@EarthTonesElectronics
Dave,
Christian's (@homeswinghome) post reminded me of another feature that would be nice to have in the PETTaLS app would be the ability to add a support on the exciter. We often debate the effects of that on this forum so it would nice to have modelling results to refer to.
Also, in your reply on YouTube you mentioned your somewhat anomalous results for posterboard, you wrote:
These simulations tend to be tricky for non-isotropic materials like cardboard or posterboard. For example, I published some comparisons between measurement and simulation here (https://acoustics.org/how-to-find-the-best-material-for-making-exciter-based-plate-speakers/) showing how weird posterboard acts. Posterboard simulations are very accurate at low frequencies, and very inaccurate at high frequencies. I'll have to do more research into how to best model that - maybe a frequency-dependent Young's modulus or something along those lines.
The usual rationalization (on this forum) is that when the panel is too "soft" (like posterboard may be) we envision the exciter sinking into the panel at high frequency rather than driving it as it should. But I wonder if the real cause might be the low shear modulus of the foam core. As you probably know, at high frequencies the speed of bending waves is limited by the shear modulus of the core, as Stephen Hambric shows (below). So an isotropic panel and composite panel of the same stiffness and areal density might behave the same at low frequency, but not high frequency. I'm thinking that the lower wave speed in the posterboard composite at high frequencies could be reason it doesn't model accurately at high frequencies, if your model doesn't take that into account. What do you think?
Eric
https://www.hambricacoustics.com/IN16_5_Hambric_rev1.pdf
Dave,
Christian's (@homeswinghome) post reminded me of another feature that would be nice to have in the PETTaLS app would be the ability to add a support on the exciter. We often debate the effects of that on this forum so it would nice to have modelling results to refer to.
Also, in your reply on YouTube you mentioned your somewhat anomalous results for posterboard, you wrote:
These simulations tend to be tricky for non-isotropic materials like cardboard or posterboard. For example, I published some comparisons between measurement and simulation here (https://acoustics.org/how-to-find-the-best-material-for-making-exciter-based-plate-speakers/) showing how weird posterboard acts. Posterboard simulations are very accurate at low frequencies, and very inaccurate at high frequencies. I'll have to do more research into how to best model that - maybe a frequency-dependent Young's modulus or something along those lines.
The usual rationalization (on this forum) is that when the panel is too "soft" (like posterboard may be) we envision the exciter sinking into the panel at high frequency rather than driving it as it should. But I wonder if the real cause might be the low shear modulus of the foam core. As you probably know, at high frequencies the speed of bending waves is limited by the shear modulus of the core, as Stephen Hambric shows (below). So an isotropic panel and composite panel of the same stiffness and areal density might behave the same at low frequency, but not high frequency. I'm thinking that the lower wave speed in the posterboard composite at high frequencies could be reason it doesn't model accurately at high frequencies, if your model doesn't take that into account. What do you think?
Eric
https://www.hambricacoustics.com/IN16_5_Hambric_rev1.pdf
Hello group,
Finally I have made good progress with my dml panel, thanks to Earthtoneselectronics in his first video (showing effect of voice coil diameter), the Amina patents, Spedge's advice, and this https://patents.google.com/patent/US11012784B2 patent (with finally clear non-wooley explanation).
I printed a cone with ventilation holes for my xt32 exciter.
When comparing it with the standard setup there's a lot more high freq, and the mid peak that was troubling me disappeared. It really is a big difference. I can now use the panel without eq. Nothing for nothing though: overall spl is less, and the bass is also a bit quieter. Cannot understand why that would happen.....
Also not clear what caused what : maybe the ventilation holes made the mid peak (cavity?) resonance disappear? or was it drum-head kind of resonance dampened by the new contact pad?
Anyway I'm quite content and plan to make a version with a plastic bolt to go through the panel - because I worry the small contact surface will make the skin shear from my nomex.
If anybody with a 32 mm exciter wants to try this I could share the stl file, or send some printed parts.
Happy new year, cheers, Hans
Finally I have made good progress with my dml panel, thanks to Earthtoneselectronics in his first video (showing effect of voice coil diameter), the Amina patents, Spedge's advice, and this https://patents.google.com/patent/US11012784B2 patent (with finally clear non-wooley explanation).
I printed a cone with ventilation holes for my xt32 exciter.
When comparing it with the standard setup there's a lot more high freq, and the mid peak that was troubling me disappeared. It really is a big difference. I can now use the panel without eq. Nothing for nothing though: overall spl is less, and the bass is also a bit quieter. Cannot understand why that would happen.....
Also not clear what caused what : maybe the ventilation holes made the mid peak (cavity?) resonance disappear? or was it drum-head kind of resonance dampened by the new contact pad?
Anyway I'm quite content and plan to make a version with a plastic bolt to go through the panel - because I worry the small contact surface will make the skin shear from my nomex.
If anybody with a 32 mm exciter wants to try this I could share the stl file, or send some printed parts.
Happy new year, cheers, Hans
I've got a few things to respond to - sorry if these aren't in order. Also happy to start a different thread if this technical stuff is getting too off-topic for this thread.
I added the ability to model the acoustic output relative to electric power input - so sensitivity would be relative to 1W. I have no idea right now how accurate this will be, and I have my doubts without doing the exciter characterization myself (like I did with exciter resonant properties), because the BL factors for the Dayton exciters (which I think are mostly from BillionSound) are probably inflated. I'll also add in the ability to have a support on the exciter - that was modeled in my 2017 paper, so it's a quick addition. I'm trying to maintain a nice GUI without too many options/checkboxes/etc., so hopefully I can organize all of these options in a coherent way!
@Veleric your first question was about measuring shear modulus using the tap method - this is usually along the lines of what I do, but with accelerometers and an Audiomatica CLIO system that does sine sweeps to get the IR. As I'm an EE guy (not a materials specialist), I've mostly focused on modeling exciter-panel-box interactions rather than the panel materials. The Hambric paper you posted above looks like a great model for the sandwich panels... I've always assumed that these panels have a mostly homogeneous set of properties at long wavelengths, but, like you said, the inner core material is too squishy at high frequencies to translate motion from the back of the material to the front. I bet that the Hambric model can be easily incorporated into my PETTaLS model and it'll just be a matter of empirical measurement for the material constants. My thesis advisor Mark was very enthusiastic about a special type of panel (here) that definitely sounds really good, but I'm not sure if it could be modeled in the same way. Accelerometer IR measurements will definitely help to figure that out!
@homeswinghome you had asked about dipole behavior, and everything you said sounds right on. This is one of the reasons that I prefer surface vibration measurements over acoustic measurements, because the panel radiation properties are so complex. You can end up in an acoustic node for a specific mode and miss it altogether, and/or the dipole cancellation at low frequencies can be pretty impactful. This paper is the only one that comes to mind where this effect was measured. I like to think of each mode as its own independent "speaker," and it's mostly the lowest modes that are impacted by either dipole cancellation or have their resonances shifted by an enclosure. I'm not aware of a comprehensive, simple model for dipole effects right now, but I can add in the effects of a enclosure to my model (again, this was modeled in my thesis and my 2017 AES paper).
Also hoping to make the software available soon. Gotta clear it with the University (because they sponsor my research)!
I added the ability to model the acoustic output relative to electric power input - so sensitivity would be relative to 1W. I have no idea right now how accurate this will be, and I have my doubts without doing the exciter characterization myself (like I did with exciter resonant properties), because the BL factors for the Dayton exciters (which I think are mostly from BillionSound) are probably inflated. I'll also add in the ability to have a support on the exciter - that was modeled in my 2017 paper, so it's a quick addition. I'm trying to maintain a nice GUI without too many options/checkboxes/etc., so hopefully I can organize all of these options in a coherent way!
@Veleric your first question was about measuring shear modulus using the tap method - this is usually along the lines of what I do, but with accelerometers and an Audiomatica CLIO system that does sine sweeps to get the IR. As I'm an EE guy (not a materials specialist), I've mostly focused on modeling exciter-panel-box interactions rather than the panel materials. The Hambric paper you posted above looks like a great model for the sandwich panels... I've always assumed that these panels have a mostly homogeneous set of properties at long wavelengths, but, like you said, the inner core material is too squishy at high frequencies to translate motion from the back of the material to the front. I bet that the Hambric model can be easily incorporated into my PETTaLS model and it'll just be a matter of empirical measurement for the material constants. My thesis advisor Mark was very enthusiastic about a special type of panel (here) that definitely sounds really good, but I'm not sure if it could be modeled in the same way. Accelerometer IR measurements will definitely help to figure that out!
