

For assembly, I plan to use 6×30 mm wooden dowels for alignment and a strong white mounting glue. Hopefully, I won’t need to do much (if anything) for the surface finish afterwards.

I’ll also design a temporary internal plate for the assembly process, so I can screw the two halves together at the beams from the inside. Once the glue has set, I’ll remove the structure.

I’ll tweak my slicer settings so that each part uses around 2.2 kg of filament, since my spools are about 1.1 kg each. Current settings: 4 perimeters, 40% infill, and supports only where absolutely necessary. Bridging should handle the rest. 3,5 Days Printing time is a bit scary to be honest ^^

I’m printing in this orientation with variable layer height to get smooth round corners. If I flipped the part 180°, the outer edges wouldn’t look as clean

 

[Torin Krell]

Perhaps this is helpful for this project or for anyone else filling hollow 3-D printed speaker cabinet walls with plaster. Apparently, as discussed and demonstrated in the video below, adding some PVA glue to liquid plaster will, when it hardens, create a compound with significant resonance damping, in significant contrast to pure plaster alone which has a tendency to ring.