DIY quadratic diffuser

Man, it is absolutely amazing when you punch in a word into a search program. I punched in DEFRACTAL.

http://www.acousticdiffusers.co.uk/prices.htm

Just amazing how many people are making these.

Heres some sideline info. Not totally related to DIY, but still related.

http://www.mech.canterbury.ac.nz/info/labs/Reverb Room.htm

Damn, I knew someone would use chemetals(laminates) before I could get to it

http://www.pcimillwork.net/display/p_sony_4.html

Thats what I want mine to look like.....dream on rick!
fitz
PS. These interior fixtures were built by Commercial Architectural Millwork Co.
I work for one. I am a detailer, and draftsman, and see this level daily. The principles are easy. The actual millwork is complex. Large cylindrical diffuser sections are most likely made of either sonotubes, or laun ply, and you can order them bare, and post laminate them. Sonotubes are 5/16" carboard tubes used for concrete form work, and are also available. Any diameter, any length to 20'. The quadratics are most likely rectangular MDF tubes, with side dados for knockdown assembly. That allows prelam of the well bottoms and sides, with fronts being postlamed. Lamination sequence is the real trick when it comes to this stuff. Usually, we prelam whole sheets, cut parts so as to dado through the laminate, rabit the edges, assemble and post lam the faces. I'll post a section drawing soon. Well widths and depths are the acousticians responsibility. The design firm would supply the basic section drawings, my job is to detail it for construction. Everything. Cabinets, tables, millwork, trim, doors and jams
and anything dictated by the millwork package.
 
Last edited:
Re: square versus honeycomb pattern

I totally agree that the honeycomb pattern has some cool advantages in terms of pipe spacing, but I have no idea how to calculate the length of the various pipes in that case. All of the calculations that I've seen assume a square grid which I would be approximating with circles within each square.

Re: weight

I don't think it will be a big problem. I'm only talking about putting a 1/2" deep plaster plug on the outside end of the pipe. Hmmmm, I wonder if I could use a shrink-wrap film instead? That would be even lighter, but possibly more expensive. Maybe a varathane varnish could end up working as an end, kind of like blowing bubbles. That would be great if it worked - just dip the end of the pipe in the varnish and let it dry... Anybody have good thoughts?

Re: cutting
I'm really leaning toward plastic sprinkler pipe -- REAL CHEAP! ;) Gotta test cutting methods to see what works.

Rick -- the metal versions look very cool, but I can't help thinking that the 2-D versions would give a lot smoother diffusion than 1-D strip versions. I also think (as per the BBC paper) that even covering 20% of a roof with these 2-D beasts would give a huge diffusion boost.

-lee-
 
that even covering 20% of a roof with these 2-D beasts would give a huge diffusion
Click to expand...

Hello laptop, I havn't read the BBC stuff yet. No time. So I don't know what your refering to. I'll read it tonight. But in regards to the pipe, they make pvc plugs for this(sprinkler pipe? right). Just solvent weld them in place. If I'm thinking on the same plane as you.
fitz
 
I've seen caps for the ends of the pipes, but haven't happened to see plugs -- doesn't mean they aren't readily available, just I haven't seen them. I don't like the thought of the caps because that would increase the spacing between the pipes. If the plugs were nice and flat, they could work out well.

The spray foam stuff could be a great idea - just squirt a tiny bit in the end, then cut it flat with a pocket knife after it dries. Sounds quick and easy. You still might want to cover the end with varathane to avoid absorbency, but thats an easy quick process too. With hundreds of pipes to do, I'm looking to make a very simple workflow, and this might be a great solution - thanks!

-lee-
 
I'd love to use the math sequence from the unplugged site, but I'm not sure how to translate the 1-d sequence properly into a 2-d square.

I'm thinking I'll use the 2-D sequence specified in the BBC paper (http://www.bbc.co.uk/rd/pubs/reports/1990_15.html). They used 45mm square tubing (just shy of 2" for us metric-challenged folks ;) ). One advantage to this size tubing is it is very reasonable to make a good sized diffuser quickly.

I haven't been able to find square pvc tubing yet -- all I have found so far is the round. I think I can find 2" tubing -- but I need to check out how the price compares to 3/4" sprinkler pipe.

Off to home depot for another look!

-lee-
 
Cutting sprinkler pipe

Went to the hardware store, sprinkler pipe is
pretty thin walled and not too hard, I still think
that if I were to cut a bunch at home I would
use a rotary blade of some sort. Power mitter
or some form of table unit that I could rig against
or jig to make quick square cuts with length control.

As for blades I would try a plastic specific model,
I saw several. they were fine toothed at about
180 to 200 for a 7 1/4". The most interesting one
was made to cut plastic siding. again sprinkler
pipe is pretty thin. I like the idea of cheap, available
and simple, There are a ton of expensive, exotic
choices out there already.

Have you thought about plastic taping one end,
setting these things on end and just pouring
liquid two part epoxy down them. If you use the
right tape when it is set you just pull it and are
done, no finishing, I think it would bond ok.

Good luck, Bob
 
we have a square plastic cable duct in OZ - I'm, sure you must have the same. It has a clip in lid and comes in varying sizes.

cheers
JOhn
 
John Sayers said:
we have a square plastic cable duct in OZ - I'm, sure you must have the same. It has a clip in lid and comes in varying sizes.
Click to expand...

Absolutely -- I'm here in Southern California, and I suspect that I can find sources for all sorts of stuff.....

On the other hand, what I'm cooking up here in my demented little mind is a project that I'd like to be usable and transferable to all sorts of home recorders in all sorts of areas and economic situations. That means -- super readily available and *cheap*.

I found some stuff in my local Home Depot that might work out just fine. Its square PVC tubing. The dimensions are slightly larger than the ones used by the BBC -- so it may not diffuse quite as well at high frequencies, if I'm not mistaken. On the other hand, they were getting extra response up there, so I think some second and third order interference effects were probably helping the higher frequencies.

The tubing is actually used for rain gutter downspouts and should be available all over. Its about 2 1/4" square. (sorry about the english units, John! ;) ) Its about $6. for a 10 foot length.

I'm still working out the length dimensions, but right now I'm thinking that I can make a diffuser with 4 of these 10 foot long tubes cut up. Because they are square, they'll lay out and glue up very easily -- much more easily than the round stuff. For glue I'm undecided between normal old PVC cement and hot glue. The hot glue sticks are solid as soon as they cool - very nice and quick. The PVC glue would be a bit cheaper, but not much. The PCV glue actually melts the PVC slightly, so the bond should be outstanding. I've never actually tested the hot glue on PVC, so I'm not 100% sure how well it would work. I'll probably go with PVC cement.

Capping the ends is still an interesting dilemma. I found the expanding foam insulation for about $6. a can. There are different kinds, both flexible, and a kind that hardens to the point where you need to sand and paint it. On the other hand, I like the recent input about putting the tape on the end. I don't know how cheap the 2-part epoxy is -- the stuff I've used is pretty expensive for small volumes and we're gonna need a bunch to fill the pipe end. On the other hand, maybe we could dribble something else down the pipe end to make the seal after taping, like varathane. Or maybe tape the end and spray in the insulation - no cutting/sanding needed that way. Or maybe cut a wood plug to fit the end and glue it in.

Lots of choices -- anybody else got more ideas? Keep them coming, this might be a great tool in our acoustic arsenal!!

-lee-
 
foreverain4 said:
i am still not quite understanding how this thing works. there are no angles to diffuse sound. am i missing something?
Click to expand...

Check out the BBC article for huge details including math that is currently stumping me, but lets see if I can describe it this way --

Imagine sound as a wave. Think of dropping a pebble into a pond -- you can see a set of round spreading waves going away from the pebble. Now put a set of sticks of varying lengths in the way of the wave. Orient the sticks so that they are lengthwise to the wave. I'm gonna try an ascii art version below -- no promises it will work well...

wave sticks
> xxxxxxxxxxxx
> xxxxxxxx
> xxxxxxxxxxxxxxx
> xxxxxx
> xxxxxxxxxxxx

When the wave hits the sticks, there will be a lot of reflections going in lots of different ways, all interfering with each other and spreading all over the place. Some of the reflections are early, others are later -- all spaced out. This causes reflected waves at different times.

This beast does the same thing, only in two dimensions. Think of a square peg board with a bunch of square pegs of different lengths in it. The BBC report has some pictures.

The key to the best scattering is to get the size and shape of the "sticks" right. If you do that, you can get good scattering across lots of different frequencies of waves. This scattering diffuses the sound and makes everything nice and smooth across the frequencies without problemmatic reflections that all arrive at the same time, or frequencies that build up too much in one place.

At least that's MY understanding of this stuff! ;)

-lee-

Darn --- the ascii art didn't work out like I was hoping. Turn the sticks in the art 180 degrees and it looks like what I was trying for.
 
Last edited:
what do you guys think of this? it will be made from hardwood maple. the ends are pieces of crown molding that will be ripped to size, given that the original piece was about 6 " wide. do i gain anything by putting this at the end of each slot? if not, i wont bother with all the work. i have a bunch of this moulding i got for free. just thought i would gather more info.
 

Attachments

  • another diffusor3.jpg
    another diffusor3.jpg
    4.9 KB · Views: 218
foreverain4 said:
what do you guys think of this? it will be made from hardwood maple. the ends are pieces of crown molding that will be ripped to size, given that the original piece was about 6 " wide. do i gain anything by putting this at the end of each slot? if not, i wont bother with all the work. i have a bunch of this moulding i got for free. just thought i would gather more info.
Click to expand...

Looks interesting! How are you determining the length for each piece? Is it strips or 2-D with other blocks in there too?

I don't know about the moulding -- offhand it seems like it would scatter very high frequencies, *but* I don't know in what way -- the heavy duty math is so that the scattering is in a smooth, non-peaky way. The moulding *could* end up causing very high frequencies to all peak at the same time -- I just don't know.

I suspect the best solution would be like a fractal -- a 2-D board where the top of each block is a tiny 2-D reproduction of the bigger board.

-lee-
 
gatorhaus said:
Sure wish I didn't throw those egg crates out DAMN. lol....
Click to expand...

:::laugh:::: Yah, that thought occurred to me too. How can something this cheap help sound as much as it promises?

I believe eggcrates would give a tiny amount of very high end absorbtion, and an even smaller amount of high end scattering. The repeating pattern could end up creating nodes of frequencies that were worse than the original. As I understand it, the strange 2-D pattern is supposed to break things up smoothly, without creating reinforced nodes.

After all -- its GOTTA be good. We have incomprehensible math to support it! ;)

I really hope to be able to rent some tools so that I can *measure* my room before and after tweaking. I want to avoid just thinking it sounds better because of the audio placebo effect.

-lee-
 
OK, so here's my latest musings...

I can't find the exact lengths of the tubes used in the BBC paper, but it looks like 2 "units" are the same as one width of one block. So if I use 2 1/4" square PVC pipe, then one "unit" is 1 1/8". The numbers in the BBC paper range from 0 to 4 for height, but it looks like 0 is actually full height, and 4 is deepest, or least amount of height. So 4 would be zero inches, 3 would be 1 1/8", 2 would be 2 1/4", 1 would be 3 3/8" and 0 would be 4 inches.

I counted up the numbers of each length in the BBC version and it looks like I only need THREE of the 10' PVC pipes (about $18.) to make all the blocks.

I'm thinking that I'll use a power miter saw to cut the pipe. I'm not sure what blade would give the least shredding on the edge. This should let me set up a jig and cut a BUNCH to a given length.

I am still puzzling over the best way to close the end of the pipe. I like the thought of the two-part epoxy, but I think the cost would be prohibitive. Wood plugs would be great, but I'm not sure how to make a thousand of them (!) easily. I really want an easy workflow. Slapping tape on, and dribbling in some varathane or squirting in some spray foam insulation seems like the way to go. I think the foam would be dry sooner.

Once one end is sealed, I can glue the bunch together with PVC pipe cement. I'll probably glue the whole mess onto some soft of very lightweight sheet -- maybe thin plastic, like what has been used in drop down kitchen ceiling lights. The whole assembly should be very very light.

Once the cement dries, I can paint it, then hang it up with velcro strips.

Of course, I still have to figure out how many diffusers versus bass traps versus absorbers to use!!!

-lee-
 
I have been following this thread for a while and I was wondering if the square peices have to be hollow? I could see why hollow might be better from an absorbption standpoint, but if this is to function as a diffusor, couldn't it be made from wood?

Also would you only use this on the ceiling or could it be used on the walls too?

