Hi guys,

I've been reading the posts around here for a while, and I figured you'd be interested in this (my apologies if it's a repeat, but I did not see it posted previously).

At this link you will find the patent for RPG Inc's BAD Panel, with detailed drawings and lots of detail. It's nice to have a printed copy on hand, so I utilized my printer.

The build does not look complicated -- just time consuming.

Also, here's some additional information that is not included in the plans that I found by rooting around on the internet:

1) In the plans, it looks like the binary holes are "rectangular." However, it looks like RPG is actually building these using 1/2 inch or 5/8 inch circular holes instead of rectangles, probably for ease of manufacture. The measurements on their site are for the 1/2 inch hole variety.

2) The 1/2 inch holes seem to offer both the best absorption and the best diffusion.

3) I believe I read on RPG's Website that their panels are made from high density fiberboard with a density not less than 90 pounds pcf. That's some heavy stuff, and I have not been able to find a good match at Home Depot or the like. According to the patent, in an ideal world, the absorping parts of the panel would be 100% absorbent and the reflective parts 100% reflective, but they do acknowledge that in the real world, we'll obviously have to settle for less. Still, they are making them from 90 lb. pcf material, so I would think that's a good target.

Enjoy: Patent US5817992

Some additional info from the patent document that is of interest:

1) Figure 7 and Figure 8 represent inverses of one another, with the reflective and absorptive parts switched. The doc mentions that the inverse panel is done to lessen the effects of lobing which come from repitition. Perhaps the barker sequence could be used?

2) There's less work involved with the pattern shown in figure 4, but diffusion is cut off at 4Khz.

3) The pattern shown in figure 6. looks like the standard RPG pattern. It is more work, but has a 9Khz cutoff.

The diffusion coefficients shown on the rpg website and in the patent doc do not extend throughout the entire frequency range. It would be nice if we could dig up a document that shows the diffusive properties above 5Khz so we can better understand how well this thing diffuses the higher frequences.....

We know the standard RPG binary pattern has a high frequency cutoff of 9Khz, but I have not been able to verify the diffusion coefficient that high up.

Some previous threads on the subject:
DIY Binary Amplitude Diffuser anyone?
Flat Fusor

My apologies if I should have posted this in one of the previous threads. I couldn't really figure out which previous thread was most appropriate and the plans seemed worthy of their own thread, but I am new here and do not know the standards yet.

I find interesting that the grating is not a rigid thing like a piece of wood or
metal with holes in it, but either a paper/FRK facing with holes stuck to the
fiberglass or else the pattern is 'painted' onto the fiberglass through a mask.

One way to DIY would be to make one panel out of plywood or metal with the
inverse hole pattern drilled in it and then use it as a mask to spray paint the
pattern onto the fiberglass.

I wonder if there is any airspace between the decorative fabric covering and
the fiberglass/BAD facing.

Paul P

No, it's perfectly fine that you started a new thread in this situation. The previous two are just good to reference for people who are interested in the subject but may not have been aware of the previous discussion.

Perhaps there are some differences between what's found in the patent and the way rpg is actually building these things.

I pulled the following text from the word doc at this link: www.rpginc.com/products/badarc/BAD_Arc_Fabric_CSI_Spec.doc

A. Core: Semi-rigid fiberglass, density not less than 6 pounds per cubic foot and not more than 8 pounds per cubic foot.

B. Template: 1/8” thick high-density fiberboard, density not less than 95 pounds per cubic foot.

C. Scrim: White, open weave, fire-rated cloth (required for light-colored fabrics).

D. Fabric: Open weave, Class A, polyester, without backing such as pattern FR701-2100 by Guilford of Maine (Grand Rapids, MI). Color per architect selection.

E. Frame: ¾” thick Plywood.

Based on info I've found, it looks like the standard BAD Panel template is made from 1/8 inch HDF @ 95 lbs pcf. 1/2 inch holes are drilled into the template mathing the pattern in the patent. That is, unless I'm totally misunderstanding what I'm reading!

Here's a helpful doc with some stats, including the diffusion coefficients higher than 5Khz that I was looking for earlier.

http://www.silentsource.com/PDF/badpanel.pdf

Hey guys, I just posted this link in another thread on its own, but I wanted to post it in this thread too because it's obviously relevant. In fact, if you're planning to build your own BAD Panel, there is a nifty template in Figure 1B that you could probably have blown up to 2x2 size, then lay it over your plywood (or MDF or whatever) and drill away, no need to draw out the template. HUGE TIME SAVER!

Unlike the drawings in the previous patent attached above, Figure 1B in this patent has the proper circular drilling regions instead of the odd shaped rectangular regions.

See Figure 1B in this document: Patent US6112852

Also of interest is figure 2, which would be a WHOLE LOT EASIER to build. I mean sheesh, it's like 4 rip slices on the table saw, but unlike the more complicated BAD pattern in Figure 1B, I cannot find any measurements for what I'm seeing in figure 2. Too bad, as it looks like it would provide some 1D diffusion in addition to absorption.

Guys, I'd like to get your input on this idea. I'm thinking that we don't need to go through all the trouble of drilling all these holes unless we need 2D diffusion. If 1D diffusion is desired (or good enough), all we need is a series of patches of hard material of various sizes placed over our absorbers, perhaps in a pattern similar to what is shown in Figure 2 (in the document I linked to in my last post).

Here is some info that I pulled from the RPG Website which clarifies that the BAD panel is for 2D diffusion.

Quote:

Originally Posted by RPG Website Text

Acousticians have known for some time that surface treatment consisting of patches of reflective and absorptive areas provide some measure of sound diffusion. The problem has always been that the normally used patch size of roughly 2' x 2' only provides diffusion up to a frequency whose wavelength is approximately equal to the width of the panels. Thus a 2' x 2' panel only provides scattering up to about 500 Hz. Higher frequencies are specularly reflected from the reflective areas. Therefore, if one were to reduce the size of these patches to roughly 5/8Ó one could raise the upper limit of useful scattering. The question is what is the best distribution of reflective and absorptive areas? The answer is easily understood once a mapping is done in which the reflective areas are a one in a binary sequence and the absorptive areas are a zero. Once this digital mapping is recognized, it is apparent that one approach would be to use an optimal binary sequence of zeros and ones. In this case one would obtain a flat power spectrum and achieve uniform scattering. If omnidirectional scattering is desired one would need to derive an optimal 2D binary sequence. The 2D sequence of zeros and ones define the locations of the reflective and absorptive areas. In 1998 RPG developed the first binary amplitude template shown in Figure 1.

So, the BAD Panel is a 2D binary sequence of holes and solid areas, which will produce 2D diffusion. However, if all you need is 1D diffusion, you need a 1D sequence of of holes and solid areas, which is just a series of vertical slats! Easy as pie... Now, if I can just find me some measurement data.

Diffusion (acoustics) - Wikipedia, the free encyclopedia #Maximum_Length_Sequence_Diffusors

What am I supposed to be looking for at that link, Lupo?

Specifically, what I am trying to figure out (assuming your post was a reponse to my last one) is what sort of spacing/sequencing would be required for a 1D array of hard surface laid over a field of porous absorption in order to gain useful 1D sound scattering while maintaining low frequency absorption. See figure 2 in the last doc I linked to.

Was trying to be brief.

The 1D binary diffuser is typically maximum length sequence diffuser. Examples of MLS's can be found in various places. Pick one with a length that suits you! A quick search gave this 15 length MLS as an example: 1 1 1 -1 -1 -1 1 -1 -1 1 1 -1 1 -1 1. The Book have an example with this 7 step sequence: 0010111.

It was supposed to give a hint as to the underlying construction of 1D binary diffusion. I bet there's a RPG patent for that as well. It usually is. :D

Using the 0010111, we could assign "absorptive" or "reflective" characteristics to either 0 or 1. So for the sake of conversation, let's make 0 reflective and 1 absorptive.

Could I take this pattern and apply it to a panel of any width? If so, let's apply it to an absorption panel 2 feet wide. Using a 0010111 would result in this setup (all sections described below run the full vertical length of the panel):

1) At the right edge of the absorption panel, we have a vertical slat of reflective material covering the face of the panel which is about 6.8 inches wide.
2) Next, we have an open absorptive region where the panel absorption material is uncovered and clearly visible, which is about 3.5 inches wide.
3) Next, we have a reflective region 3.5 inches wide.
4) Finally, we have an absorptive region about 10.25 inches wide.

Am I thinking of this correctly?

Excellent thinking!

Add this bit from the wiki link into the equation: "The width of the strips is smaller than or equal to half the wavelength of the frequency where the maximum scattering effect is desired."

You'll probably want each of the units in the sequence to be thinner than 3.5". I can't think imperial, so i'll do it with meters: 3.5(inches) times 2.54 is 8.89 centimeters, bringing it to metric land. Then 343(meter/second ~air velocity) divided på 0.0889 meters (width) divided by two (half wavelength) equals 1929Hz. The upper limit for the diffuser in question.

MLS brings us back to the first diffusion attempts by Manfred Schroeder! :-)

Could you take a look at Fig 2 in this doc: Patent US6112852

I'm trying to figure out what prime number is being used here. I copied the the pic into MS paint and started chopping it up trying to determine the unit being used, and all I've been able to figure out is that it's some number bigger than 15, and probably bigger than 31. This, of course, is assuming that every patch measures a width that is a multiple of the same common unit.

The reason I'm asking is because it's probably a pretty safe assumption that Fig 2 is a legitimate sequence for a good diffuser, since that same doc contains legitimate BAD Panel sequences. So if I can figure out the prime number, I can determine the common unit, and from there can apply it to my specific panel.

Well, I never figured out the prime number, but what I did do was copy the pic of fig2 from the document and paste it into MS Paint along with an online ruler. Then I scaled both the ruler and the pic up to a 21 inch width (the actual absorptive area in my panel), and I ended up with this sequence:

1) 2.25 inch absorption
2) 1.75 inch reflection
3) 6 inch absorption
4) 3.75 inch reflection
5) 2.25 inch absorption
6) 1.5 inch reflection
7) 2 inch absorption
8) 1.5 inch reflection

Upon second thought, I don't think Fig2 is an MLS sequence at all, since it does not follow the conventions.

I found an MLS sequence for m=4 that is 000100110101111, and I think I'll be using that, since it gives me a common unit of 1.5 inches, which should be effective at diffusion up to around 4500Hz, and reflective above that. Because I'll be using 1/8 hardboard for the reflective areas, I'm guessing sub frequencies will pass right through it and into the absorption material behind it.

Are there any reflective materials for which we know for certain the lowest frequency that will reflect? If so, I could calculate for certain which frequencies would fall into the following 3 areas:

1) absorption
2) diffusion
3) reflection

Howdy!

Maximum length sequences are powers of two minus one, not primes. Ie, a seven length sequence is based on (2^3)-1 and (m=4) 15 length is 2^4-1. The next steps are 32-1, 64-1, 128-1 and so on.

Don't know about the material properties. Seems like a good idea to find one that serves dual purpose! Although it probably doesn't need to be that transparent at all, as bass will have little problem finding its way around relatively small obstacles.

I remember looking at those drawings once, there was something about the wording that made me suss.

If you go to the section titled 'brief description of drawings', it says "figs 1A and 1B show schematic representations of masks"

It all depends on how you interpret the word representations...I took it to mean an 'example' or 'illustration' rather than an 'exact must copy' thing.

My main interest is because I have a mate with a cad router, so if I can get an exact plan then it is a piece of piss. But as I said, I interpreted that word a certain way.

Andreas, get cracking with the pen and pencil would you and whip up a prime sequence for me??

Now, if it could be confirmed that those figures are indeed an exact representation......

I'm thinking the same thing Terry. I have a friend with access to a small laser engraver at his work. It can only handle 18" wide and 12" tall material, but I could break my drawings into sections that size, and give him sheets of veneer to cut out on the engraver, that I could laminate onto some 3pcf material. I was going to use the patent drawings as my guide.

Otherwise, it's the old .5" bit and drawing it out by hand on 1/8" MDF. FUN!

You only have to do one, it is after all a mask. According to the patent
the pattern is painted on the fiberglass or made out of paper.

Paul P

I'd say the drawings in the doc are accurate. here's why. if you carefully read the patent I linked to originally, it says this:

"Fig 5 shows a two dimensional MLS array where m = 6.
Fig 6 shows a two dimensional MLS array m = 10."

You will notice that Fig 6 matches the pattern in the drawings exactly.

Furthemore, the doc goes on describe the high frequency cutoff of each pattern, and earlier in the doc it even gives exact instructions for converting a 1D mls sequence into a 2D array. Yeah, I'd say those pics are real.

In the doc where it uses the word "representations," I'd say that's just a way of pointing out the obvious: that it is, in fact, a schematic representation (or depiction, or illustration) of a mask, which is true.

According to the patent, you're correct. But one thing that it says even in the patent is that in an ideal situation, the reflective patches would be 100% reflective and the absorptive patches 100% absorptive, so I think it would be difficult to achieve acceptable performance (even accounting for real world limitations) using paint or paper. This may be why the panels that rpg is selling apparently use fiberboard of a high density with holes drilled into it.

The pattern appears to be the ideal perforation pattern for absorbers with
a perforated front, à la BBC. If you use something thin at least you can
drill a lot of the panels at the same time.

Paul P

You may be right, will have another look. For example, I cannot recall seeing that stuff you were working out with the blank spaces etc. the mls sequence? so maybe I was looking at something else, or didn't study it properly the first time.

Re the 'first attempt by schroeder' the mls sequence, there is something in the alton everest book that (from meory) 'shows' that the mls sequence is so limited it is not worth pursuing.

I will have to look it up and quote it for you. (I could be wrong too, we'll both see when I do it)

LR, I find it easy to work you out. You are determined (just for the hell of it) to make at least ONE representation of every type of room treatment there is, aren't you.

Kinda like a museum of room treatments!

So hurry up with this one wouldya? (how are the subs going??)

You mean like yours ?

Heck, even your house could be in a museum ...



Paul P

The space sequence stuff I was trying to work out were from the second patent i linked to, further down the page. There is a pic of Fig2 in the second doc with what appears to be a one dimensional MLS sequence of absorptive/reflective patches. However, after measuring it out, I don't think it's an MLS sequence at all. I'm really not sure what it is, although it is described in general terms later in the doc.

To make the most sense of either/both of those patent docs, I had to print them out and read them -- both docs have several pages of very detailed text at the end, after all the pics.

I'm interested in whatever you can find about MLS sequences not being effective in 1D binary patterns....

p 292 of the h/book of acoustics re the mls diffusor

'experience has indicated that reasonable diffusion results over a band of plus or minus one-half octave of the frequency around which the diffusor is designed...consider an mls diffusor with a sequence length of fifteen. A design freq of 1000 hz gives 1/2 wavelength of 7.8 in, 1/4 wavelength groove depth of 3.9 in. A single period of this diffusor would be about 5 ft in width and effective from about 700 hz to 1400 hz.'

That to me means not much bang for the buck in terms of space and diffusion.

I guess historically it appears as it was the first (and proof) of what schroeder was trying to develop.

I think the technology has moved on considerably since then and personally see little point in revisiting that particular cul de sac.

Hope that helps.