I am trying to push the envelope based on the idea/theory that the best panel for a DML is the lightest, least dense and most rigid that can be built. If a solid isotopic material ( think Foamular 150,250 Ext) could fit that bill you would be in this way of thinking, correct.
However, as many others have pointed out, if the above theory is correct, lighter and stiffer is better, composites that utilize “beam theory” in their construction are much lighter and stiffer than any solid material. These composites, ie carbon fiber skins over foam or aramid honeycomb fit that bill.
Read about concepts of beam theory and monocoque theory. It’s why jet airplanes are made of nomex honeycomb and aluminum.
As hobby builders we are trying to use engineering principles and modern materials to build the best we can. Within reason and budgetary limits.
To your question though, Foamular panels are pretty darn good!!
However, as many others have pointed out, if the above theory is correct, lighter and stiffer is better, composites that utilize “beam theory” in their construction are much lighter and stiffer than any solid material. These composites, ie carbon fiber skins over foam or aramid honeycomb fit that bill.
Read about concepts of beam theory and monocoque theory. It’s why jet airplanes are made of nomex honeycomb and aluminum.
As hobby builders we are trying to use engineering principles and modern materials to build the best we can. Within reason and budgetary limits.
To your question though, Foamular panels are pretty darn good!!
Lighter is certainly a obvious goal. Lighter means less inertia. Easier to control with mechanical -electrical drivers. More efficient
Stiffer or “ stiffer is always better” in DSL panels is up for debate. Rigid pistonic motion isn’t the goal for a DSL. A loose, floppy panel isn’t either. The best is somewhere in between.
Stiffer or “ stiffer is always better” in DSL panels is up for debate. Rigid pistonic motion isn’t the goal for a DSL. A loose, floppy panel isn’t either. The best is somewhere in between.
Thanks davelang. I had a look, and pretty quickly realised why I am not a physicist. There was also a link in one place I looked to plate theory ( Plate theory - Wikipedia ) which also might hold some interesting information, but again it's a bit above me.
I haven't tried foamular/xps yet. Would like to, but imagine it is not stiff enough in the 5-7mm variety, and have always wondered if a bit of high frequency gets absorbed in the thicker sheets? Plus, whenever I see a sheet for sale in a retail outlet here in Australia it always looks like its been runover by a truck.
I've been playing with some DML's made with 4mm thick twinwall polycarbonate glazing sheet. It weighs about 800g a square metre, and is extremely stiff in the verical, and very stiff in the horizontal. I have mentioned it here a few times, but have never seen anyone else talk about having tried it. Maybe Veleric mentioned looking for it once, not sure if he tried it though? I think there was some talk of a similar material a few pages back. It is obviously not a solid sheet, and its construction is possibly based on the beam/monocoque theories? It is however all constructed from the same material and extruded so there are no glue seams/lines involved
Not trying to cause any argument or put an opinion across, just thinking about it. Totally open to all panel/diaphragm construction suggestions. Sadly, don't have the time or funds to try all possibilities.
Last edited:
A commonly cited index for panel material is stiffness divided by density^3. So density is most important. On that basis alone, CF composite and wood are actually pretty close. But when you factor in that you are not using solid wood or solid CF composite, but rather are making a skin/core/skin composite, the fact that you can make CF composite much thinner than wood veneer favors CF composite over wood for a skin material.
Consider this: Assume any given core. With wood veneer (at about 0.5 kg/m3) you can perhaps get a 0.5 mm x 10 GPa skin. But with CF composite (about 1.6 kg/m3) you can get a .1 mm x 100 GPa skin. In that case the CF skin composite is both lighter (.16 VS .25) and stiffer (10 VS 5) than the wood version.
Still, I think wood veneer is a great choice for a skin, bit not as good as CF composite.
Eric
Thank you Eric for that really complete reply. My question is still but how stiff is best. I recognize that a very stiff light panel can be made but since we are not aiming for pistonic motion some flexibility is needed. Your industry index above (stiffness/density^3) is the first objective guideline I have come across . Where does it come from?
Thanks again!
Thanks again!
Last edited:
From the Parts Express Exciter guide:
"The ideal material for mounting an exciter is a thin, lightweight sheet of material with high compressive strength and moderate to high bending strength. The compressive strength of the material has the greatest effect on the treble extension of the resulting ‘speaker’ (affecting ‘detail’ and ‘air’), while the bending strength of the material influences the midrange and low frequency efficiency of the ‘speaker’.
The amount of acoustic output from the panel over its effective bandwidth is determined by the weight of the panel material and the available force from the exciter. For thicker or heavier materials, a larger and more powerful exciter may be needed, or multiple exciters may be used. Whenever possible, the weight of the surface to be excited should be minimized.
Excellent materials to use for exciter mounting include:
HD EPS and XPS are kind of unique in that they are relatively stiff, yuet bendable, extremely light, and yet can have a higher compressive strength in the HD versions which seems to make them quite suitable for a panel....yet we also seem to see a lot of HF muting going on with these materials with the typical cheaper exciters we've all been using over the last 5-10 years. I'm not saying the material is not good, just that we don't seem to find the right excitor/panel material combination just yet.
And even for the CF/nomex honeycomb composite the TI guys tested it and it didn't make there top 4-5 materials to use...probably due to availability, price and workability...but they didn't really say.
geo
"The ideal material for mounting an exciter is a thin, lightweight sheet of material with high compressive strength and moderate to high bending strength. The compressive strength of the material has the greatest effect on the treble extension of the resulting ‘speaker’ (affecting ‘detail’ and ‘air’), while the bending strength of the material influences the midrange and low frequency efficiency of the ‘speaker’.
The amount of acoustic output from the panel over its effective bandwidth is determined by the weight of the panel material and the available force from the exciter. For thicker or heavier materials, a larger and more powerful exciter may be needed, or multiple exciters may be used. Whenever possible, the weight of the surface to be excited should be minimized.
Excellent materials to use for exciter mounting include:
- Aluminum, Nomex/Kevlar, or resin-impregnated paper honeycomb sandwich composites
- Structural or syntactic foam sandwich composites (e.g. Rohacell)
- Fabric-reinforced phenolic plastic panels
- Fiberglass-reinforced resin panels
- Corrugated or honeycomb cardboard sheet
- Corrugated plastic ‘signboard’ material
- Foam-core poster board with paper backing
- Unreinforced plastic panels
- Mylar sheet
- Plexiglas or Lexan sheet
- Glass windows
- Mirrors
- Acoustic drop-ceiling tiles
- Wallboard
- Plywood/MDF/OSB sheets
- Metal panels
- Metal structural members
- Concrete
- Wooden beams
- Soil"
HD EPS and XPS are kind of unique in that they are relatively stiff, yuet bendable, extremely light, and yet can have a higher compressive strength in the HD versions which seems to make them quite suitable for a panel....yet we also seem to see a lot of HF muting going on with these materials with the typical cheaper exciters we've all been using over the last 5-10 years. I'm not saying the material is not good, just that we don't seem to find the right excitor/panel material combination just yet.
And even for the CF/nomex honeycomb composite the TI guys tested it and it didn't make there top 4-5 materials to use...probably due to availability, price and workability...but they didn't really say.
geo
The index I mentioned comes from the patents of Kenneth Heron and the UK Dept. of Defence in the 1990's.
One of the earlier ones is here:
https://patentimages.storage.googleapis.com/cc/2f/c0/ab18da804ded46/EP0541646B1.pdf
Here is a quote from the patent:
The invention claimed herein is a panel-form loudspeaker comprising:
a resonant multi-mode radiator element being a unitary sandwich panel formed of two skins of material with a spacing core of transverse cellular construction. wherein the panel is such as to have ratio of bending stiffness (B) to the cube power of panel mass per unit surface area (µ) in all orientations of at least 10;
Reading deeper you will find that this criteria is based on the idea of providing a panel with a low "coincidence frequency" and hence allow for efficient sound radiation over a broader bandwidth (by extending the low end deeper).
Whether or not this criteria (B/µ^3) applies to what we are doing is something you may have to decide for yourself. But my experience is that this index is a pretty good indicator of just how much sound power you can get from a panel with a given exciter input.
And how stiff is too stiff? In theory, that depends on how large you are willing to make your panels, and how low you want to be able to go. One source says that the low end is limited to about 2.5 x the fundamental frequency of the panel. But for a given size panel, the stiffer and lighter it is the higher it's fundamental becomes. Driving the fundamental back down to compensate requires a panel with more area.
Finding the right balance between stiffness, density and size seems to me to be our basic challenge. That's how I see it, anyway.
Eric
Hi davelang, they sound a bit like this, plus a better soundstage, more air around the music, and a less 'phasy' quality. These recordings were done late last year. (I removed some hiss from that tracks that came from my homemade mics, it also removed a bit of presence from the tracks)
I am going to get them out and re-measure them again properly, and play with some eq'ing and a sub-woofer this time. I have already found out through some quick measurements that I can halve the blu tack that I have over the voice coil centres and maintain the smoothing it gives between 5 - 10k with less drop in responce around 300 - 500hz.
Last edited:
I've mentioned Polypropylene (PP) a few pages back. At least what I've seen it's available also in 2mm resulting in 250g/m^2 - so approx. 3 times as light as Polycarbonate (PC) in 4mm what you've used. Although I'm not sure if 2mm does already flex/bend too much. As PC is used a lot for green houses and PP for transport protection you can get them quite cheap between 2-6€/m^2.
Test Honeycomb Panel
I decided to dive in and get feedback from the forum on the process I am using. If it works I will do a summary process posting
I had a little .25 mil Mylar from my ESL days and some Nomex honeycomb 1.8 lb
I couldn’t find my old ESL Mylar stretching frame so I compromised with tape.
My plan is to make a 1/2 size panel to avoid to much wasted expense if it fails. Two high density MDF plate are used for glue up. Mylar was stretched over the first plate with the goal of removing wrinkles. I didn’t want significant tension on the Mylar to avoid buckling of the panel. The second plate of MDF will be used to clamp.
Now to mix up glue. West’s System 2 two part slow cure epoxy. Will post more after first side is glued and dry. My plan is to use a ink brayer to spread out epoxy onto a piece of glass. Next will be to lay the nomex on the spread out epoxy to transfer “just enough” to the nomex. Finally, I’ll glue up the sandwich.
Feedback is welcome, please!
I decided to dive in and get feedback from the forum on the process I am using. If it works I will do a summary process posting
I had a little .25 mil Mylar from my ESL days and some Nomex honeycomb 1.8 lb
I couldn’t find my old ESL Mylar stretching frame so I compromised with tape.
My plan is to make a 1/2 size panel to avoid to much wasted expense if it fails. Two high density MDF plate are used for glue up. Mylar was stretched over the first plate with the goal of removing wrinkles. I didn’t want significant tension on the Mylar to avoid buckling of the panel. The second plate of MDF will be used to clamp.
Now to mix up glue. West’s System 2 two part slow cure epoxy. Will post more after first side is glued and dry. My plan is to use a ink brayer to spread out epoxy onto a piece of glass. Next will be to lay the nomex on the spread out epoxy to transfer “just enough” to the nomex. Finally, I’ll glue up the sandwich.
Feedback is welcome, please!
Attachments
Dave,
Be sure to let us know what you decide to get if you do pull the switch. I've been calculating for a while and just started making very small prototypes of carbon skin panels. Presently I'm leaning towards a construction with 3K carbon plain weave skins (which end up about .009 in thick) and a lightweight core of 1/16" or at most 1/8" thick. I fear that either thicker CF or a thicker core than this will make the panel too stiff, and hence need to be too large.
The core will likely be balsa or nomex, or other. It may depend on what I can get easily at the right thickness. Most (but not all) of the common core materials tend to be 1/4" thick or thicker. What works best with my limited construction capabilities is also a big factor.
Eric
Cool. I would not expect that something you can stretch with just tape will provide a stiff enough skin for a DML panel, but I guess you'll let us know!
Eric