Triple-leaf question for the experts

Folks, I'm hoping Brian Ravnaas, John Sayers, Rod Gervais, or any other isolation experts here can help me to understand something. As many of you know, on my personal web site I have THIS article from Electronic Musician magazine showing plans to build wood panel membrane bass traps. It has been suggested that this type of bass trap can reduce isolation between rooms because the front panel of the trap creates a triple leaf system with the existing wall. The back of the trap is the interior room wall, and the trap's front panel adds a third "leaf" inside the room. It seems to me this will not reduce isolation between rooms because the third leaf does not reduce the spacing between the existing two leaves. In all of the two- versus three-leaf comparisons I've seen, they show the same fixed distance with either two leaves or a third added between. Clearly, adding a wood panel trap inside the room will not reduce the wall's interior spacing, so it's not clear why having an extra leaf attached outside the existing wall will reduce isolation. But this is outside my area of expertise so I'm asking the guys who really understand the theory behind this stuff. Thanks! --Ethan

hey everybody, i'm really glad to get a chance to post, this forum is always talked about but i've never been a member. Ethan asked me to comment on this thread, so i joined. The basic query is: is a triple leaf as bad when its deeper than the comparable double leaf... Its risky to answer questions concretely without testing them first so i dug for some data. In IR811 the NRC has a ceiling that is OSB/joists/drywall screwed directly, with R19 in the cavity. And the same wall with RC and drywall added to form a triple leaf. The triple leaf in this case is not worse, but also not much better. (if you build a double leaf with the exact same amount of materials, the performance gains are massive). This data both points out the importance of not building triple leaves, and supports Ethans suggestion that adding to a wall to make a triple leaf isn't necessarily bad. That is just one set of data, and i will look for more, and post them if/when they come to mind. Some (many, actually) exampes of triple leaves are found in IR-586, which studies concrete block with added drywall. In that case, the triple leaf has a distinct positive/negative effect (dpeending on frequency), but with lighter walls where the added mass may be significant, the negative effect may not be as pronounced. I'll add more if i can, i hope that's at least a bit interesting.

now to add to that, the situation above has a mass-spring resonance from the triple leaf that falls about at the same frequency as the mechanical resonance of the reference ceiling. ... two resonances in series at about the same frequency aren't always worse than 1, and it is certainly possible to get much worse results from adding the triple leaf than what was had above (which was actually an improvement). So, while clearly double leaf is the way to go, this data supports the thought that adding a 3rd leaf isn't necessarily bad. Still, i would recommend caution when messing with walls... as once you start venturing into the realm of the untested you're sort of asking for trouble (as you don't know what's going to happen). i'll look for more data.

Hi Brian: I added another .gif to the bottom of (scroll down a bit) http://www.bobgolds.com/WallCharts/Quad ... eafSTC.htm

sorry to triple-post. One thing(s) to think about in the case of adding a panel trap to a wall are these: -the panel traps resonance freq depends on the mass both in front of and behind it, so the trap might react differently to being put on a lightweight wall -vs- a massive concrete floor -the panel trap won't probably be 100% of the area of the wall. And as such, the negative effect of the trap won't be as bad (if a negative effect occurs) as if it was over the whole area of the wall. An open door isn't as bad as no wall at all (but its still very very bad, of course, from a sound isolation standpoint). I once entered a spreadsheet at studiotips that lets one calculate the effect of a (for example) poor quality door of XX area in a wall of YY area, that could be used to estimate the overall potential effect of a panel trap in a wall. :)

z60611 wrote:

Hi Brian: I added another .gif to the bottom of (scroll down a bit) http://www.bobgolds.com/WallCharts/Quad ... eafSTC.htm

hey, Z you might change the 2-layers-each-side to the 16" OC example, as the triple leaf was 16" OC. but that's a great graph and very illustrative. if you read the IR761 text, it says those inner panels weren't attached to any studs... if they had been attached, the results would have been alot more disastrous. i have some old owens corning data about this that you can have if you like, but it only goes to 125hz.

Brian: I made a minor edit to that new .gif, to add a Staggared Wood Stud wall.

you might change the 2-layers-each-side to the 16" OC example,

Also done.

Brian, Excellent. Thanks very much for your detailed replies! And thanks to Bob "Z" too. > This data both points out the importance of not building triple leaves, and supports Ethans suggestion that adding to a wall to make a triple leaf isn't necessarily bad. < That's the only way it made sense to me too. But since I don't have a good grasp on the underlying theory, all I could go by was a "common sense" analysis of what might be happening. I tend to see this stuff in elecrical terms, so a two-leaf system is like a passive filter made from one inductor and one capacitor. Adding another leaf inside the wall puts two capacitors of half the value in series where there had been one capacitor twice as large. That obviously reduces the "blocking" effect. Or something like that. So I could sort of see how adding it inside harms isolation. > the panel traps resonance freq depends on the mass both in front of and behind it, so the trap might react differently to being put on a lightweight wall -vs- a massive concrete floor < Agreed. The low-bass trap in my plans has a center frequency of about 100 Hz. It's hard to know for sure, and when I measured a portable version I made at IBM's acoustics lab it peaked at 100 Hz. But reverb room absorption measures at only third octaves, and that hides a lot of detail. None the less, this low-bass panel trap uses 1/4 inch plywood for the front surface, so that's still pretty "far away" from the mass and self-resonance of even thin (1/2 inch) sheet rock. > the panel trap won't probably be 100% of the area of the wall. < Well, it could be a significant part of it. I learned how to build these traps in the late 1970s from the acoustic consultants I hired when I built a large pro studio. These traps are installed as a "system" that goes all around the room. However, the trap types alternate between low-bass, fiberglass-only, high-bass, fiberglass-only, low-bass, and so forth. So 2/3 of the wall (and optionally ceiling) could potentially have wood panel traps. I have to think the high-bass type with only 1/8 inch plywood affects this even less though. Thanks again. If you find any further data about adding a third leaf outside the existing wall please let us know. --Ethan

Just to enter a link to the original thread: http://www.johnlsayers.com/phpBB2/viewt ... 8355#48355 Ethan, I think you didn't ask a question here but just were looking for a confirmation. Brian is mainly active at AVS where you also are very active. Then you ask Brian to come here, who specially needs to register. Brian's reply is the typical but rightful careful scientist's reply:

Brian Ravnaas wrote:

The basic query is: is a triple leaf as bad when it’s deeper than the comparable double leaf... It’s risky to answer questions concretely without testing them first so i dug for some data.

Brian Ravnaas wrote:

This data both points out the importance of not building triple leaves, and supports Ethan’s suggestion that adding to a wall to make a triple leaf isn't necessarily bad. That is just one set of data, ..........

How triple leaf systems behave in practice can be very unpredictable sometimes looking strange. Brian carefully states: "isn't necessarily bad". He did not say: it can't harm TL. That making the total outside measures of a tripple leaf system larger logically will be less damaging than dividing an existing space, but that's still no automatic synonym with it can't harm TL. You TUNE your traps (logically) specific to dominant audio frequencies, and they are designed to maximize resonance. I read between the lines that you want public confirmation that panel traps can't diminish insulation. For double leaf systems they do. How much depends on the mass ratios but it doesn't matter that the existing wall is a very heavy wall or not. For triple leaf systems, as said they can behave very unpredictable, but by definition creating a higher resonance will influence TL. This all depends of a lot of factors. And the relative significance in a project also depends.

Ethan Winer wrote:

In all of the two- versus three-leaf comparisons I've seen, they show the same fixed distance with either two leaves or a third added between. Clearly, adding a wood panel trap inside the room will not reduce the wall's interior spacing, so it's not clear why having an extra leaf attached outside the existing wall will reduce isolation.

Here you start from the assumption of an existing wall assuming you just add something, without altering the existing wall. Well whether you work with a double or a triple leaf system you alter the properties of the existing wall, because you create a new system.

Brian Ravnaas wrote:

-the panel traps resonance freq depends on the mass both in front of and behind it, so the trap might react differently to being put on a lightweight wall -vs- a massive concrete floor

From all those rightful and careful question marks in Brian's replies, you seem to interpret this as a confirmation that panel traps can't hurt TL.

Ethan Winer wrote:

None the less, this low-bass panel trap uses 1/4 inch plywood for the front surface, so that's still pretty "far away" from the mass and self-resonance of even thin (1/2 inch) sheet rock.

Even just a gypsum board in front of a heavy brick wall will alter the TL properties significantly of that wall. In the tables I entered and very good analyzed, and presented by Bob you can see the influences on the MSM, showing you it creates different systems. Triple leaf systems can behave strange, but you can't conclude anything more from this than the valuable but very careful comments as expressed by Brian. Which certainly does not support the almost (free translation): "You see, I thought so too, common sense ....." In your common sense, you saw this as something added to an existing situation which you interpret about as some constant. The tables I entered make clear that adding something to an existing wall (whether single or double) alters the system, based on strict MSM calculations (ignoring other factors as mounting, damping etc.). http://www.johnlsayers.com/phpBB2/viewt ... ght=#48870 In how far this all translates to real live circumstances should be tested. And important is that you tune a trap ALWAYS to dominant disturbing audio frequencies. Bob replied rightfully in the other thread that his reply/presentation was relevant to your question. http://www.johnlsayers.com/phpBB2/viewt ... ght=#49275 It should indeed be very interesting to see representative real live measurements. That suggestion of Brian is very nice, gentle and welcome.

A very relevant but simple experiment - http://forum.studiotips.com/viewtopic.php?t=1198 FWIW... Steve

Knightfly: Cool memory.

Yup; 1 gigabyte, to be precise :mrgreen: (It's called [ctrl]b in Mozilla, lets you search for stuff you've stupidly filed in the wrong folder... :? )

Hi again everybody. Just a couple of notes, if i may: Ethan, you are sure welcome for my contributions, and i'm grateful that you thought to invite me. Eric, your comments and bolding of "necessarily" are very fitting and prudent. Certainly the only honest answer i could give is "i don't know for sure". Imagine a 4' x 8' sheet of drywall or plywood freely floating in the air. If you struck that sheet of plywood, it would ring a bit like a bell. you could hear the reaosnance slowly die down. Now go to a wall near you - a boring old conventional 2x4 wall - and do the same thing. Strike the wall and listen. You won't hear a ringing slowly die down. The reason for this is that lots of little things add damping to walls. The freely floating sheet has very little damping, but a wall has losses introduced by screws and friction and rubbing and minute air leaks and so forth. Probably 10x higher damping for the wall than the free sheet is conservative. Still not great damping, but alot better than the free sheet. Now imagine a panel trap on a wall. Its closer to the free floating panel than to the wall for damping, i'd guess. I know the insulation adds some damping, but i don't know how much. That isn't meant as anything other than an observation that a panel trap may have higher potential for severity-of-resonance than a typical wall. If the damping of a real wall was as low as raw drywall, TL at the low-freq resonance would be very nearly zero, regardless of mass.

In about 2 weeks i will be in a transmission loss test lab with some floor/ceiling installations. I am open to suggestions of any kind, but i just propose this: i'll test a wall (which would have to be tested anyway) -then screw a 4' x 8' frame with weather stripping on the back to the wall -add plywood and insulation to make a simulated panel trap -test again And report the unofficial (its more expensive to get an official test) results in this thread. Seem reasonable? My questions would be what thickness of plywood and what depth of trap and what type of insulation. And it will have to be 16", 32", or 48" wide.

Brian Ravnaas: Without insulation the peak will be higher, and probably more easily detectable. What are the modal resonances of the test room? We could target a frequency that will ring. What's the exact wall construction you have in mind? (e.g. {2G16, SWS with insulation, 2G16}). {No plans to 'build' it the way Paul described (have someone stand there and hold the thing)?}

Brian, > you are sure welcome for my contributions, and i'm grateful that you thought to invite me. < I figured all the other folks here would also appreciate hearing your perspective on this stuff. > Certainly the only honest answer i could give is "i don't know for sure". < And that's exactly why I asked you - 'cause I don't know either! :) If it turns out that building a wood panel membrane bass trap onto a wall really does harm isolation by a meaningful amount, I want to know that and I'll add a note to my Plans article. > Now imagine a panel trap on a wall. Its closer to the free floating panel than to the wall for damping, i'd guess. I know the insulation adds some damping, but i don't know how much. < It adds a fair amount of damping. If you rap it with your fist it does ring, but not for as long as if there were no insulation inside. In fact, rather than try to describe it with words, here's a small (83 KB) Wave file of exactly that: www.ethanwiner.com/misc-content/lb_trap_struck.wav You can load this file into Sound Forge or whatever and run an FFT on it too if you'd like. Or just listen. :)

i'll test a wall (which would have to be tested anyway) -then screw a 4' x 8' frame with weather stripping on the back to the wall -add plywood and insulation to make a simulated panel trap -test again And report the unofficial

That would be awesome! But best is if you could build a trap that's more typical of the size and shape people make them. I was told to make them two feet wide, which I assume avoids additional vibration modes and keeps the plywood vibrating more like a piston. Look at all the different ways a panel can vibrate here: www.kettering.edu/~drussell/Demos/Membr ... quare.html I assume the (1,1) mode as shown is the only vibration wanted for a wood panel trap. So if you use a 2 by 8 foot piece of 1/4 inch plywood, 4 inches deep, with 1 or 2 inch thick 703, that would be closest to how I've build these traps. Again, this is truly awesome - I owe you much more than just a beer! --Ethan

Brian, Also, here is the link again showing how to build these traps: www.ethanwiner.com/basstrap.html The plans are in a GIF file linked near the bottom of the article. Please let me know if you have any questions about the plans or materials etc. If you were near me I'd come in person and help! --Ethan

Brian, If you do a test, it should be based on a good base wall/seperation, where the MSM is not or as little as possible defined by structural resonances but by the MSM of the air cavity. I can imagine that in your standard comparisons versus GG you compare with walls/separations where the GG damping shows best. You know what I mean here. Not only the trap is defining to analyse the effect but the base wall as well. And make them standard as they are standard made. Those original RealTraps measurements which were published at the time must still exist. if that was 2' x 8' then measure them that way, else whatever size they had. (I think 2' x 7 1/2') Test report IBM: S030227 - February 27, 2003 Warm regards Eric

Eric_Desart wrote:

Brian, If you do a test, it should be based on a good base wall/seperation, where the MSM is not or as little as possible defined by structural resonances but by the MSM of the air cavity. I can imagine that in your standard comparisons versus GG you compare with walls/separations where the GG damping shows best. You know what I mean here. Not only the trap is defining to analyse the effect but the base wall as well. And make them standard as they are standard made. Those original RealTraps measurements which were published at the time must still exist. if that was 2' x 8' then measure them that way, else whatever size they had. (I think 2' x 7 1/2') Test report IBM: S030227 - February 27, 2003 Warm regards Eric

OK, so you vote 2' x 8', as does Ethan, that's what i'll go with. On a side note relating to Green Glue, i think that the site has same-lab, same-configuration tests for wood studs and steel studs comparisons and GG on resilient channel including short circuited RC walls is also online, and GG on conventional decoupled assemblies like stagg and double studs will be up by the years end. The steel stud comparison is not what i expected, but i think the data is online. I maintain that criticisms of that sort aimed at GG Co aren't very fair, as I don't know of another company who has put forth as much effort to show apples to apples. back to the topic: In thinking about it over the weekend, i was thinking that i'd build one of Ethans traps, and then just screw it to the wall, and probably over two different "walls" (which will actually be floor/ceiling combinations), one with conventional assembly where drywall is screwed to the joists, and one with resilient channel. so PANEL TRAP Subfloor joists with R19 fiberglass drywall and PANEL TRAP subfloor joists with R19 fiberglass resilient channel drywall And probably NOT test the trap on a Green Glue based assembly as the GG might, by virtue of being part of the same mechanical circuit, have a damping effect that could affet the outcome.

an FFT of the file you posted a link to, Ethan, seems to show the overlap of acoustic (air cavity based) and mechanical (panel) reosnances, quite a bit like a wall. the damping levels are apparently similar to a wall as well. WAs this trap mounted on a wall? You mentioned that you'd come if you were closer, if anybody else is near the minneapolis area, you're welcome to send a note and stop by as well for these tests, should be fun.

Thanks Brian, The tested paneltraps were 2' x 7 1/2' as was puiblished by Ethan. Brian, I wasn't critisizing anything. I know you measured more than anyone. The question was how does an added trap being a third leaf behave in function of the total system. Therefore it's good to see the effect on a separation were the standard air MSM applies and is dominant. Eric .

Brian: This is Ethan's thread, and ultimately he's interested in his own website's bass traps -- a worthy venture. But I always tell people to build membrane traps that are as big as possible 4'x8' -- so that the panel is a normal mass-only membrane, and to minimize the stiffness/structure component. So, if you find yourself with time on your hands, could you try a 4'x8' version as well?

z60611 wrote:

Brian: This is Ethan's thread, and ultimately he's interested in his own website's bass traps -- a worthy venture. But I always tell people to build membrane traps that are as big as possible 4'x8' -- so that the panel is a normal mass-only membrane, and to minimize the stiffness/structure component. So, if you find yourself with time on your hands, could you try a 4'x8' version as well?

If time allows, i can. As i'm envisioning, it should be a pretty fast process. As a sideline that i know Eric will follow but might sound like greek to those who haven't been involved in sound testing, a few questions: 1. would placing the trap in the "recieve" room so that its absorption contribution would change recieve room absorption and correction be preferable? 2. would simply taking a noise reduction and forgoing the process of testing absorption and correcting for it hide any information if the panel trap was on the "source" side and so not contributing to recieve room absorption? I do get where you are coming from, and its a good point. The mechanical resonance of a non-decoupled assembly is sometimes interfered with when adding a triple leaf, resulting in lesser effect. I totally get your drift.

Brian, > an FFT of the file you posted ... WAs this trap mounted on a wall? < Yes, that is a low-bass type (~100 Hz) and it's built directly onto a standard sheet rock wall. I have 12 traps (6 low-bass and 6 high-bass) built onto the walls, and I also have 5 of the traps I used to sell hanging on the wall firmly using French cleat hooks. If you'd like me to record a fist thump with any of those let me know. --Ethan

Hey Ethan, ever built a membrane trap(panel) across a corner, either wall/wall, or wall/cieling, like a "superchunk"? Just curious. Seems like that would be the place to put them. :? Of course, what do I know. fitZ