@homeswinghome you had asked about dipole behavior, and everything you said sounds right on. This is one of the reasons that I prefer surface vibration measurements over acoustic measurements, because the panel radiation properties are so complex. You can end up in an acoustic node for a specific mode and miss it altogether, and/or the dipole cancellation at low frequencies can be pretty impactful. This paper is the only one that comes to mind where this effect was measured. I like to think of each mode as its own independent "speaker," and it's mostly the lowest modes that are impacted by either dipole cancellation or have their resonances shifted by an enclosure. I'm not aware of a comprehensive, simple model for dipole effects right now, but I can add in the effects of a enclosure to my model (again, this was modeled in my thesis and my 2017 AES paper).
Also hoping to make the software available soon. Gotta clear it with the University (because they sponsor my research)!
Nice work!
Also consider how the coil is wound, as the sound is believed to depend on that.
The first and the third are the standard methods of winding. The last one is said to be Manger's design. It's all about how the sound is transferred from the coil to the diffuser surface in order to minimise distortions and provide full bandwidth.
Hello Hans
Excellent! This should be an important step to improve our panels!
2 questions (or problems) have been identified here for a long time : the exciter/panel interface (which shape to transfer the force from the exciter) and the cavity noise (the "noise" from the air trapped in the voice coil area which is associated to the peak in the FR mainly visible from the back side).
In the idea to design some adapter (I can't print in 3D...) I made some tests to evaluate which weight it is possible to add on the voice coil without reducing the highs. For a DEAX25FHE, 0.5g should be no problem, 1g probably possible. This might be panel material dependent.
Hans, could you post a cross section of your part showing the height, the diameter at the panel and the position of the holes regarding the voice coil? Maybe the weight also?
Christian
Hi Christian,
I just guessed the thickness and surface area, and made a piece that looked logical to me (stiff enough and with enough surface area, and big ventilation holes.
I printed the piece without trying to get it as light as possible with standard wall thicknesses and a bit more infill than standard (25%,3d honeycomb).
Nothing scientific, just wildly trying without knowing what I'm really doing🙂.
The part is about 2,5 g (that's what the software predicted), so a lot heavier than your calculated optimal. It feels light, and once it's connected to the panel would the weight have such a big influence?
anyway, I'll attach two pictures-sorry pdf.
Cheers, Hans
In relation to 3D Printing, approx' 1 year ago I met a individual on a train journey who I was to become engaged with in a conversation.
In the talk I was informed the person was retired from a career in IT. The last few years of their employment was designing Software for 3D Printing that would produce the lightest weight parts with structural properties that were acceptable to be used for Military Purposes.
I was intrigued and did ask further questions, but due to the Client the individual was not allowed to be too forthcoming with their acquired knowledge.
There is software in use that will show for certain print substrates how the lightest and structurally acceptable components can be produced.
In the talk I was informed the person was retired from a career in IT. The last few years of their employment was designing Software for 3D Printing that would produce the lightest weight parts with structural properties that were acceptable to be used for Military Purposes.
I was intrigued and did ask further questions, but due to the Client the individual was not allowed to be too forthcoming with their acquired knowledge.
There is software in use that will show for certain print substrates how the lightest and structurally acceptable components can be produced.
No calculation on my side, nothing highly scientific. It was just a set of tests of the on axis response while adding some weight on the central area (large enough to cover almost all the central area). I haven't posted them because I think further investigations are needed : from those measurements, I had as 1st conclusion that increasing the weight doesn't mainly cut the highs but creates some dip before. I found that strange. The weights were added on the front face so maybe not fully representative. My evaluation was for a DAEX25 with 3mm plywood.
Nevertheless, as your interface seems to show it also, there is a certain possible weight allowance to create a part.
In addition, there are probably more benefit to eliminate the cavity noise than to keep the highs (if both targets can't be satisfied at the same time)
I am also thinking to have a nut at least for testing (diameter? nylstop?) or to add an axial pin glued to the interface going through a hole into the panel and then glued to it. In case of a honeycomb composite, maybe a small additional "washer" with a hole of the pin diameter... or maybe a pushpin inserted from the front side.
Here is a sketch of your interface with the exciter to give an idea (I made the assumption of 8mm holes).
Last edited:
No idea which year of production, but it’s from Japan and old. It features wood panels on a wooden baffle and leather surrounds...
Another one, also from Japan, features a flat round panel and an interesting back with a wooden ribbed cone (?).
And a 12-year-old Japanese video featuring a flat polystyrene panel. An interesting surround.
An, those holes in the middle cone...
Another one, also from Japan, features a flat round panel and an interesting back with a wooden ribbed cone (?).
And a 12-year-old Japanese video featuring a flat polystyrene panel. An interesting surround.
An, those holes in the middle cone...
Last edited:
A pushpin with a drop of 3 second glue might be a very good (and practical!) idea. One downside : less easy to remove without damage.
Thanks.
Last edited:
@EarthTonesElectronics
Dave,
Another feature that would be great to have in PETTaLS would be the ability to simulate the electrical impedance of the flat panel speaker. In your reply to Christian, you mentioned how surface acceleration is in some ways better (less complex) than acoustic measurements. Impedance seems to be like that too, while still being is sensitive to every aspect of the flat panel speaker's design. But impedance is far easier to measure without expensive equipment, so more accessible to DIYers.
I think the main value of having an impedance simulation would be to provide a way for a user of PETTaLS to "reality check" their simulation. That is, if the measured impedance doesn't resemble the simulation, then something is off. Any mistake in setting up the simulation, like using the wrong panel properties, or dimensions, or exciter position, or boundary conditions, or damping level, and so on, would all cause the simulation to differ from the measured impedance, and the user would realize that there is problem somewhere in the simulation.
Participants in this forum know how much I love my impedance tests. It's great for a) identifying natural frequencies , b) comparing internal damping of panels, c) assessing damping of perimeter materials, and more. It would be great to have yet another use for it.
Eric
Dave,
Another feature that would be great to have in PETTaLS would be the ability to simulate the electrical impedance of the flat panel speaker. In your reply to Christian, you mentioned how surface acceleration is in some ways better (less complex) than acoustic measurements. Impedance seems to be like that too, while still being is sensitive to every aspect of the flat panel speaker's design. But impedance is far easier to measure without expensive equipment, so more accessible to DIYers.
I think the main value of having an impedance simulation would be to provide a way for a user of PETTaLS to "reality check" their simulation. That is, if the measured impedance doesn't resemble the simulation, then something is off. Any mistake in setting up the simulation, like using the wrong panel properties, or dimensions, or exciter position, or boundary conditions, or damping level, and so on, would all cause the simulation to differ from the measured impedance, and the user would realize that there is problem somewhere in the simulation.
Participants in this forum know how much I love my impedance tests. It's great for a) identifying natural frequencies , b) comparing internal damping of panels, c) assessing damping of perimeter materials, and more. It would be great to have yet another use for it.
Eric
Excellent work on this, Hans! Regarding your concept, could similar benefits be had from a concentrated force, even if the contact area is not round? The pictures below show a concept for using 2 exciters to excite the same modes as using 4 exciters. Would there need to be 4 contact points or could a single contact surface work? This would have more attachment area. I really appreciate the recent progress from this group!
Thanks, Bruce
Ähhh, sigh.... How quickly we forget.😢😜.. If you search for yellow trace, maybe 300 pages ago, you'll see I have been using stressed skins from the get-go.
Eucy
The drum skin is always under tension. They have been used for centuries. 🙂
That's a real problem with this lengthy tread - one tends to forget what was shared 2 or 3 years ago. But I still remember the Eucydome and might give it a go later this year.
Ogitone speakers. The person who created it is Keikichi Ogi.
However, he won't disclose which paint he used.
(There is someone in Japan who experimented with these —he made them himself—adding a Dupont Corian round baffle around it and fixing it onto an acrylic board, with the back left open. I cannot find his website at the moment.)
Last edited: