Pway,
The bending stiffness B and the aerial mass µ have a role in the efficiency (B/µ³ or E/rho³), the 1st mode (B/µ), the coincidence frequency (µ/B). So in my understanding, the first parameter is the density that leads the efficiency. Then thickness and dimensions are adjusted to reach the targeted 1st mode. If density and stiffness are the drivers, the poplar plywood is the winner among the most common plywood. Birch plywood has a good reputation for cheap or DIY music instrument. So there is perhaps an other parameter like the absorption but this hasn't been put forward here up to now. It would be interesting to see a FR of a marine plywood; what happens because of its treatment, how are the HF?
Christian
That looks pretty neat, and gives me an idea...
Assuming you could work out the magnets interference with each other...couldn't one glue two exciters back to back and then sandwich that between two panels with a flexible but enclosed frame to creat a unipolar DML? I picture two identical art canvases back to back.
You get no reverse polarity from the rear...the masses of the exciters should cancel and remain mostly neutral causing no distortions or energy loss from the counter-mass of the coil. The pressure changes in the sealed space between the panels could be a feature rather than a bug as long as it's included in the tuning since it would add to local bending forces.
Had to take some time from playing around with plates to sort out the subwoofer I got hold of. Replaced caps, but no luck, so had to get a new amp. But now finally I can run the plates as intended, and it sounds glorious!
I tried pushing them really hard with a HPF before, but of course that doesn't really give you an accurate impression of how it would sound like in a PA setup. But now with the subwoofer set up even my wife, who usually doesn't have strong opinions about sound quality, was dropping her jaw in awe.
So now I have gotten back in to see what I can achieve with the EPS plates. I got some hide glue and shellac and prepared a 500x330x10mm neopor plate. Seems like obvious choices to try out since that is what is used for instruments and soundboards since ages, and instrument making is one of the few crafts where they haven't gone over to PVA and other synthetic adhesives with many claiming it is inferior acoustically.
Both was a pleasure to work with really. It takes a few minutes to dissolve the glue, but on the other hand it dries completely in a couple of hours.
And the shellac dries extremely quickly...I just had it on and it seem ready to use already. Interestingly the whole plate became a bit convex when applying the shellac to one side, and the stiffness of the plate has increased drastically. Weight before applying the two coatings of hide glue on both sides and shellac on exciter side was 43g, and after it is 63g.
Will wait for tomorrow and then fix exciters using hide glue as well and post back with results.
I tried pushing them really hard with a HPF before, but of course that doesn't really give you an accurate impression of how it would sound like in a PA setup. But now with the subwoofer set up even my wife, who usually doesn't have strong opinions about sound quality, was dropping her jaw in awe.
So now I have gotten back in to see what I can achieve with the EPS plates. I got some hide glue and shellac and prepared a 500x330x10mm neopor plate. Seems like obvious choices to try out since that is what is used for instruments and soundboards since ages, and instrument making is one of the few crafts where they haven't gone over to PVA and other synthetic adhesives with many claiming it is inferior acoustically.
Both was a pleasure to work with really. It takes a few minutes to dissolve the glue, but on the other hand it dries completely in a couple of hours.
And the shellac dries extremely quickly...I just had it on and it seem ready to use already. Interestingly the whole plate became a bit convex when applying the shellac to one side, and the stiffness of the plate has increased drastically. Weight before applying the two coatings of hide glue on both sides and shellac on exciter side was 43g, and after it is 63g.
Will wait for tomorrow and then fix exciters using hide glue as well and post back with results.
If the material stiffness is high, you can use a thinner panel for a given B, thus reducing the areal density. In a realised panel, I don't know you can say that µ 'dominates' based upon B/µ³. Could you give again the reference for this efficiency relation? Small correction: 'aerial' in English means 'of the air', the correct word would be 'areal'.
Yes would be interesting to try marine ply. I am just now using some phenol formaldehyde coated ply (not true marine ply) to replace the old boards on some scaffolding. It's very hard, and is the same material as that original plastic, bakelite, which is hard and brittle. Whether the internal glue layers would get the same degree of hardness I don't know - it could be a different formulation. But its promising for the HF response of marine ply.
I just had a look at the different types of plywood. Aircraft plywood seems to fit the bill. More layers, best quality veneers, light and strong. A 1200x1200 sheet of 3mm costs around $AU 160
http://www.misterplywood.com.au/Products/Plywood/Aircraft-Plywood
http://www.misterplywood.com.au/Products/Plywood/Aircraft-Plywood
Tried fixing the exciter first with the hide glue, but somehow it didn't see to stick to the plastic on the exciter. Maybe if I would been more patient it will adhere, but when the glue was sticky enough to adhere to the shellac surface and my fingers really well.
I also tried shellac, and it seemed to be working ok at first, but after playing loud for a while one exciter fell off. Again, I might have been a bit impatient, and I let it harden overnight if I try it again.
The results with the 500x330x10mm neopor with hide glue and shellac was both disappointing and encouraging.
The disappointing part is that drop in HF is still there, and the FR doesn't look better, but is less flat:
It is weaker between 3-7k which makes it sound duller.
However, despite that it is the best sounding plate I have so far, especially when pushing the volume. The sound is more focused and it distorts less. Things sound punchier and interestingly when playing in stereo with the 25mm styropor, even if it has less HF and slightly lower sensitivity one gets the impression that sounds come more from the side the neopor. I would have though loudness and HF response would be more important, but I guess it has tighter transients despite the lower treble.
Also it seems to respond better to EQ, and when I try to adjust to get a flatter graph, the different between the neopor plate and styrpor becomes very big.
I think some tweaking can smoothen out the FR a bit, but my hopes of getting better HF response with the harder shellac surface was obviously false.
I also tried shellac, and it seemed to be working ok at first, but after playing loud for a while one exciter fell off. Again, I might have been a bit impatient, and I let it harden overnight if I try it again.
The results with the 500x330x10mm neopor with hide glue and shellac was both disappointing and encouraging.
The disappointing part is that drop in HF is still there, and the FR doesn't look better, but is less flat:
It is weaker between 3-7k which makes it sound duller.
However, despite that it is the best sounding plate I have so far, especially when pushing the volume. The sound is more focused and it distorts less. Things sound punchier and interestingly when playing in stereo with the 25mm styropor, even if it has less HF and slightly lower sensitivity one gets the impression that sounds come more from the side the neopor. I would have though loudness and HF response would be more important, but I guess it has tighter transients despite the lower treble.
Also it seems to respond better to EQ, and when I try to adjust to get a flatter graph, the different between the neopor plate and styrpor becomes very big.
I think some tweaking can smoothen out the FR a bit, but my hopes of getting better HF response with the harder shellac surface was obviously false.
Hello Leob
The reaction of people around us is really interesting. I had the same reaction from my wife with the DML. No question about the technique, the modal behavior or some accidents in the FR... just a reaction on the quality of the sound.
It will nice to see some pictures of the result of hide glue plus shellac.
Christian
Hello Leob,
I am sorry but I don't remember what is your situation about HF... I just found in previous posts you got some DAEX25, I remember you made tests in a 4 exciter configuration. With only one do you have some results reaching 10k at least. Even wit a small panel (HF doesn't change with the panel dimensions)?
Christian
Hello Paul
Thank you for the correction : yes "areal mass".
The question about the role of B/µ or B/µ³ was started a long time ago by Veleric. When I joined the discussion, we (Veleric and I) exchanged about that.
I have in mind 2 sources :
- one thesis about piano soundboard : Piano Sound board thesis
- one patent : WO1992003024A1
You can have a look to my previous posts I did to answer to a question from Cheapvega some months ago.
I reported about some quick tests (acoustic experiments in a living room!!!) that linked SPL to the B/µ³
The main outcome was the thickness is not a 1st order parameter of the efficiency (efficiency in the meaning of a SPL at some distance for a given input). The parameter is E/rho³ (E Young modulus, rho density) which is the same as B/µ³
Christian
Thanks Christian I will have a read.
I think the most useful formulas use the intensive material properties as far as possible (ie those that are dependent on the material itself, and independent of extent), and leave the dimensions (width, height, thickness) explicit in the formula. Thus E is to be preferred over B, volumetric density over areal density etc. That way, you're not bundling in the effect of eg, thickness which is critical and should be explicit.
Hi pway
As I said I was a plywood sceptic...and I still believe it is a compromise material compared to solid wood but it has the obvious advantages of thickness control, workability and availability....Just need to find the best type and then thickness for a particular panel.
Laserply is laserply largely because it uses nontoxic glue and they choose timber densities and thicknesses suitable for laser cutting. If panels are sealed (lightly) and used indoors, I doubt that the glue type will be critical.
A benefit is that the internal layer/s is/are claimed to be free of fillers/gaps (which can blow out under laser cutting). It does seem to be good quality. Another benefit is that it comes in 1200x600 sheets so you can try out a few types without investing in large sheets.
As you are in Aus you could order one or more sample packs for free - good for testing.
In the 3mm poplar I am currently using, each layer is the same thickness and they all seem to be poplar which makes sense as it's a low cost base material. It is truly light - 3000 kg/m3 - about 2x balsa density. Selectively adding more mass may benefit low frequencies but I haven't tried to date.
I'd only use the 3mm Poplar for smaller panels and if it's adequately supported - otherwise go for the 4mm Birch(/poplar - see below)
I checked the 3 ply birch 4mm sample against stated densities and the numbers tell me that the core is poplar - again, it has equal layer thicknesses.
Marine ply - Hoop pine would definitely be worth a try.
Laserply is available in Hoop Pine. They didn't include Hoop Pine in my sample pack so I couldn't test it. Then it was de-listed, now it's back.
More to find no doubt but really all too time and cash consuming.
Christian, as for Revply, it's not a product available in Australia - my comments were influenced by what I saw in the photo as excessive flexibility and the seemingly odd construction - (transverse 5 ply with extremely thin surface layers). I couldn't find any posts for panels made from it. If it works for you then good - no more need be said.
Cheers
Eucy
pway - I'd beware more layers --> stiffer, but more glue > more damping - 3 ply seems preferable to me. If it's too flexible transversely there are ways to fix it
Good reading! Let us know after.
I understand what you mean... an other way is also way to think is we may work not only with plain material but also composite. Do E or rho have a meaning then... I have some knowledge on mechanics and material but not enough to be fully at ease.
Christian
Very interesting Eucy. Thank you! I agree on your description about plywood use. The availability in the different countries with similar characteristics makes it in my opinion as the best "entry" material.
About the choice of the wood specy, the bending stiffnes (coming from the young modulus and the thickness) is also a parameter for the low frequency. I mean if both increase keeping the ration B/µ (I don't know if it is possible...), no benefit... Additional mass will then only reduces the efficiency.
About birch, I mentioned seeing it use in DIY music instrument. Is it because of a good balance availability/performance (as the poplar for DML) or other characteristic?
About the revply, I just tried to make a synthesis of what I saw about it. I am in France where in my knowledge it is not available or at least not in the DIY stores where I go.
Christian
Pway,
I also appreciate it when the solution can be reduced to the material's intrinsic properties. But Christian is spot on. Possibly the most efficient materials (other than PS foams perhaps) for dml construction are sandwich composites with thin high modulus skins and low density cores. For those materials, there is not really a single elastic modulus or density. You can talk about an "effective" modulus or density, but they are not longer intrinsic properties since they change if the ratio of the core to the thickness changes. So in that case, B and areal density are more generally applicable, as they apply to both both homogeneous and composite materials, while the E and rho formulas really only apply for homogeneous materials.
Eric
By adding mass I don't mean a heavier panel - I mean adding some form of localised mass in one or more spots. F0 is directly related to panel stiffness and inversely related to mass - so adding a controlled mass to a panel without increasing it's bending stiffness will lower F0. Usable low frequency response is said to be 2.5*Fo. The problem is that it may/will cause interference to other frequency ranges. There's no free lunch as we say
I mentioned birch because I tap tested a 4mm sample and it sounded OK - I haven't tested it in a full panel. Birch has been used by others with reportedly good results. It is considered to be a good(ish) tone wood (which is the most important starting point for me in considering materials). I suspect that all of the 'birch' ply used is actually birch outer skins with a poplar core like my sample - that's not an issue imo as the poplar is very efficient. Solid birch (SG 0.67) is 2x Poplar density (0.3-0.35) and 1.5x Cedar (0.38-.45) so having the poplar core brings it to about 0.55 which will help with low power exciters- It's still on the dense side however.
Regards
Eucy
pway,
That's a good point. I've been looking at the stuff I can find at the hardware store, but haven't recently considered the aircraft plywood suppliers. Was just looking now and basswood looks like a good candidate:
https://www.wicksaircraft.com/shop/basswood-and-basswood-plywood-aircraft-grade-mil-p-6070/
According to my reference book, basswood has a density of only 0.37, and a modulus (with the grain) of 10 GPa. That makes it about equivalent to Sitka spruce, which is he gold standard for guitar tops. Only problem is the shipping cost is more than the cost of the panels!
Eric
Eucy,
Revply is a US thing, so Christian has not tried it. But the 5 ply construction with thin outer layers is actually pretty common in the US for "underlayment" plywoods. These plywoods commonly use quite light core layers, which are surprisingly free of voids and such (for their cost). And curiously, they use an attractive reconstituted veneer face later. This construction leads to a pretty light and stiff panel, with similar stiffness in both directions and prettier face than most plywoods. So overall they make decent inexpensive panels.
I'm not sure what photo suggested "excessive" flexibility, but generally they are much stiffer in the "cross" direction than most any three layer plywood having roughly equal thickness layers.
Sadly, of late, the supply and quality of these plywoods has diminished considerably, in my area at least.
Eric
Eric,
This is a plywood answering to a MIL standard.
Have you considered the plywood coming in different thickness, species used in aircraft modelling? I found some interesting possibilities at french retailers (site only in french). Their plywood is said "aircraft" but with no certification for real aircraft use. I haven't search for more international retailers... This one in the US : https://shop.balsausa.com/category_s/1131.htm ? It is the kind of the kind of products I have seen in France.
I recorded also in my list of possibilities interesting store this one in Spain specialized in products for guitar making : https://www.madinter.com/en/ . No plywood but other products that might be interesting one day.
Christian
I read the first chapter of the piano soundboard paper (the only one in english). Very interesting how many concepts are common between musical instruments and our goals here. I split off the appendices and tried to translate, but the equations were mangled. Those appendices cover a lot of the important concepts for our purposes.
The factor that has been the subject of much discussion here (and forgive me, I havn't read all the correspondence on the matter!) appears early in the paper (p8) as 'radiation ratio' , 'acoustical constant' or 'Shelleng's vibration parameter
R=sqrt(E/rho^3)
I was about to refer to this paper https://sci-hub.se/https://doi.org/10.1121/1.4906186 which relates radiation efficiency for f<fc to damping factor eta. Then I read that the piano paper also draws a connection to damping. And indeed, with Eric's graph of acrylic vs plywood vs foam giving increasing SPL seems to suggest it could be damping playing a role.
I have to read these more carefully, as Im not sure I understand anyway. There are at least 3 sources of damping mentioned in papers, which is really just talking about energy losses - internal frictional damping, conductive damping (energy lost eg to the frame and surroundings) and radiation damping - the energy lost to radiated sound. If the relation is just between this factor and radiation damping, then it may be just another way of stating the same thing. And I cant see why the factor would have a relation to viscoelastic damping anyhow. OTOH the paper I cited above is talking about internal, viscoelastic/frictional damping related to greater efficiency at low f.
PS Inline html for greek characters doesn't seem to work. Have to use unicode entry I guess.
Last edited: