Is that all? I have no doubt that you can do it.
Eric
Lekha:
There are indeed not too many entirely new ideas in the universe of things.
Experimenters in this forum know from tests that exciters with smaller diameter coils and hence contact area/diameter are superior in high frequency performance than larger ones. The difficulty lies in determining whether that is solely due to the contact area or the exciter.
I believe that my oft mentioned "Eucydome" functions in a similar manner by encapsulating the area of the coil between the panel and the dome, creating an air spring while also allowing the aluminium dome some vibration characteristics.
Anyway, I think reducing the contact area is a good idea and is worth more experiments.
I did notice that the patent diagrams did NOT show the case of -ve movement with the force concentrator added.
Eucy
Not a bad idea at all!
I've seen it being tested by certain hobbyists elsewhere quite thoroughly, initially as a "dome made from can bottoms" and eventually transitioning to other configurations, yielding interesting results. Aluminium foil, used as a sort of cover around the sound actuator or around the entire surface, has also been tested on both sides and is still being explored in this forum—there are many threads on the topic.
Oh, that patent mentions everything except the kitchen sink, yet it remains quite vague on all points. It's strange how patents are awarded for what seems to be very little at all. 😉
Hello Dave
I've only just managed to watch your videos, and I think that your work is game changing for most of us in the DIY world..
I am somewhat concerned that forces may come to bear to prevent or restrict access by folks like us when the software is released, and I can only hope that my concern will not be realised.
I have a number of questions and suggestions for you which relate to issues covered in this forum of over 600 pages (!!)
Not in any order of importance:
1/ Panel materials.. Most people here have avoided metals and solid acrylic for sonic reasons.
The sandwich materials you're working on will be of benefit if we can effectively model the characteristics of the range of materials available to the common buyer. I'm not sure if that will be a stumbling block or not.
Many members have had success with tonewoods and plywood, and it would be great to get some solid theory behind these materials, particularly in thicknesses around 3 to 4mm and that leads to:
2/ Aspect ratios and Anisotropic/Orthotropic Materials
There have been vigorous discussions here in the past regarding the effects of aspect ratios and anisotropy. Your software will allow rapid checking of the effect of aspect ratios, but if would be great if anisotropic effects can also be modelled.
3/. The effect of the "corners only" support on the response graph was quite stunning I thought. For panels of 3:1 or 4:1 aspect ratios (which are quite popular/common here), some other side supports are usually necessary on thin panels for stability. Are you able to model sectional supports with differing degrees of stiffness? ie: moderate hold near the corners, and some softer restraint at mid points on the long side etc etc.
4/ Harking back to tonewoods, it is well known that Spruce has a brighter tone than Cedar. I can envisage that this software would allow examination of the reasons for this provided that the timber characteristics can be accurately modelled.. That would be fun to do I think, and useful in establishing just what a warm or bright sound looks like in response terms. Maybe a bit outside your area of interest but...
That'll do for now, but I'll have more questions I'm sure
Thanks for your input.. It's a massive leap.
Eucy
I've only just managed to watch your videos, and I think that your work is game changing for most of us in the DIY world..
I am somewhat concerned that forces may come to bear to prevent or restrict access by folks like us when the software is released, and I can only hope that my concern will not be realised.
I have a number of questions and suggestions for you which relate to issues covered in this forum of over 600 pages (!!)
Not in any order of importance:
1/ Panel materials.. Most people here have avoided metals and solid acrylic for sonic reasons.
The sandwich materials you're working on will be of benefit if we can effectively model the characteristics of the range of materials available to the common buyer. I'm not sure if that will be a stumbling block or not.
Many members have had success with tonewoods and plywood, and it would be great to get some solid theory behind these materials, particularly in thicknesses around 3 to 4mm and that leads to:
2/ Aspect ratios and Anisotropic/Orthotropic Materials
There have been vigorous discussions here in the past regarding the effects of aspect ratios and anisotropy. Your software will allow rapid checking of the effect of aspect ratios, but if would be great if anisotropic effects can also be modelled.
3/. The effect of the "corners only" support on the response graph was quite stunning I thought. For panels of 3:1 or 4:1 aspect ratios (which are quite popular/common here), some other side supports are usually necessary on thin panels for stability. Are you able to model sectional supports with differing degrees of stiffness? ie: moderate hold near the corners, and some softer restraint at mid points on the long side etc etc.
4/ Harking back to tonewoods, it is well known that Spruce has a brighter tone than Cedar. I can envisage that this software would allow examination of the reasons for this provided that the timber characteristics can be accurately modelled.. That would be fun to do I think, and useful in establishing just what a warm or bright sound looks like in response terms. Maybe a bit outside your area of interest but...
That'll do for now, but I'll have more questions I'm sure
Thanks for your input.. It's a massive leap.
Eucy
Regarding thin panels, you might want to take a peek here at those very high-priced speakers made with thin alumina (ceramic) panels. Oh, and forget about the "moulded" part. Thin alumina sheets are quite inexpensive from China, by the way. These alumina (ceramic) panels can be integrated with other thin panels, including homemade paper. It might be of interest to explore why the metal grille cover is designed in that manner.
The panel surface...and it is convex...
The panel surface...and it is convex...
Last edited:
You might also want to have a look at this old article on standing waves. You can see that idea used in planet10's speakers. It could certainly be used in flat panels; someone might have already utilised that in this lengthy thread. 🙂
EDIT: Since I found that article, I have also come across a lengthy thread right here, which appears to be by the author of that article.
EDIT: Since I found that article, I have also come across a lengthy thread right here, which appears to be by the author of that article.
Last edited:
In response to the great feedback from @Veleric, @homeswinghome, and @Eucyblues99 - here's a screenshot of some updates. I redid the code so that the IR is shown as a spectrogram (not CWT, yet, but I can add that in the future too). I also figured out how to calculate two (very related) things... exciter impedance and exciter magnet displacement. I doubt anyone cares much about magnet displacement besides me, but there was one company that had a real issue with the magnet whacking against the plate at a certain frequency, and we had to determine where that was to filter it out effectively. Usually that frequency where the magnet is moving like crazy is one where you don't get much acoustic power output anyway, so it's usually worth filtering out.
In response to some of the other questions...
I've always stuck to homogeneous materials like acrylic or glass simply because this is what most of my corporate sponsors were interested in! You'd be surprised how many defense contractors are trying to integrate exciters into things that need to be deployed into high heat environments or places where there might be bombs - so traditional speakers won't work, and the radiating panel needs to be a specialty plastic or metal material. I'm very happy and excited to work more on integrating anisotropic and non-homogeneous materials into this model, but it'll probably take time to verify that everything works. I'll post again later today when I'm in my lab about how I'm modeling the sandwich boards - I think it'll work okay for right now 😉 And yes, I need to figure out a different name for the cutoff frequency, since F_c is already used for coincidence frequency. Oops!
I'll plan to work in adding edge damping and damping at the supports (what Eric called reflection coefficient). This isn't difficult to add into the code - it's actually more difficult to figure out how to organize the GUI effectively now that it's becoming quite crowded. I'm not a UI/UX developer, unfortunately!
The tentative plan for the software from the University is that there will be a free version for DIYers with minimal features, a low-cost version for DIYers with full features, and a version that will cost more for companies but comes with additional features (especially piezoelectric exciter support, which most of the companies I've worked with have used, rather than dynamic exciters). When I say "low cost," I also basically mean a small donation that helps support further development in my lab at work!
If I've missed any questions - sorry, there's just been a lot to respond to! Feel free to bug me again about them.
I should also note that the patent @lekha brought up about the force concentrator came from my colleagues at UR. I was around during those discussions, but I left that school in 2018 and the patent is from 2021. I'm not named on the patent... for legal reasons, I'll refrain from saying anything else right now 🙂. Here's an interesting read from almost 10 years ago, when we were thinking about starting a company.
In response to some of the other questions...
I've always stuck to homogeneous materials like acrylic or glass simply because this is what most of my corporate sponsors were interested in! You'd be surprised how many defense contractors are trying to integrate exciters into things that need to be deployed into high heat environments or places where there might be bombs - so traditional speakers won't work, and the radiating panel needs to be a specialty plastic or metal material. I'm very happy and excited to work more on integrating anisotropic and non-homogeneous materials into this model, but it'll probably take time to verify that everything works. I'll post again later today when I'm in my lab about how I'm modeling the sandwich boards - I think it'll work okay for right now 😉 And yes, I need to figure out a different name for the cutoff frequency, since F_c is already used for coincidence frequency. Oops!
I'll plan to work in adding edge damping and damping at the supports (what Eric called reflection coefficient). This isn't difficult to add into the code - it's actually more difficult to figure out how to organize the GUI effectively now that it's becoming quite crowded. I'm not a UI/UX developer, unfortunately!
The tentative plan for the software from the University is that there will be a free version for DIYers with minimal features, a low-cost version for DIYers with full features, and a version that will cost more for companies but comes with additional features (especially piezoelectric exciter support, which most of the companies I've worked with have used, rather than dynamic exciters). When I say "low cost," I also basically mean a small donation that helps support further development in my lab at work!
If I've missed any questions - sorry, there's just been a lot to respond to! Feel free to bug me again about them.
I should also note that the patent @lekha brought up about the force concentrator came from my colleagues at UR. I was around during those discussions, but I left that school in 2018 and the patent is from 2021. I'm not named on the patent... for legal reasons, I'll refrain from saying anything else right now 🙂. Here's an interesting read from almost 10 years ago, when we were thinking about starting a company.
I posted quite a few patents by you and your colleagues few pages back. A patent is just a paid article until it is utilised in production somewhere.
Anyway, there were some interesting thoughts a few years ago...
Last edited:
Dave,
That's great stuff. I do like the new spectrogram. What does CWT mean? I'm really glad to see the impedance plot, and the magnet displacement is also of some value, because, yeah, it can start slapping the board. For "final" builds I usually use a spine support for the magnet, which eliminates that issue without any apparent detrimental effect for my usual designs (where the panel fundamental is well above the magnet resonance). But I'm interested to use the model to see what it says about the effects of a spine are with more varied designs.
I'm glad to hear that you are excited about anisotropic materials. From my perspective, one of the most interesting aspects of designing flat panels is trying to tailor the elastic properties and density of the panel to optimize performance, so that pretty much requires sandwich structures. I'm eager to hear more about your current sandwich model. I'll mention again that I suspect the transition from the frequency dependent wave speed to frequency independent wave speed can be handled "seamlessly" by an orthotropic thick plate model, where the three shear moduli are used as inputs.
One other suggestion that occurred to me is that it would be nice if the acoustic response plot and the polar radiation plots were aligned directly above and below each other with the frequency scales matching. It would then be more obvious how the features of each line up with each other. For example, how the region of sharply spreading directivity in the polar plot aligns with the coincidence frequency in the acoustic response plot. Also, it would be interesting to see clearly which peaks in the acoustic response are responsible for each of the "lobed" frequencies in the polar response.
That sounds more than fair to me! I'd be glad to do it.
Thanks,
Eric
Thanks Dave.
I second Eric's interest in the effects of adding a spine, as support for exciters, esp the heavier ones, becomes an issue to prevent sag. I've been adopting a system where the exciter is tethered to the panel with a tension strap which has worked ok, but it would be excellent to lay to rest questions regarding the use of spines on the response.
Also, how difficult world it be to incorporate panel splitting? Ie.. Superimpose the results of two panels of independent size, materials and exciters. One may be optimised for bass, and one for mid and high frequencies.
The proposed pricing model also seems fair to me. Good news
Thanks
Eucy
I second Eric's interest in the effects of adding a spine, as support for exciters, esp the heavier ones, becomes an issue to prevent sag. I've been adopting a system where the exciter is tethered to the panel with a tension strap which has worked ok, but it would be excellent to lay to rest questions regarding the use of spines on the response.
Also, how difficult world it be to incorporate panel splitting? Ie.. Superimpose the results of two panels of independent size, materials and exciters. One may be optimised for bass, and one for mid and high frequencies.
The proposed pricing model also seems fair to me. Good news
Thanks
Eucy
So, Mark Bosko was the boss.
Shouldn't he be introduced to the group here? https://www.hajim.rochester.edu/ece/people/faculty/bocko_mark/index.html
He is the foremost specialist in this area at the university; perhaps someone in this group could make contact with him. He has a keen interest in flat panel speakers and exciters.
8 years or so ago...
And, at ResearchGate, https://www.researchgate.net/profile/Mark-Bocko
Last edited:
Using this earlier post as inspiration, here's a graph of the wavespeed in my gator board PETTaLS simulation with two different cutoff frequencies (which roughly defines the "corner" where the model becomes primarily nondispersive). Blue line is cutoff=infinity, yellow line is cutoff=20kHz, red line is cutoff=4kHz. I'm sure the "thick plate" model is a better fit, but this might work okay for the time being!
Also, @Eucyblues99 I did build the "spine support" into the software, I was just calling it "fixed" exciter vs "free" exciter. It assumes that the exciter magnet is totally unable to move - spines can often be resonant themselves, but hopefully those effects can be neglected.
Continuous Wavelet Transform. It is a time to frequency transform to get spectrogram. A wavelet has interesting mathematical properties that go probably beyond our need here. Basically a wavelet is a band pass filter. It allows to build a filter bank that will decompose the signal (the impulse response) over the frequency domain.
Hi Dave, thanks.
Can you then briefly summarise the general effects of spine vs no spine on a typical frequency response, or is it not so simple?
Intuitively, the effect will vary with panel mass and size, and with the mass of the exciter.
Eucy
Last edited:
Eucy,
Can't you wait until Dave publishes the model so you can see for yourself (haha)? I have a million similar questions for Dave but have been trying to hold off and see what the model shows!
Eric
I would direct those questions to Mark Bosko himself. He is still in charge of the faculty, and the "patents" are his publications, with students contributing as well. He has more than 41 years of experience experimenting with this subject: flat panel speakers.
Last edited: