From my side, I would have assume the cells of the honeycomb to be small compare to almost all the wavelength in order to play a role in the stiffness and density but not as resonant cavity. And even if there is a role, I would more think to a disturbance than a benefit.
The patent US20030081799A1 (I have it in mind because I read it recently) gives a cell distance 4mm.
Which data do you have about "very large" holes? of panel response with different cell sizes?
Good question!
About one year ago I had the same question to start... At that time I made the decision to choose an exciter with a good BL, low inductance and a low mass. Not necessary the highest power handling, not too expensive and available. I made a choice in the 25mm diameter because of what happens within the voice coil diameter. So I bought some Dayton Audio DAEX25FHE-4. Nothing to say against. As there are so many things to do with the membranes, I haven't searched for other possibilities.
Have a look to post 3433 and 4419 about the exciter Spedge uses
Loeb,
These are great questions! Really making me think. For now, consider this:
You don't need air chambers to create resonance. In fact, the resonating modes of DML panels are resonating modes of flexural vibration, not resonating modes of air cavities. Air cavities certainly can resonate. It's called Helmoltz resonance. But that's not the the primary (or intended) resonating mode of a DML panel. Helmoltz resonance requires an opening in the cavity, which nomex core panels would not typically have.
Probably you have heard of the Tacoma Narrows bridge, otherwise known as "Galloping Gertie". It collapsed due to wind driven resonance (no cavities required) within months after construction. DML panels likewise resonate quite well with or without cavities, just like Galloping Gertie.
Eric
+ @Leob
DML works mainly in the frequency domain above the 1st mode where the wavelength in the panel is shorter than the physical dimensions. All the points don't move in the same direction or at the same speed at the same time. The exciter "doesn't push" against the whole mass but just what it is in front of it. Then the movement is transmitted at a certain speed by the conversion of the energy between inertial and spring effect. It is a propagation mode. The modes occur because of the reflexion of the energy at the edges. As Veleric already said, the mode frequencies (the resonances) are linked to the conditions at the edges (free, simply supported...)
A one dimension analogy to that is the coaxial cable in electronics (or the telegraph or telephone lines). When the period of the signal is above the propagation time to reach the end, the total capacitance or the total inductance don't drive the propagation. the propagation is driven by the capacitance per meter and the inductance per meter which combined in the cable impedance. You can increase (of course in the limit of the attenuation by the cable) the length of the cable with no big problem.
In a previous post, I shared my tests with different materials, different dimensions. I didn't test a low density, high mass panel against a low mass, high density. It is perhaps not so easy to do because the low density, high mass panel might be very large or the high density, low mass to small... If you want to make a test Loeb, what about poplar plywood with a density of about 540kg/m³ or birch which a bit more dense and acrylic at 1200kg/m³?
After the tests I did, what I read about DML in academic papers (or try to read...), even if I have intellectual difficulties with it, I have no big doubt on the results.
When I did those tests, I was really surprised to get almost the same SPL with a big 10mm thick plywood than with a thinner one.
Christian
Christian.
When driving a 3mm ply panel full range with my 10watt exciters the unit is pushing its limits.
If I was to try this with a 10mm ply panel this would push the exciter coils into meltdown.
You need a lot more power to drive heavy panels full range in a real room .
Steve.
When driving a 3mm ply panel full range with my 10watt exciters the unit is pushing its limits.
If I was to try this with a 10mm ply panel this would push the exciter coils into meltdown.
You need a lot more power to drive heavy panels full range in a real room .
Steve.
I think that cavity or chamber is not really helpful ways describe what makes up a resonant body. For most people a cavity means something that they can see with their bare eyes. They don't think of the capillary structures of wood as cavities, nor the microscopic bubbles in XPS foam. For an object to resonate well you don't need any cavities in that sense. We known since ages how to make a resonant body to create passive acoustic amplification. You need something that is hollow in the sense that core density is low. Air is low density and works well, but will not hold very thin skins together. In fact in one of the Tectonic videos they basically describe the nomex core as a way of creating an empty space, but they need something to hold the structure together.
I was also surprised that thickness didn't negatively affect efficiency that much, but it makes more sense when you think about it as a resonant body, which has waves bouncing around in a 3d space, rather than a plane that flexes in 2d.
Reading the Wikipedia article on Tahoma Narrows Bridge, there is some interesting notes about the meaning of resonance, since it seems to be differing opinions if in fact resonance caused the failure. It goes in to that linear resonance and self exciting systems are different, and I think that difference correlates to what we are discussing here. At least from the way I understand those terms, initially I though about it more like linear resonance, when in fact the plate is a self exciting system, at least for large parts of the spectrum.
Perhaps 3mm is enough to allow the self exciting behaviour, and as you go thicker efficiency will not decrease as rapidly as expected, but in the end you just adding mass and moving the strong modes out from the useable spectrum. And if you put the exciter on just thin veneer you will start reducing efficiency again.
Now of course the density of your material, the dimensions of the plate and position of the exciter will determine where you have your modes, and having strong modes in the right part of the spectrum seems to be what makes an efficient DML. So it is hard to derive conclusions from only looking at thickness of the plate for example...even if you also specify the type of material there are a lot of variables missing.
I know you tried lots of materials, so not trying to argue against that experience, just sharing how I theorize around yours and others findings when trying to understand DML efficiency.
Leob.
on pages 165 and 166 I made some recordings of a 1mm veneer test panel , one was a room walk around.
If you have read these pages you would know that they work perfectly well and sound very good in a normal room.
These recordings were made while playing the music as loud as I could stand at 3m ,just to prove that they could handle continuous high db without distortions, using only a 10watt exciter.
The reason I started making small panels in the first place was because some people kept saying they won't work.
I'm glad I did experiment with small panels as I have learnt a lot more about dml and managed to experiment with vastly more materials, it's a lot cheaper !
I noticed on page 165, that I made a 14x24inch 1/2 inch test panel out of very low density eps, which to my surprise sounded very good, and with a little EQ as shown produced a very good frequency response at 11ft into the room.
not that I would ever drive that or any of my panels down to the 40hz point .
About 100hz is the limit I would recommend driving any dml panel in real room situations.
That 1mm veneer panel was at that time rolled off at 170hz ,so only 70hz difference from my max position.
If there was a choice of having a 7ft panel that rolled off at 100hz or a 9inch panel that rolled of at 170hz which would most people's other halves prefer ?
Steve.
Steve.
Hello Steve
Oh yes for sure. I have no intention to use 10mm plywood! It was just for test.
It would really be interesting to have the relation between the SPL, exciter displacement.
Your reply reminds on point : not less 10W exciter with a plywood panel. It is an element in the exciter choice.
Christian
Leob,
What do you mean by "linear resonance", "self exciting system".
It is known that mechanical parts, any metallic bracket around a gazoline engine, a bridge, a plate have modes. when an external cause (an exciter is one for DML membrane) applies a force, if the force is at the frequency of a mode, an important displacement occurs. if the force is too high, the limit of resistance are reached damaging the part.
No need of cavity in that. No need of low density
The natural flexibility of the part, its mass are enough.
You could make a DML with aluminium, glass, some of us use acrylic.
Christian
It was interesting to read the Wiki link about the bridge failure. I did not know (or at least recall) that the failure was driven by "aeroacoustic flutter", rather than a simple case of resonance.
I must admit I'm confused still about your meaning when you say "resonant body". But the resonating modes in a DML (also know as bending wave speaker), are the bending modes of a vibrating 2D plate. It's the motion of the surface of the plate creating waves in the air that creates sound waves. Not some kind of waves bouncing around through the thickness of the plate.
Putting nomex between skins is about making a stiff structure with minimal addition of weight. It's basically the same concept as a structural I-beam. The goal is to make a stiff structure with the minimum amount of material. If it were possible to make a plate that was paper thin, but still had the same flexural stiffness and (areal) density (and perhaps I should add damping) as the carbon/nomex sandwich, such a plate would make virtually the same DML as the carbon/nomex sandwich.
In other words, it's not the space itself between the skins that's important. Rather, the space between the skins is what creates a structure with a high bending stiffness, without adding significant weight.
Eric
Leob,
What do you mean by strong modes?
"efficiency will not decrease as rapidly as expected" : what is your expectation?
Why do you think the efficiency will be reduced with a thin membrane? having a kind of maximum?
Have you had a look how the modes are distributed in the theory of the thin vibrating plates? There is a first mode, almost nothing happens below and the other one linked to the 1st one. So I don't see how it is possible to reject modes out of the useable spectrum.
Yes the material and the dimensions but also the edge conditions determine the modes but not the exciter position. The exciter position determines if a mode can be activated or not. OK there might be a limit to that simplification because the exciter is not a pure force generator. It adds a mass and a local spring (the spider).
Oh boy, now it feels like we are ganging up on Leob.
Loeb, rest assured we are all trying to be helpfull!
Eric
Loeb, rest assured we are all trying to be helpfull!
Eric
It is my understanding too.
If the Tacoma bridge is not the good example (I haven't read the wiki page...), there are other examples with bridges broken because of mechanical force at the resonance frequencies. It is asked to soldiers to stop walking at pace on a bridge (a bit old perhaps...) and there are tests of bridge with heavy loads (trucks) at different speeds.
I think it might be possible to find examples in building construction... Let think to a washing machine as an exciter and the noise heard in some plate far away because of single concrete floor or the noise of the metallic walls of a store impacted by the wind.
No cavity in all of that. Only flexibility and mass.
4mm is massive compared to the void sizes in wood or PS foam. The Tectonic plates appear to be closer to 5 or 6mm
I think that the main problem I have with corrugated or honey comb panels is the distance between the two driving surfaces .
I have heard this problem on corrugated cardboard and on the podium speakers , and I also believe this is a problem also with thicker ply panels.
The sound waves travel across the rear exciter drive surface freely , but on the secondary front drive surface ,direct sound has to travel through the honey comb before reaching the front drive surface.
Basically the rear drive surface should be at the front, but then you have the problem of the exciter covering up the main driving area and all the problems that go along with that.
The only way I can think of that could one day solve the problem would be to have some sort of piezoelectric plate driver on the front surface that would be able to radiate directly into the air and also drive the main front driving surface?
Steve.
I have heard this problem on corrugated cardboard and on the podium speakers , and I also believe this is a problem also with thicker ply panels.
The sound waves travel across the rear exciter drive surface freely , but on the secondary front drive surface ,direct sound has to travel through the honey comb before reaching the front drive surface.
Basically the rear drive surface should be at the front, but then you have the problem of the exciter covering up the main driving area and all the problems that go along with that.
The only way I can think of that could one day solve the problem would be to have some sort of piezoelectric plate driver on the front surface that would be able to radiate directly into the air and also drive the main front driving surface?
Steve.
If anything I feel a bit silly for perhaps at times questioning conclusions that come with a lot of more experience as well as knowledge than what I possess. But I'm glad if my reflections are thought provoking enough to warrant you all spending your time replying, and just wish I had more time to think about your input and reply to you all 🙂
But I will make another attempt to clarify what I mean.
I'm not previously familiar with the terms "linear resonance" and " self exciting system" as used in the Wikipedia article.
But in the context of the bridge, I would interpret it like this; with linear resonance you have in this case a swinging object being pushed by a strong force, the wind. You don't need any oscillation of the wind to start swinging, it pushes the bridge until it cannot be pushed more. The bridge then falls back because it has lost momentum, until it again looses momentum after swinging back, and is pushed back by the wind again. The important thing is that the frequency and amplitude of the oscillations doesn't need to coincide with oscillation of the external force, and the movement will be proportional to the strength of the external force.
In a self exciting system you have an external oscillation that coincides with a the resonant frequency of an object. So instead of the wind just constantly pushing, causing the object to swing, it is like when you consciously is pushing a swing with the right timing to help the swinging motion. With very little force applied at the right time you can then generate a very strong motion. There will be a stronger correlation to the amount of resulting motion and at what frequency the force is applied, than how strong the force is.
So going back to sound, guitars and DML plates are then self exciting systems. But you have lots of different swings with different rope lengths that will swing strongly when pushed at a certain interval. Of course rather than swinging back and forth due to gravity, they travel through a medium and reflect back at the boundaries.
It seems that most agree that DML depends on resonant modes, and also that they are self exciting systems as described above. I think where it gets really hard to follow and agree is how the waves actually resonate in the plate.
