PURIFI finally did a fully purified passive speaker design! The SPK 16 prototype is here - with a PTT tweeter

[wineds]

Hi all, I've started preliminary work on the 3D printable version of the cabinet.

Based on a survey someone conducted on Reddit, it looks like the 80th percentile of the 3D printer consumer 'install base' has a print bed size of no more than 220 x 220 x 250mm, so I will be designing for that.

I seem to have been overly optimistic in my initial posts, it looks like it will take at least 8 prints per cabinet at this size. ("at least" because I also think it's important not to have seams cross the woofer or PR cutouts.) I will say people should start printing as soon as they order the kit.

I also would like feedback from the group and/or @Lars Risbo :

It seems pretty feasible to print at a density similar to MDF - someone check my math on this, but a 19x19x19mm cube with heavy settings of 6 walls and 15% gyroid infill has 6859 cubic mm of volume and takes ~5.1g of PLA to print. This apparently yields a density of 737.72 Kg / M^3 which is in the range of MDF.

This would be pretty costly in terms of material (around 11kg for the pair of cabinets, so anywhere from $100-250+ for PLA) and take a long time to print. However, it means no extra step of filling the model with plaster would be needed, at least for the sake of mass / density.

A sturdy-but-light setting of 4 walls and 5% gyroid would take more like 7kg of material for the pair, doable for less than $100, but the density would need to be enhanced with a filler to match MDF. So you'd have to get some plaster, pour it into a space in the wall of the cabinet after assembling it, and hope that step goes well. You can really keep the cost down with plaster as it's very cheap per kilo.

My gut tells me that for a $2200 kit, spending 5-10% on PLA is not a dealbreaker, and simplifying the build process is important. Based on what I've seen so far this is still cheaper than having MDF milled at a shop.

An advantage of designing it with a plaster layer is that it could help with airtightness. Otherwise, you will be relying on glue for that. I am planning on designing a special glue applicator to help with that part, but a solid layer of plaster would be more confidence-inspiring.

I don't think the bracing is a concern with a lighter print, because PLA is stiffer than MDF by quite a bit. I just want to make sure the walls are dense enough so the cabinet behaves as expected in that regard.

Anyone have strong opinions on the above?

(semi-related: anyone know how hot we would expect the crossover parts to get in real use? I'm thinking it would be nice to have a little printable template to attach the crossover parts to, but I lack any and all experience in that domain, and I could imagine PLA failing due to heat in that usage.)

For the money being spent it really should be a PCB for the crossover.

 

[Roland68]

Ps. Are you sure you have the price for the 19mm Valchromat?

A single sheet of 19mm black, of 2.44m x 1.22m, is £278 here.

This is probably also due to the high shipping costs to England for the large Valchromat plates.

 

[Luvchampagne]

For the money being spent it really should be a PCB for the crossover.

As long as we can get the timing, delays and crossovers, why not active over passive? Better performance. Use amps you want. Better efficiency

 

[Kaladin]

Hi all, I've started preliminary work on the 3D printable version of the cabinet.

Based on a survey someone conducted on Reddit, it looks like the 80th percentile of the 3D printer consumer 'install base' has a print bed size of no more than 220 x 220 x 250mm, so I will be designing for that.

I seem to have been overly optimistic in my initial posts, it looks like it will take at least 8 prints per cabinet at this size. ("at least" because I also think it's important not to have seams cross the woofer or PR cutouts.) I will say people should start printing as soon as they order the kit.

I also would like feedback from the group and/or @Lars Risbo :

It seems pretty feasible to print at a density similar to MDF - someone check my math on this, but a 19x19x19mm cube with heavy settings of 6 walls and 15% gyroid infill has 6859 cubic mm of volume and takes ~5.1g of PLA to print. This apparently yields a density of 737.72 Kg / M^3 which is in the range of MDF.

This would be pretty costly in terms of material (around 11kg for the pair of cabinets, so anywhere from $100-250+ for PLA) and take a long time to print. However, it means no extra step of filling the model with plaster would be needed, at least for the sake of mass / density.

A sturdy-but-light setting of 4 walls and 5% gyroid would take more like 7kg of material for the pair, doable for less than $100, but the density would need to be enhanced with a filler to match MDF. So you'd have to get some plaster, pour it into a space in the wall of the cabinet after assembling it, and hope that step goes well. You can really keep the cost down with plaster as it's very cheap per kilo.

My gut tells me that for a $2200 kit, spending 5-10% on PLA is not a dealbreaker, and simplifying the build process is important. Based on what I've seen so far this is still cheaper than having MDF milled at a shop.

An advantage of designing it with a plaster layer is that it could help with airtightness. Otherwise, you will be relying on glue for that. I am planning on designing a special glue applicator to help with that part, but a solid layer of plaster would be more confidence-inspiring.

I don't think the bracing is a concern with a lighter print, because PLA is stiffer than MDF by quite a bit. I just want to make sure the walls are dense enough so the cabinet behaves as expected in that regard.

Anyone have strong opinions on the above?

(semi-related: anyone know how hot we would expect the crossover parts to get in real use? I'm thinking it would be nice to have a little printable template to attach the crossover parts to, but I lack any and all experience in that domain, and I could imagine PLA failing due to heat in that usage.)

I wouldn’t print this in PLA personally. I would expect that creep will be an issue over time under the constant weight of the drivers.

For filled cabinet walls you could reduce the number of perimeters down to 2 or 3. Geometry is important here to limit potential spots where the printer could leave a gap for the fill material to leak out. A fill test with water beforehand is recommended. If you do decide to fill with plaster you will want to add something like pva glue into the mix.

 

[kemmler3D]

I wouldn’t print this in PLA personally. I would be worried that creep will be an issue under the constant weight of the drivers.

It's a fair concern and something I will look into a bit more to see how likely and how much could be expected to happen. Of course you can just print it in PETG or whatever you have on hand, but I've had PLA in mind because I think the extra stiffness is desirable for this application.

FWIW I've experienced some PLA creep on a large (70lb) diffuser I built, but it was with fewer walls than we're talking about here. It was slow but noticeable, like 1-2mm per week.

But now that you mention it, I imagine creep is pretty likely if I put threaded inserts for VESA mounting on the back, so realistically those should go on the bottom if anywhere...

For the money being spent it really should be a PCB for the crossover.

I wouldn't necessarily argue but that's out of scope for what I can contribute here.

For filled cabinet walls you could reduce the number of perimeters down to 2 or 3. Geometry is important here to limit potential spots where the printer could leave a gap for the fill material to leak out. A fill test with water beforehand is recommended. If you do decide to fill with plaster you will want to add something like pva glue into the mix.

Good points. There will be a lot of potential for leaks, especially since the end user will be gluing together 16 pieces, so even if the geometry is perfect in the STL or even off the print bed, (my printer is good about warping, not sure everyone's is...) there's plenty of opportunity for tiny gaps to open. Realistically I would expect minor leaks in most cases, the plaster would be how those leaks are ultimately plugged...

BTW welcome to ASR!

 

[Roland68]

I wouldn’t print this in PLA personally. I would expect that creep will be an issue over time under the constant weight of the drivers.

For filled cabinet walls you could reduce the number of perimeters down to 2 or 3. Geometry is important here to limit potential spots where the printer could leave a gap for the fill material to leak out. A fill test with water beforehand is recommended. If you do decide to fill with plaster you will want to add something like pva glue into the mix.

It's a fair concern and something I will look into a bit more to see how likely and how much could be expected to happen. Of course you can just print it in PETG or whatever you have on hand, but I've had PLA in mind because I think the extra stiffness is desirable for this application.

FWIW I've experienced some PLA creep on a large (70lb) diffuser I built, but it was with fewer walls than we're talking about here. It was slow but noticeable, like 1-2mm per week.

But now that you mention it, I imagine creep is pretty likely if I put threaded inserts for VESA mounting on the back, so realistically those should go on the bottom if anywhere...

I wouldn't necessarily argue but that's out of scope for what I can contribute here.

PLA is theoretically rigid, but in practice, it's also very brittle in such structures, especially regarding layer adhesion along the Z-axis.
The material is completely unsuitable for technical applications.

PETG and PET offer significant advantages and aren't even more expensive. I currently pay around €80-85 for 11 kg of Jayo HS PETG (approx. €7.50/kg). For brand-name PET, it's around €11/kg.

PC is significantly better, but also more expensive. However, it bonds much more easily, even with cold welding.

 

[kemmler3D]

PLA is theoretically rigid, but in practice, it's also very brittle in such structures, especially regarding layer adhesion along the Z-axis.
The material is completely unsuitable for technical applications.

PETG and PET offer significant advantages and aren't even more expensive. I currently pay around €80-85 for 11 kg of Jayo HS PETG (approx. €7.50/kg). For brand-name PET, it's around €11/kg.

PC is significantly better, but also more expensive. However, it bonds much more easily, even with cold welding.

Agree that cost is not a concern for PETG vs PLA. At first I was going to disagree on suitability, on the grounds that the forces involved aren't that serious, but after attempting to do super basic calculations for the force applied to the baffle I'm not so sure.

I also realized that PLA and PETG are both much stiffer than MDF so there isn't a compelling reason to go with PLA, at least from that POV.

I think the air pressure from sound inside the cabinet is not a concern... 153dB SPL ~ 10 millibar ~ .014 PSI. Any remotely realistic SPL is not going to do anything bad to the cabinet by itself.

On the other hand, moving mass for the woofer is 28g and xmech is +/- 14.5mm - according to my (likely wrong, what am I missing here?) calculations this means it applies a force of well over 100lb if it's hitting Xmech at 1Khz+...? In which case I would have real concerns. At maximum SPL this driver might be pretty demanding on the housing... yes it's distributed over some area, but I don't think I would sit on a chair supported by such a small bit of PLA... and the force is at least loosely comparable to that.

I think in normal usage the force would be a tiny fraction of that, but I don't like the idea of technically being able to drive the speaker hard enough for it to break the housing.

For now I will consider PETG the material of choice, but my guess is that if someone chooses to print in PLA it won't truly affect outcomes unless the user is trying to break the speaker.

I would say that PC is not a likely choice for me personally just because of the cost, but I could imagine someone using it for this. IME effectively bonding parts with 2 part epoxy or even CA glue is not a big deal.

 

[Kaladin]

Good points. There will be a lot of potential for leaks, especially since the end user will be gluing together 16 pieces, so even if the geometry is perfect in the STL or even off the print bed, (my printer is good about warping, not sure everyone's is...) there's plenty of opportunity for tiny gaps to open. Realistically I would expect minor leaks in most cases, the plaster would be how those leaks are ultimately plugged...

BTW welcome to ASR!

Thanks!

I have a Bambu H2D currently and it’s a nice printer but issues still pop up. A recent example is where I used dowel plugs to help align parts for gluing. Some of the cylindrical mating holes were missing the very bottom walls leaving gaps that were easy to miss with black filament. The low infill percentage can help create issues that you wouldn’t normally worry about. In general, I’ve had success with spraying a couple coats of polyurethane to seal off any minor leaks before filling a print.

For filaments, PETG-CF is a solid candidate to consider and can be found around $18 a kg.

 

[wineds]

It's a fair concern and something I will look into a bit more to see how likely and how much could be expected to happen. Of course you can just print it in PETG or whatever you have on hand, but I've had PLA in mind because I think the extra stiffness is desirable for this application.

FWIW I've experienced some PLA creep on a large (70lb) diffuser I built, but it was with fewer walls than we're talking about here. It was slow but noticeable, like 1-2mm per week.

But now that you mention it, I imagine creep is pretty likely if I put threaded inserts for VESA mounting on the back, so realistically those should go on the bottom if anywhere...

I wouldn't necessarily argue but that's out of scope for what I can contribute here.


Good points. There will be a lot of potential for leaks, especially since the end user will be gluing together 16 pieces, so even if the geometry is perfect in the STL or even off the print bed, (my printer is good about warping, not sure everyone's is...) there's plenty of opportunity for tiny gaps to open. Realistically I would expect minor leaks in most cases, the plaster would be how those leaks are ultimately plugged...

BTW welcome to ASR!

Interesting also that looking at the build guide there doesn't seem to be a place for the crossover inside the cabinet.

 

[JanRSmit]

I wouldn’t print this in PLA personally. I would expect that creep will be an issue over time under the constant weight of the drivers.

For filled cabinet walls you could reduce the number of perimeters down to 2 or 3. Geometry is important here to limit potential spots where the printer could leave a gap for the fill material to leak out. A fill test with water beforehand is recommended. If you do decide to fill with plaster you will want to add something like pva glue into the mix.

For preventing creep i would choose petg, my experience is that after a little initial creep it does not anymore.
Plaster is brittle and not strong, not a good choice in my view, as cracks are air filled thus a barrier for transmission of structure borne vibration/energy.

 

[3ll3d00d]

A single sheet of 19mm black, of 2.44m x 1.22m, is £278 here.

seems a bit much, e.g. https://www.slhardwoods.co.uk/product/moisture-resistant-valchromat-mdf-2440-x-1220mm-black/

 

[Fredon37]

Reading the data sheet, it mentions a 'hard anodised aluminium dome', does that mean some kind of type III 'hardcoat' anodising or is it some kinda long bath type II anodise?

 

[Lars Risbo]

the 3D CAD files have been added to the downloads

 

[RolandRa]

@Lars Risbo, could you please share the criteria behind the choice of the selected damping material?


I’m having difficulty finding any technical documentation or datasheets describing its acoustic properties (e.g., absorption coefficients, flow resistivity, density, etc.), which makes it hard to evaluate and compare it against more commonly available materials such as melamine foam (e.g., Basotect).


Given its limited availability outside Denmark, a bit more insight into the selection rationale and expected performance would be very helpful.

 

[kemmler3D]

A recent example is where I used dowel plugs to help align parts for gluing. Some of the cylindrical mating holes were missing the very bottom walls leaving gaps that were easy to miss with black filament. The low infill percentage can help create issues that you wouldn’t normally worry about. In general, I’ve had success with spraying a couple coats of polyurethane to seal off any minor leaks before filling a print.

For filaments, PETG-CF is a solid candidate to consider and can be found around $18 a kg.

I am also a fan of dowels for aligning and slightly strengthening parts, but haven't had the problem with missing walls / layers... sounds like it could be a slicer issue?

15% gyroid leaves gaps of a few mm, but I think more importantly 6 walls / top / bottom layers will take care of any sketchy bits.

 

[Lars Risbo]

@Lars Risbo, could you please share the criteria behind the choice of the selected damping material?


I’m having difficulty finding any technical documentation or datasheets describing its acoustic properties (e.g., absorption coefficients, flow resistivity, density, etc.), which makes it hard to evaluate and compare it against more commonly available materials such as melamine foam (e.g., Basotect).


Given its limited availability outside Denmark, a bit more insight into the selection rationale and expected performance would be very helpful.

the open cell recycled foam is annindustry standard so should be available internationally. Not saying it’s optimal in any way. it proved very effective in increasingthe apparent volume (lowering fs is a sealed box). It is less effective than Rockwool is damping standing wave in the sub 200Hz range. We later found that the new building insulation materials based on wood fiber /cellulose are quite effective. close to rockwool but not irritant/allergen. The Basotect was disappointing. it has a too low density for low frequency effectiveness and it breaks easily and makes screechy sounds when rubbing it. It’s very important that the damping material can be fixated mechanically so that it does not move from the air flow forces. Also the sounds it makes gives a clue. the open cell foam locks in place effectively and does not make sounds when rubbing it. one day we built an impedance tube to test all this.

 

[Kaladin]

I am also a fan of dowels for aligning and slightly strengthening parts, but haven't had the problem with missing walls / layers... sounds like it could be a slicer issue?

15% gyroid leaves gaps of a few mm, but I think more importantly 6 walls / top / bottom layers will take care of any sketchy bits.

It was a combination of issues. I could’ve reinforced the cylindrical holes with additional material around them but I just used Bambu slicer to split the box and create the connectors/holes and called it good. I also used a lower infill (in the 5-7% range) with 3 walls to save filament and leave plenty of room for the epoxy mix to flow through. The low infill percentage didn’t always provide enough support for the bottom layer of the cylindrical holes and Bambu slicer didn’t care.

 

[Roland68]

Agree that cost is not a concern for PETG vs PLA. At first I was going to disagree on suitability, on the grounds that the forces involved aren't that serious, but after attempting to do super basic calculations for the force applied to the baffle I'm not so sure.

I also realized that PLA and PETG are both much stiffer than MDF so there isn't a compelling reason to go with PLA, at least from that POV.

I think the air pressure from sound inside the cabinet is not a concern... 153dB SPL ~ 10 millibar ~ .014 PSI. Any remotely realistic SPL is not going to do anything bad to the cabinet by itself.

On the other hand, moving mass for the woofer is 28g and xmech is +/- 14.5mm - according to my (likely wrong, what am I missing here?) calculations this means it applies a force of well over 100lb if it's hitting Xmech at 1Khz+...? In which case I would have real concerns. At maximum SPL this driver might be pretty demanding on the housing... yes it's distributed over some area, but I don't think I would sit on a chair supported by such a small bit of PLA... and the force is at least loosely comparable to that.

I think in normal usage the force would be a tiny fraction of that, but I don't like the idea of technically being able to drive the speaker hard enough for it to break the housing.

For now I will consider PETG the material of choice, but my guess is that if someone chooses to print in PLA it won't truly affect outcomes unless the user is trying to break the speaker.

I would say that PC is not a likely choice for me personally just because of the cost, but I could imagine someone using it for this. IME effectively bonding parts with 2 part epoxy or even CA glue is not a big deal.

I just did a very rough quick test to see if a one-piece design with internal structure is an option.
It should work.

 

[Salt]

I just did a very rough quick test to see if a one-piece design with internal structure is an option.
It should work.

View attachment 518963View attachment 518964View attachment 518965View attachment 518967View attachment 518968

Now a version with separated space for a plate amp ...

 

[Paul7777x]

Now a version with separated space for a plate amp ...

A plate amp will necessarily remove some of the bracing and foam will it not?

A couple of power amps on the outside, or a four channel Purifi amp perhaps?

I’ve never seen the need for a plate amp really.

This is all assuming someone intends to do a madly magnificent active set up (and be kind enough to share it) for a Minidsp or similar.

I’m certainly not up to it, sadly, but a three way active crossover for, for instance, the Minidsp Flex Eight would make me drool.