Making Your Own 2D Diffusors


I've been wanting to try some diffusor panels behind my Quad 988's (to tame the back wave), but have been put off by the cost of the industry standard 2D diffusor panels. So, I've started researching making my own in a similar style. I think I've found a solution that saves roughly 75%-80% of the cost and requires just a table saw and some time to put it together.

For the purpose of comparison, let's consider the specs of the leading product. They are roughly 24"x24" in area and 6" deep, weigh 4 lbs, and comprise a 12x12 matrix of variable depth 2"x2" columns made from an made from extruded polystyrene and painted white. Based on these dimensions, and assuming an average column depth of 3", yields a polystrene foam density of 3 lbs/cubic foot.

Given my adversity to the cost of these diffusor panels ($150 each), I set about looking for a way to make their equivalents. At first, I thought I could maybe make a mold, buy some liquid pour foam, and fill the mold multiple times. However, the pour foam is extremely sticky and the mold has very complex geometry, and it would be nearly impossible to seperate the two unless you used either visquene coating or motor oil (neither of which are workable in this case).

Instead, I found a source of solid foam sheets that are cut to specified thickness and can be ordered in densities ranging from 4 to 20 pounds per cubic foot. The foam is described as "fine-celled high-density polyurethane modeling foam", and the manufacturer tells me it can be cleanly cut with a table saw, and finished with paint, polyurethane or lacquer. (Note that both the #7100 and #6700 foams are applicable). By buying a 24"x96"x2" sheet of 4 lb/cf foam for $150, one could send it through the table saw 24 times to make 2"x2" columns, and then cut the columns into various lengths (depths). A 2'x2' wooden frame could be constructed with a 1/4" plywood backboard, and the frame could be filled with the variable depth columns, gluing each one to the others with contact cement. A formula/schematic for assembling a random combination of columns (5 different depths) is available online. One could either assemble the diffusors within the frame and then take them out of the frame once the glue has dried or leave them in the frame and stretch acoustically transparent fabric over the frame to make them more WAF pallatable. In the first case, one could make 5 panels for the cost of one commercial panel, in the second case 4 panels.

All this looks good to me, and I may give it a try soon. The foam manufacturer has a minimum order of $250, which makes too many panels for my room, so I may go at the project with a fellow audiophile. Also, if you have no foam manufacturer nearby, they could cut sheets to a shipable dimension (e.g. 2'x4') and send you multple sheets. Interesting, huh?
peter_s
Peter s,

"...a random combination of columns..."

Neither the heights of columns nor their arrangement is random. They are determined by a precise mathematical formula in Number Theory.

"...stretch acoustically transparent fabric over the frame..."

Best Regards,
John

My understanding is: That will reduce the effectiveness of the unit.
Nsgarch - I'm not sure that the foam is meant to be absorptive due to open pores. RPG paints its diffusors, and I would imagine that this clogs up the pores. RPG does state that "the solid expanded polystyrene core of the skyline provides useful low frequency absorption", but also states that "the large prime number design offers the acoustic industry's largest reflection number density". Unless you are trying to make an "abfusor", I think you are after a reflective surface but with multiple angles of incidence and reflection, and not overly concerned with absorption.

John - you are right, the combination of columns is not random, and I was just being lazy in describing this. RPG uses "advanced primitive root number theory to design the most powerful two dimensional omnidimensional diffusing surface in the acoustical industry". Are their any skilled mathematicians out their (equally skilled in normal communication) who can explain what primitive root number theory is? I don't know how the link I provided arranges the various columns, but it is not purely random, as can be noted by the distribution of various column heights (at the bottom of the page). My hope is that it came directly off a Skyline, but I have no way of knowing!

Anyhow, I've referred to the discussion of diffusors in F. Alton Everest's "Master Handbook of Acoustics", specifically the discussion of "reflection phase-grating diffusors", or which the Skyline seems like a variant. At the time of writing the 4th edition (2001), the book seems to prefer quadratic residue diffusors to primitive root diffusors, and notes that diffusors built with separators between wells are most effective. The skyline doesn't use seperators, nor do the popular Auralex TFusors, but the RPG Omnifusor does. I'm not sure which of the 3 design is actually best. The RPG products are more evenly 2D than the TFusor, which seems to have less 2D effectiveness, but the TFusors are certainly most affordable. Back to the point: it would be great to know if the DIY design on the link is the same as the Skyline. It certainly isn't purely random, and recipes are available in Everest's book for both quadratic residue diffusors (which are symmetrical) and primitive root diffusors (which are not symmetrical, like the skyline and the DIY design). That just leaves me saying "hmmmmm".

Regarding covering a diffusor with fabric, yes I've heard that this compromises its effectiveness too (e.g. on the Audio Asylum acoustics forum). But the way it was stated seemed like here-say. Can anyone point to a basis for or against covering (seemingly ugly) diffusors with acoustically tranparent fabric?
I did what you are describing using styrofoam. You can purchase styrofoam in 2"X2"X12" sections then cut to the desired length. The drawback with styrofoam is that you can not paint it. I also found a primative root mathematical formula that supposedly provides the optimum length and placement of the columns. I think it works quite well but with most things audiophile it is difficult to measure. If you are interested let me know and I can send you more information. Good luck.
Peter s,

The following "translation" of your primitive root link may be easier to follow: Let p be a prime number and let b be any number greater than 1 but less than p. Divide each of the following p powers of b by p: 1 (=b^0), b, b^2, b^3, ..., b^(p-1). If there are p-1 distinct non-zero remainders, then b is a primitive root of b. Otherwise, it is not a primitive root of p.

For p=7 and b=3, the 7 powers of 3 are 1, 3, 9, 27, 81, 243, 729 and the corresponding remainders are 1, 3, 2, 6, 4, 5 ,1. Since 6 (=7-1) of these remainders are distinct, then 3 is a primitive root of 7. However, if b=2, the 7 powers of 2 are 1, 2, 4, 8, 16, 32, 64 and the corresponding remainders are 1, 2, 4, 1, 2, 4, 1. Since there are only 3 distinct remainders, then 2 is not a primitive root of 7.

John
Sgunther,

You can paint them with latex spray paint, or any latex based paint. I have done this on several of my sons school projects.

Richard