John Sayers' Recording Studio Design Forum Archive

Hello,

Sorry, my search turned up zero results.

It's my understanding that a diaphragmatic(membrane) absorber is built like a speaker, so the front can vibrate, but the back and sides are stiff, massive, cannot move.

Let's say one were to build a diaphragmatic absorber with the following attributes:

-rectangular, wood, let's just pick a size to picture it - 24"W X 48"H X 12"D.

-Front AND rear are both resonant, made of the same material and thickness. So for practical purposes, let's say it's 1/8 or 1/4" plywood, 24 X 48.

-Sides are very stiff, made of MDF.

-Let's say it's placed perpendicular to the wall, so that, in theory, bass energy from the room reaches both membranes at exactly the same time.

Would the two opposing panels attempt to push on each other, and therefore would each behave as if it were in an enclosure in which the back wall were infinitely stiff? And would that serve to effectively absorb the sound energy?

Or is that not how it works?

Or is that not how it works?

Correct! That's not how it works....

A diaphragmatic trap (a.k.a. membrane trap, panel trap, etc.) works by having some type of membrane held in place over a sealed cavity of fixed volume and depth. It is tuned to it's operating frequency by choosing the surface density of the membrane, and the depth of the cavity. The equation (simplified) is:

f = 60 / sqrt(md)

Where "m" is the surface density, and "d" is the depth of the cavity.

So, if the cavity does not have a fixed depth (because you made the other side into a membrane as well), then it won't have a stable tuned frequency, and since the cavity is basically not sealed (the volume changes), it won't be very good at resonating.

OK, so you could probably figure out how much the depth will change, and allow for that in your calculations, so you would end up with a range of possible tuned frequencies... but that indicates your trap has a low Q! It is not tightly tuned, and therefore not efficient at trapping any of the frequencies it might be tuned to.

Then again, this would only work with a panel type membrane trap (pistonic), not with a limp membrane type. With a flat panel, it acts like a piston moving in and out, so the entire depth changes by the same amount. If you have limp membrane, then it moves in and out a lot more in the middle than it does along the edges, so the depth changes a lot more towards the middle, with very little change at the edges. Once again, that makes your Q lower and broader, implying a broader range of frequencies and lower efficiency.

Finally, there's the issue of reality and construction. I don't know if you have ever tried to actually build a tuned trap of any type, but if so then you already know that what the text books tell you, and what happens in the real world, do not match very well! The REAL tuned frequency of any tuned trap (Membrane, Helmholtz, etc.) is almost never the same as what the equation said it would be, because real-world building materials are not perfectly uniform, real-world tools are not perfect and don't cut perfectly straight, real-world people are not perfect, and don't build perfectly, and worst of all, the air in your room is not uniform temperature, pressure, and humidity, and does not match the perfect assumed conditions in the equation. So your tuning will pretty much always be off. So you need to have some method for changing the tuning the finished trap to get the actual frequency that you wanted. How would you build a device such as you propose, while still being able to tune it after it is finished?

-Let's say it's placed perpendicular to the wall, so that, in theory, bass energy from the room reaches both membranes at exactly the same time.

That's a big assumption! How would you ensure that this really is the case, and that the energy arrived at both sides, in phase? In the real world, sound waves don't arrive perfectly normal to all surfaces, all the time. They bounce around the room, and arrive at all types of angles. It is quite feasible that the energy would NOT arrive at both sides of your device in phase, because the wave is passing by from left to right, or right to left, or perhaps two versions of the wave are arriving at that point from different directions... what happens then?

But let's assume that you got lucky, and both sides of your trap did, in fact, see the exact same phase of the sound. So both sides "press in" at the same time, then get "sucked out" at the same time.... thus creating the largest possible change in cavity depth! Thus lowering the resonant frequency when they are "sucked apart" and raising it when they are "pushed together".... In other words, low Q, low efficiency, broad tuning, blah blah blah...

And all of this begs the question: Why? Why would you want to give up a very large area and volume of your room for a device that only deals with one very specific frequency, or a small range of frequencies? If you have a dozen modal issues in your room that all need treating, would you really have several on each wall, sticking out a meter or more into the room, and also have a few hanging down from the ceiling, and some more sticking up from the floor? What if you can't get the device to the right location?

In general, it is far, far better to treat rooms mostly with absorption, which is effective down to low frequencies if done right, and is broadband. It treats all frequencies at once, at every point over its surface, and it can be "tuned" as well, by placing reflective or diffusive surfaces in front of it. I'm not a big fan of trying to treat a room with highly tuned devices, for all of the above reasons.


- Stuart -

Fascinating! Thanks Sensei for your prompt reply.

Allow me to respectfully challenge.

Soundman2020 wrote:So, if the cavity does not have a fixed depth (because you made the other side into a membrane as well), then it won't have a stable tuned frequency, and since the cavity is basically not sealed (the volume changes), it won't be very good at resonating.

Yes it's possible that the design could be less effiicient than one with rigid back, because when one is pushing in, the opposite side is pushing against it. But as long as it's more than 50% as efficient as a rigid back, there should be a net benefit from having twice the area resonating.

Soundman2020 wrote:... but that indicates your trap has a low Q! It is not tightly tuned, and therefore not efficient at trapping any of the frequencies it might be tuned to.

My understanding is that a membrane absorber is already low Q by design. As I understand it, the depth doesn't determine the centre of the frequency, but instead the LOW CUTOFF, and that it goes up from there. So a membrane absorber might do from 60Hz up to a few hundred Hz - higher if there is 703 in front of it but not touching.

If I have that part correct, then I care not if, as it's vibrating, the plywood is moving my cutoff from 60Hz to 59 or 61. Actually, I would suspect it would be fractions of a cycle, but either way, who cares...if it's working?

Soundman2020 wrote:Finally, there's the issue of reality and construction. ... So you need to have some method for changing the tuning the finished trap to get the actual frequency that you wanted. How would you build a device such as you propose, while still being able to tune it after it is finished?

Irrelevant to this specific discussion because these are realities whether I'm building a hole panel Helmholtz, a slot resonator, a rigid backed diaphragmatic, or this double walled diaphragmatic absorber. None of them can be practically adjusted after they are cut, built, and glued - with the exception of the Helmholtz, in which one can drill more holes.

Soundman2020 wrote:

-Let's say it's placed perpendicular to the wall, so that, in theory, bass energy from the room reaches both membranes at exactly the same time.

That's a big assumption! How would you ensure that this really is the case, and that the energy arrived at both sides, in phase?

The whole reason we are attacking this problem is because we have waves that are 15-20' long. Really, even if we turned one side away, 12" difference on a 20' wave is going to be close enough to the same part of the cycle that each will be compressing or rarefacting at the same time.

I would think phase issues would only come into play when we are looking at short wave, mid and high frequencies. This would suggest that some 703 or similar in front would be of benefit - not touching the membrane of course.

Soundman2020 wrote:And all of this begs the question: Why? Why would you want to give up a very large area and volume of your room for a device that only deals with one very specific frequency, or a small range of frequencies? If you have a dozen modal issues in your room that all need treating, would you really have several on each wall, sticking out a meter or more into the room, and also have a few hanging down from the ceiling, and some more sticking up from the floor? What if you can't get the device to the right location?

My answer is this is a theory I'm curious about. I'm trying to determine if, theoretically, there is a net benefit, or a net detraction to the concept of a double walled absorber to tackle low end problems. I have room in my studio for 2' panels protruding from corners and ceilings, though many wouldn't have that extra room.

I suppose the way to do this would be to build 1/2 dozen panels, do some REW tests, then screw a rigid back on one side and test again to see which works better.

Cheers
Keith

Yes it's possible that the design could be less effiicient than one with rigid back, because when one is pushing in, the opposite side is pushing against it. But as long as it's more than 50% as efficient as a rigid back, there should be a net benefit from having twice the area resonating.

The BBC does not agree with you! They have an extensive library of acoustic research, both theoretical and practical, and many of their papers highlight issues that they ran into when building tuned devices and the box wasn't rigid enough. The devices did not perform as designed, and had to be modified with thicker, stronger, more rigid boxes. And we are not talking about replacing a rubber membrane with plywood! Rather, things like having to replace 12mm plywood with 16mm plywood, because the 12mm plywood back was flexing too much, and de-tuning the box. I'm not sure if you have built many small boxes with 12mm plywood, but it's not something that one would think of as "flexing too much".... yet it did. Enough to mess up the tuning. When they replaced that with the thicker stuff, the problems went away.

Check out their library for all the details.

But as long as it's more than 50% as efficient as a rigid back,

If the rigid-back version has a coefficient of absorption of, for example, 0.86 at the resonant frequency, and yours has only 0.43 then it won't be a lot of use.

Which of these two do you think would make a good membrane trap, designed to resolve a modal issue at 80 Hz?
I know which one I'd prefer!

Notice the Q in each case....

there should be a net benefit from having twice the area resonating.

.... nope! You'd have twice the area, yes, but that area is now NOT all sharply tuned to the same frequency. Only a part of the area is actually tuned to the frequency you want, since the Q is low.

And even if you do get 0.4 instead of 0.8 with twice the area, all that you have accomplished is getting the exact same effect with a device that takes up a lot more space inside your room.

My understanding is that a membrane absorber is already low Q by design

Really? So the green curve in the above diagram is low Q? Ummmm..... I guess I must not be understanding "Q" very well....

As I understand it, the depth doesn't determine the centre of the frequency, but instead the LOW CUTOFF, and that it goes up from there.

Then your understanding is wrong. Not mine! The equation defines the resonant frequency, which is usually a bell-shaped curve with the center of the bell at the resonant frequency.

There is no such thing as a "low cutoff" with a resonant trap. That term is normally applied to numeric-based diffusers (Schroeder, skylines, QRD, PRD, BAD, etc), where there really is a fairly sharp fall off at a certain point, but even then the effect does not "go up from there". It only goes up to the high cut-off, where the same issue happens again. Diffusion is also usually a bell-shaped curve, albeit broader, tilted, and not so well defined. The ONLY device that has a low frequency cut-off of sorts, and then "goes up from there" is a purely porous absorber. Tuned devices have bell-shaped curves, with one peak absorption frequency, and the absorption falling off on each side of that (higher and lower). some devices might have more than one peak, but each peak behaves roughly the same.

So a membrane absorber might do from 60Hz up to a few hundred Hz -

I'm not aware of any such device. Maybe you could show the design for one, and the equations that describe it's operation? It certainly wold not be a membrane trap. Normal membrane traps are tightly tuned, high Q.

higher if there is 703 in front of it but not touching.

Huh? What "703 in front of it"? I have no idea what you are trying to describe, but it is not like any membrane trap that is commonly used! Membrane traps have insulation inside, not in front of them, and 703 might not be such a god choice, depending on what the trap is tuned to.

If I have that part correct, then I care not if, as it's vibrating, the plywood is moving my cutoff from 60Hz to 59 or 61.

There is no such cut-off, so your point is moot.

Actually, I would suspect it would be fractions of a cycle, but either way, who cares...if it's working?

Do the math, and see. Or check the BBC papers, since they already did the math, then built them, the re-built them when they did not work. IT will save you the trouble if you read up on what they did, and the results they got, and how they fixed that.

Irrelevant to this specific discussion because these are realities whether I'm building a hole panel Helmholtz, a slot resonator, a rigid backed diaphragmatic, or this double walled diaphragmatic absorber.

Sorry, but it is ENTIRELY relevant! If you don't see why, then I'm not sure that I can help you much.

None of them can be practically adjusted after they are cut, built, and glued

Yes they can.... Sorry to disagree with you so much, but you are just plan wrong. All of them can be adjusted.

with the exception of the Helmholtz, in which one can drill more holes.

Nope! That would not necessarily work. It might, but it might not. For example, what if drilling more holes would increase the percentage open area such that it no longer acts as series of individually tuned devices, and instead starts acting as a low-Q broad-band device? And you are talking about the specific case of a perforated panel device, not the general case of Helmholtz resonators. Many Helmholtz resonators can be tuned rather easily, by changing the neck length, for example.

The whole reason we are attacking this problem is because we have waves that are 15-20' long. Really, even if we turned one side away, 12" difference on a 20' wave is going to be close enough to the same part of the cycle that each will be compressing or rarefacting at the same time.

What makes you think you could build a membrane trap tuned to 56 Hz that would be only 12" deep? How would you go about that? What materials would you use, and what would their physical properties be? How big would it be, and how far out into the room would it protrude?

This would suggest that some 703 or similar in front would be of benefit - not touching the membrane of course.

Once again, membrane traps do not have 703 in front of them. And even if they did, it would not change the way the trap itself operates. It would merely "filter" the sound in some fashion, so the trap would not "see" the full sound field in the room.

I have room in my studio for 2' panels protruding from corners and ceilings

That's great! You have a large room! Which means that modal issues are likely not such a big problem anyway, and can easily be dealt with. How big is the room, and what is the Schroeder frequency for it at present?

Also, if you have 2 feet to spare, I would go for hanger traps, for sure. Tom Hidley style (or John Sayers style...). Proven to work down to very low frequencies, with good efficiency. With 2 feet to spare, you have enough space to do that, for sure. I have done several rooms with hangers in a bit less than that (around 20"), with good success.

I'm trying to determine if, theoretically, there is a net benefit, or a net detraction to the concept of a double walled absorber to tackle low end problems.

It's an interesting experiment, certainly, and I'd love to see the results, but I'm very unconvinced about the practicality, even if you can figure out how to keep the Q high.

I suppose the way to do this would be to build 1/2 dozen panels, do some REW tests, then screw a rigid back on one side and test again to see which works better.

If you have the time, and the money, and the inclination to do that, that would be fantastic! I'd really love to see the results.


- Stuart -

Again. Fascinating.

Where can one get hands on the BBC paper? Sounds like they performed some interesting and telling experiments.

Do you have a link to the Tom Hidley style bass traps?l of which you spoke?

Maybe I was deceived, but I was under the impression the model in this link was a broadband diaphragmatic. I believe he said in one of their videos it goes from 30 to about 300(not sure of the exact number, and that they put a 703 style insulation in front of it to absorb even higher.

http://www.acousticfields.com/product/a ... -absorber/

This is crazy! I wrote a very long, detailed reply to this yesterday, and now it is gone. I explained in great detail why that device is not what it seems to be, and why I would never use such a thing in my studio, but all of that disappeared.

Not sure what happened, but I'm not going to write it again. Sorry.

I managed to recover part of that message from the buffer:

--------------------

Where can one get hands on the BBC paper? Sounds like they performed some interesting and telling experiments.

They have a large archive of research papers going back about 60 years. Thousands of papers on all aspects of radio, TV, audio, and acoustics:

http://www.bbc.co.uk/rd/search?Type=Publications

They used to have a better system for finding stuff, but they "improved" it a while back, so it is now not very useful at all...

Do you have a link to the Tom Hidley style bass traps?l of which you spoke?

You could try Google.

Maybe I was deceived, but I was under the impression the model in this link was a broadband diaphragmatic.

That's not a membrane trap. It's a rather more sophisticated device, with multiple elements, and very little information on how it is supposed to work, or how it is built (except for a very simplified diagram). The report from Riverbank is not to exciting either:
Not exactly what you'd call a "Broadband absorber"! It certainly does not absorb everything from 50 Hz to 300 Hz. The above information comes from the manufacturer's own website, and seemingly contradicts the marketing claims on the same website.

Yes, the data tables do show another peak at 50 Hz, bur Riverbank wisely (and correctly) does not include that in the graph, since it is notoriously difficult to get accurate, repeatable, usable measurements at very low frequencies. The device might or might not have another absorption peak at 50 Hz, but even if it does, that still does not fit the claims. (Comment from one article on this subject: "Absorption measurement standards such as ASTM C423 do not address taking measurement at very low frequencies. As a result taking absorption measurements below 100 Hz is an "out of spec" proposition. ")

It's also interesting that you'd need a large truck, a crane, and a lot of muscle-bound weight-lifters to install that device in your studio. It weighs nearly 1200 pounds! And it costs US$ 1450! At more than a dollar per pound, that's rather expensive acoustic treatment, even if it does work.

So what you have there is not a true membrane trap, weighs a ton, costs a small fortune, does not offer exciting laboratory measurements, and is not broadband. All of that is evident from the website you linked too. And if you happen to have modal problems in your room at 35 Hz, 90 Hz, and 120 Hz, then that device would do nothing at all for you. It would only be any use if you happened to have a probe at 50 Hz, and another, smaller, problem at 259 Hz

On the other hand, here's a graph of coefficients of absorption for 2" OC-703:
That's a true broadband absorber. And only 2" thick, instead of 16".

For 1400 bucks you can buy one hell of a lot of OC-703, and it won't weigh a ton!

I'm always wary of web sites that make big claims for very expensive products, especially when a lot of the information provided is questionable anyway... That's the same place that claims it is impossible to have a good control room with a volume of less than 3,000 cubic feet. On the other hand: http://johnlsayers.com/Pages/Spark_1.htm http://johnlsayers.com/Studio/Mainpage/MP-Mark.htm

But to get back on topic: The device you are talking about is not a membrane trap, is not broadband, and is not very effective.

- Stuart -