3D printing a QRD N53 diffuser

[kemmler3D]

Hey all, in order to finally live up to my username I'm going to attempt 3D printing something audio-related.

I have always thought that 2D QRD diffusers look cool, and now I have a room to myself where I can hang one up without making common areas of the house look strange.

Buying a large QRD diffuser is not cheap, they're thousands of dollars at least. And building one from wood is not for the faint of heart, with a N=53 diffuser incorporating 2,809 individually cut & glued bits of wood. And the weight...

For a project like this, a 3D printer has a huge advantage. The parts can be made stiff but relatively light, the material is fairly cheap (about $20/kg $10/kg if you buy in bulk) and there is a great deal less manual labor involved. You have to wait a good while for the printer to finish its job, but it's no more work than waiting for a washing machine to finish.

It's possible that it will also benefit my room acoustics. I'm going to try and document the process in this thread.

Q) Why N=53?

I only have about 1.3m of free space on my wall, and apparently you're not supposed to make the wells less than 25mm on a side. This means that although I could do a N=101 diffuser just as easily as an N53 with this method, I don't have room to hang it up.

Step 1: Designing the diffuser:

I used the venerable QRDude to generate the well depths and size of the diffuser. Settings here:

It gives you a "report" as an export with lots of labels and helpful info. The output is geared toward someone using it as instructions to build by hand. Luckily, it's almost formatted as a CSV, so it's not totally hopeless. There are a few (janky) steps involved in turning that report into a 3D model:

(raw report in Notepad++)

After deleting everything except the well depths in mm, I loaded it into Google Sheets, to use the conditional formatting feature to turn the numbers into a grayscale image.

Step 1: Loaded as CSV

Step 2: Conditional formatting with minimum = black and maximum value = white:

Step3 : You have to paste this into a new file to get rid of the conditional formatting so you can delete the values but keep the colors... I exported that as a PDF, loaded it into Inkscape, then cropped to just the colored squares and exported as a PNG.

This gives us what's known as a "texture" and more specifically a "heightmap" in the biz. The 3D software takes the grayscale values and translates them to a displacement in 3D space. Basically white areas stick out more and black areas stick out less. You are probably seeing where this is going. I loaded the texture into Blender and used it to manipulate a grid of 25mm squares:

This actually took more screwing around than I've described here, but in the end we get something that looks right. Comparing it to the dimensions we started with in QRDude, looks like we nailed it.

A quick render with semi-realistic lighting to help imagine the final product:

Ooh, fancy.

So we've got the basic shape in place.

Next comes splitting the squares into chunks that will fit on a 3D printer, adding attachment points and holes for screws so it's easy to glue them together, and printing the pieces, painting the pieces, gluing them together, and hanging it up!

For the next step, I'm investing in a nicer printer than I have now, the Bambu Labs P1S, which is apparently set to go on sale tomorrow. It prints about 4x faster. That's important because this is going to require about 25 36 print jobs near the maximum size the printer can handle, each of which will take several hours, maybe close to an entire day. (e: having messed with the slicer a bit just now, looks like each of the sections will take about 760-1100g of filament and 12-22 hours to print. Woof. Better look for a bulk deal...)

More progress updates to follow, in the meantime happy to answer any questions about this, or take advice on what I should be building instead.

 

[soundflo_w]

Thank you for sharing this project. I'm looking forward to further progress updates!

What slicer settings are you going to use? Which filament material?

You could consider using a greater diameter nozzle, since you don't need to print any details. This could reduce the printing time further, but is limited by the heat capacity of the hot-end in that at one point the heating can't keep up with the requested flow rate (but you can experiment with using a higher print temperature to offset this).
These videos might give more insights:
-https://www.youtube.com/watch?v=UNJdv5bFGOg
-https://www.youtube.com/watch?v=ZgIlSpb-A2Y

Just a guess, but something that came to my mind is that the structure might have resonances or vibrations? Hollow wooden diffusers are typically filled with damping material from the back side.

Also, for most print jobs you have to remember that you have to interact two times with the printer, setting the new print job up and changing the empty filament mid print.

Good luck with the project!

 

[kemmler3D]

Thank you for sharing this project. I'm looking forward to further progress updates!

What slicer settings are you going to use? Which filament material?

You could consider using a greater diameter nozzle, since you don't need to print any details. This could reduce the printing time further, but is limited by the heat capacity of the hot-end in that at one point the heating can't keep up with the requested flow rate (but you can experiment with using a higher print temperature to offset this).
These videos might give more insights:
-https://www.youtube.com/watch?v=UNJdv5bFGOg
-https://www.youtube.com/watch?v=ZgIlSpb-A2Y

Just a guess, but something that came to my mind is that the structure might have resonances or vibrations? Hollow wooden diffusers are typically filled with damping material from the back side.

Also, for most print jobs you have to remember that you have to interact two times with the printer, setting the new print job up and changing the empty filament mid print.

Good luck with the project!

Thanks! From what I've read the P1S hotend can only handle so much volume and the print speed can be limited by flow rate, depending on material. But since I'm planning on using PLA it might be worth looking into a 0.6 or 0.8 nozzle and cranking it up. On the other hand, since the piece will be hanging on my wall, I don't want to go crazy with thick layer lines just to save time.

I think I'm going to go with basic white PLA and paint it later. I'm trying to come up with a method of dipping each piece in paint because I think spraying down into the deep wells, or getting a brush in there, will be virtually impossible. Anyone with clever ideas on painting, please chime in.

As it stands each piece will take about 14 hours and the whole thing (I estimate) will take roughly 300-400 hours. This is almost all passive time, I just need to go downstairs and remove completed prints and start new ones about 25-30 times.

I am going to be running the printer in a disused shower, which sounds weird but is ideal because 1) nobody really goes in there and 2) it has ventilation to the outdoors, so the microplastics / fumes are dealt with and I can sit the printer on a solid tile surface.

As for resonances, maybe I should be more worried, but the infill should act as "bracing" every inch or so, and diffusers work with reflection, not absorption, so I think it should be OK. It's important to me to keep weight low. Even with 2 shells, 3 top/bottom layers, and 2.5% infill, (I normally do 3 shells and 10% infill for functional parts) the whole thing is going to weigh about 19 kilos before paint. That's not trivial to hang on a wall, and filament starts to seem expensive at this scale! (I did find some bulk 5kg rolls on sale so the whole thing might only cost about $200-300 in material.)

 

[MCH]

Thanks! From what I've read the P1S hotend can only handle so much volume and the print speed can be limited by flow rate, depending on material. But since I'm planning on using PLA it might be worth looking into a 0.6 or 0.8 nozzle and cranking it up. On the other hand, since the piece will be hanging on my wall, I don't want to go crazy with thick layer lines just to save time.

I think I'm going to go with basic white PLA and paint it later. I'm trying to come up with a method of dipping each piece in paint because I think spraying down into the deep wells, or getting a brush in there, will be virtually impossible. Anyone with clever ideas on painting, please chime in.

As it stands each piece will take about 14 hours and the whole thing (I estimate) will take roughly 300-400 hours. This is almost all passive time, I just need to go downstairs and remove completed prints and start new ones about 25-30 times.

I am going to be running the printer in a disused shower, which sounds weird but is ideal because 1) nobody really goes in there and 2) it has ventilation to the outdoors, so the microplastics / fumes are dealt with and I can sit the printer on a solid tile surface.

As for resonances, maybe I should be more worried, but the infill should act as "bracing" every inch or so, and diffusers work with reflection, not absorption, so I think it should be OK. It's important to me to keep weight low. Even with 2 shells, 3 top/bottom layers, and 2.5% infill, the whole thing is going to weigh about 19 kilos before paint. That's not trivial to hang on a wall, and filament starts to seem expensive at this scale! (I did find some bulk 5kg rolls on sale so the whole thing might only cost about $200-300 in material.)

Love your project and I hope one day I can do something similar (I do have a printer, but I don't have yet a place to have it working so many hours without annoying the family, so I use it only for small projects).
Anyways, just chiming to suggest you that if you are going to hang 19kg of plastic on the wall, it might be good to check for some sort of flame retardant grade.

 

[voodooless]

A nice thing to try! This will take a while to print for sure

I also find your way of creating the 3D model very creative. I’m sure this could have been done easier with a bit of coding, but this work I guess

 

[kemmler3D]

if you are going to hang 19kg of plastic on the wall, it might be good to check for some sort of flame retardant grade.

Oh look, here come the fun police.

OK, so after a quick google: Flame retardant filament exists but it's $100s per kilo, so that's right out. If that's a requirement it's not worth it anymore.

However, using a coating of flame retardant paint (more like $50-100/gal) seems plausible and maybe prudent.

This has got me thinking though... none of the wood stuff in my house has flame retardant paint on it...

Maybe I will just buy an extra fire extinguisher.

 

[kemmler3D]

I also find your way of creating the 3D model very creative. I’m sure this could have been done earlier with a bit of coding, but this work I guess

Indeed, a smarter man would have used the parametric modeling features in Blender (called Geometry Nodes) to generate the QRD sequence in 3D directly. However, not having used those features before (except on this project for duplicating the cubes) it would have taken me longer.

 

[voodooless]

none of the wood stuff in my house has flame retardant paint on it

I’m sure that wood is less flammable than most plastics.

 

[kemmler3D]

Small update: received and set up the printer today. It is really shockingly fast compared to printers I've used before. And it seems to do some kind of resonance testing when it does its initial self test, vibrating the motors at a series of increasing frequencies. Really interesting.

I'm printing this to test it out.

Anyway, I have the diffuser cut into 36 pieces in Blender (36 ended up being the most even way to fit them into the printer), I need to add attachment points, (print a scale miniature to make sure I thought through the assembly correctly,) slice everything, tally up the needed filament, order the filament... and it's off to the races.

 

[kemmler3D]

I’m sure that wood is less flammable than most plastics.

Also, yes I would agree most wood is less flammable than most plastic, or at least it seems that way intuitively, but I think the real issue as @MCH mentioned is that the smoke from plastic is more dangerous than wood smoke, and I guess there's a lot if there's a fire. I am well aware that from when I used to import and sell acoustic foam, things hung on the wall are supposed to be flame retardant.

I'll just have to be very careful not to let my wall catch on fire. Luckily the house has very good, new wiring all done to code, so I'm not unusually worried about fire.

 

[DWPress]

On Etsy there's a seller who engineered skyline QRDs out of paper. I haven't finished it yet (too much like my work) as every node needs to be printed, carefully folded, glue the tabs then adhere to a base. I thought it a brilliant idea though and totally accessible DIY for most.

 

[voodooless]

I’m actually surprised how much more expensive fire retardant filaments are. With injection molding the V2 plastics aren’t that much more expensive.

Another issue may be the wear on the nozzle with these materials: usually they add carbon, and this is a lot more abrasive.

 

[kemmler3D]

On Etsy there's a seller who engineered skyline QRDs out of paper. I haven't finished it yet (too much like my work) as every node needs to be printed, carefully folded, glue the tabs then adhere to a base. I thought it a brilliant idea though and totally accessible DIY for most.

Accessible in the sense of not needing special equipment, but an N53 made of paper would probably involve something like 12,000 folds (~4 per well), so not exactly a quick afternoon project. If you wanted to do an N13 or something, proably doable. Would it be functional enough just made of paper though? I think you at least need a stiff material to reflect the sound.

I’m actually surprised how much more expensive fire retardant filaments are. With injection molding the V2 plastics aren’t that much more expensive.

Another issue may be the wear on the nozzle with these materials: usually they add carbon, and this is a lot more abrasive.

Wear would be a concern on a project this big for sure. The Bambu printers come with stainless steel nozzles, but even those wear out after a while, you'd probably want to get one of the hardened steel ones at least, and even then I imagine you'd go through a couple.

 

[DWPress]

Would it be functional enough just made of paper though? I think you at least need a stiff material to reflect the sound.

Stiff stock like tagboard and coated with multiple layers of spray enamel would be fine. The surfaces need to be solid and reflective but not necessarily full density solid. I've played with folded well diffusors with book board (cardboard chipboard) and they worked fine according to measurements.

Will your 3D printed version be solid material or just the shell?

 

[kemmler3D]

Stiff stock like tagboard and coated with multiple layers of spray enamel would be fine.

Yeah, would believe it if you used something pretty thick.

Will your 3D printed version be solid material or just the shell?

2 shells on the sides and 3 shells on top, which gets you almost a millimeter of solid plastic on the top (0.3mm layer height)... plus there's a lot of internal "bracing" via infill. So very far from solid but not completely hollow. To give some idea, you could probably break one of these things with your hands, but you'd have to put serious effort into it.

 

[Curvature]

Can you please make before and after measurements once the project is finished?

 

[kemmler3D]

Can you please make before and after measurements once the project is finished?

I definitely will, although I am not sure how visible a single large diffuser will be in typical measurements. I don't *think* my room has super-strong resonances in the frequencies it's going to affect, and I don't know that it would affect overall RT60 much?

To the group: Does anyone know exactly what type of measurement will tend to best reveal any differences from a diffuser? Will be happy to do them as long as it just takes a UMIK...

 

[fluid]

You might want to have a read of Ron Sauro's papers
https://nwaalabs.ipower.com/Papers and Publications.html

You can also read D'Antonio and Cox's book on Absorbers and Diffusers.
https://www.amazon.com.au/Acoustic-Absorbers-Diffusers-Theory-Application/dp/0415471745
The hard cover is a little overpriced at Amazon , but google provides a way to see the information if your wallet doesn't like the price.

In your own room you can measure the impulse response before and after and try and analyse it, you might see some changes as diffusers are also absorbers. It's not so easy in room to measure the relative phase differences at high frequencies which is why it can "sound" more different than the measurements might suggest, for good or bad depending on where you put the panel.

 

[Prana Ferox]

Curious how this works out. If you make the shape too hollow without enough infill, it will sound like a cowbell when struck. More walls and infill = more filament and print time, of course. Plus the pieces will need to fit together pretty positively. I wonder if it would be more practical to use 1/4" plywood or similar as a backer and attach the pieces to it to hold the bulk weight.

 

[sq225917]

If I was looking to make one I'd make it from 2" square section Eva foam, just measure and cut with a hot wire. Arrange in grid and spot glue as you go. Gotta be the quickest assembly method, still very light.

Plenty of places will cut 2" sheet foam for you.