Anyone with experience in audio is somewhat familiar with standing waves and that they relate to the size of a space. With a guitar it is pretty obvious that it has a hollow body acting as a resonator, and we understand that if you simply made a guitar of thin plate of wood it would be very quiet.
Of course a DML is different in that you excite the body directly, rather than using a body to amplify the soundwaves from a string. So I'm not saying a guitar would be the ideal DML, but I really think the plate needs mass to work well. Seeing it as 3d waves bouncing around in a body of a guitar or in a room, is probably incorrect though.
DML is more like a drum. You excite the skin, and the waves from hitting that skin can be pretty loud since you do have some resonance in the skin itself and can hit pretty hard, but without some body to amplify the vibrations from the skin it will still be a quite weak sounding drum. The panel ends up being the drumskin and body at the same time.
So far I compared similar, not identical, plates of different thickness, so can still be a misunderstanding that indeed more mass of the same material can be more efficient. I prepared prints of the same size plate as 1 mm, 3.8 mm and 6 mm thick which I will print with same thickness skin and same density infill.
My guess is that the 3.8 mm will be the loudest, but not sure why and if I'll be any wiser as to how it really works regardless of the result 🙂
Hello Leob,
No problem and don't worry. You can see with the answers that your posts are challenging and are the opportunity, at least to me to test my understanding of those strange things that are DML and are a possible source of new knowledge or even correction of my own misunderstanding.
I read quickly (meaning not all) about Tacoma bridge... interesting! so the reason seems a torsional mode triggered by aerodynamic behavior of the beam. Here I think we can speak about self oscillating system. The wind gives a constant input power, like a battery will do in electricity and the bridge oscillates on its own timing.
I haven't found wiki articles about self exciting system neither linear resonance. Do you have some links ?
In addition to the challenge of DML, I have the challenge to read and write in English... and some times there is a lack of connection with the french vocabulary.
So like that, I can't agree on the fact a DML is a "self exciting system". It is not for sure a self oscillating one because it reproduces the frequencies coming from the amplifier. It works in a forced mode. The frequency out is the frequency in.
The guitar is in my view not the best example as, as you says, there is an hollow body acting like a resonant cavity.
I prefer the comparison with the soundboard of the piano. The strings as exciter, the soundboard has panel (the plate).
I am not sure about the drum but I think it is another theory where like a string, the resulting frequency is linked to the tension in the membrane. In both case, the string or the drum membrane, an external impulse triggers the free mode (amortized free oscillations). The string or the drum membrane continue in oscillation on their on frequency.
In a DML, like in the soundboard of a piano or any soundboard (guitar, violin, harp) excluding any additional effect of an air volume, the frequency is external, coming from the string. The game is to have this frequency close enough of a soundboard mode so that the soundboard transmit the vibration to the air in an efficient way which a string can't do. In the case of a guitar or a violin, the hollow body comes in addition
Really, have a look around the piano soundboard or other soundboards (ie violin). The design constraint they have for instrument is a bit different than ours. One important is they have to resist to the string tension which is huge in a piano and they perhaps not have the same requirement of fidelity all over the audio frequency range
For your tests, if you could add the measurements of the 2 data that are the bending stiffness and the arial mass it would be really great in order to see how it fits (or not!) with the basics of the thin plate theory.
Christian
Like I said, the wikipedia article was my introduction to those terms. So I'm just doing a literal interpretation of the terms along with how it relates to the issue in the article.
Linear means you get output proportional to the input. So with the swing, you make one push and it will swing back and forth a few times, or with a metal pipe, you hit it once and it resonates for a short moment. How long it swings or loud it sounds depends on the strength of the input.
A self exciting system on the other hand describes the interaction between the external force and the resonating object. When the frequency of the external force coincides with a resonant mode, you get an output that is no longer directly proportional to the input.
If a guitar was not self exciting, you would only hear the sound of the actual string and, probably so quiet that it would be masked, traces of the wood getting excited by the soundwaves from the strings. To amplify a signal, the resonance cannot be linear.
I will see what I can come up with for measuring bending stiffness. Should at least be able to come up with some kind of rough relative figure between the different plates.