Beezoboy
 
Doh! That sound you heard was me slapping my forehead...HARD! ;)

I believe.... NO PROBLEM with using wood for the diffusors. They'd be heavier, but other than that, cheaper and easier to work with, with fewer steps.

New plans: use 2x2 lumber (actual size is 1 1/2" square) -- each board is only $1.80 at my local Home Depot.
workflow:
1) cut blocks to length
2) attach blocks onto backing board
3) paint
4) hang

Cost per diffusor panel - about $20. (!!!!) Yes, boys and girls twenty bucks for the audio equivalent of what some companies charge $300. for! (note: this is assuming my new depth calculations which are about double the previous depths. Using the old ones, the cost would be about $10.!!)

OK, so I'm left with new questions:

Question 1) What should I attach the blocks with?
Hot glue -- hardens QUICK
Wood glue -- Cheap, buy it by the gallon, takes much longer to dry (IS IT STRONGER???)
Screw/nail through backing board?
Use electric nail gun or staple gun through the backing board?

Question 2) What should I make the lengths?

With a 2x2 piece of lumber, the actual dimensions of the block would be 1.5" by 1.5" by length. This is slightly smaller than the BBC prototype. If I understand it, the high frequency performance is determined by the face size of the blocks, and the low frequency performance is determined by depth. The BBC prototype seemed to use a multiple of the face size for the depth -- but is that a requirement? I'd rather not give up the lower frequency response. The BBC paper doesn't give dimensions for the exact depths of the prototype -- at least not in my reading. It does seem to imply about 200mm (about 8").

I *think* that the real-life bouncing around of the frequencies all scrambled together, combined with the various sides and side lobes, etc. result in a diffusor where some of the exact parameters are more tolerant than might be expected by mathematical analysis of a sound wave arriving perfectly perpendicular to the diffusor. The paper seems to imply a real-world performance better than expected. With that in mind, I'm leaning towards making the blocks 2", 4", 6" and 8" long. Such blocks would be reasonable to work with and assemble. I would expect low frequency diffusion down to something like 860 Hz, and high frequency diffusion somewhat over 3.4 kHz, with significant helpful effects both above and below this frequency band. Does this sound reasonable??

Insane as always,
-lee-
 
Last edited:
Beezoboy said:
Also would you only use this on the ceiling or could it be used on the walls too?
Click to expand...

As I understand it, the diffusor is optimized for sound reflections that come head on. Off-axis performance is less quantified. This makes it ideal for ceilings that are parallel to the floor, and very good for walls which are parallel to other walls. If your walls are not parallel to each other, the performance would probably be not as good -- but still could be helpful.

-lee-
 
Please don't tell me you never even considered using wood!! :P 45mm was the square size and that comes out to 1 and 3/4 inches. Not 1.5 inches.

When you look at the layout they don't give the size but give the number of the blocks. I figure they are 45mm by 45mm, or 1.75" x 1.75". I interpret the diagram as giving the height by number blocks tall.

So to make this diffuser here is what "I" would do:

First you have to look at that diagram and fugure our how many of what cuts you have. You obvously don't want to just try and glue together bunches of blocks, although this could look cool if you alternated woods (Like pine and cedar for color contrast).

Buy a couple 2 x 4 studs and a sheet of plywood or fiberboard. Measure the short side of the 2 x 4 as use that as you block size. What ever that measurement is double it and make sure that it comes to be the width of the long side of the board.

The cut all the appropriate pieces and the plywood to size. Then lay it all out without gluing anything down. Once you get that done start gluing all of the outside pieces down until you get the perimeter of blocks done. I would then let those dry and come back and glue the rest. Then paint or varnish as you see fit.

Definately use glue and don't try nails as this would be a nightmare of flipping the thing over several times. Also this allows you to glue the whole structure to itself by putting glue on the sides and bottom of the block.

I think this project will take a good weekend to do. Once you figure out all the cuts and make one, the rest will go much faster; the first one would really be a learning experience.

I bet you could make one of these from wood for only like 10 or 15 dollars. The table saw and sander it would require obviously offsets the cost a bit. Then there is glue and paint or varnish and most importantly time.

A fun project though. I'll probably try when I get my table saw in a couple weeks.

Beezoboy
 
You must log in or register to reply here.
Back
Top