DIY Flat Panel Speaker Love

[bradley.s]

I've been going back and forth on how to support the array. Finally decided on wood railing over aluminum angle. I'm soaking the wood in mineral oil to deaden it a little.

[Unbiasedsound]

Because Don Keele said flat CBTs work and researches said MAP arrays work. That's pretty much the only reason other than DMLs have the characteristics you mentioned.

How are your CBT panels coming along?

[bradley.s]

Because Don Keele said flat CBTs work and researches said MAP arrays work. That's pretty much the only reason other than DMLs have the characteristics you mentioned.

[Kornbread]

Mine have a very diffuse kind of sound and nulls to the extreme sides but otherwise a large area in which they sound about the same seated, standing, or to the sides. There is a sweet spot though. I probably missed it, but why cbt with this type transducer?

[bradley.s]

bradley.s You might find this youtube vid interesting. If you are placing multiple exciters on a panel with this many nodes? where do you place the additional exciters so that they act constructively instead of destructively? IIRC, that paper did mention something about the damping of the panel materiel and a minimal separation distance between exciters. If only we had some modeling software ...

I'm shooting from the hip on that issue. I'm hoping my low volume level listening requirements will mitigate the majority of the problem. I'm also going to lean heavily on DSP. All the research I've personally seen investigates homogeneous single layers like a single sheet of aluminum or they use foam sandwiched with paper. The panels I'm going to try are hacked together. I have no idea what the internal dampening is on these things. It's a crap shoot. In addition, my geometry is also way off. I'm using very narrow panels relative to typical DMLs. With regard to normal DMLs, mine are a freaking mess.

What I'm hoping to accomplish is a CBT polar response in the 500hz to 6,000hz frequency range. CBT seems easy to handle below 500hz and hard to do above 6,000hz. If I can get the bulk of the frequencies working I can build a separate DML to handle below 500 and either forget about frequencies above 6,000 or try to figure out an alternative.

[Unbiasedsound]

Anyone notice something in the vid of the resonance experiment that pertains to the shape of DML panels? One of the standard proven successful techniques of DML panels is to ROUND the corners. Notice how the salt stays away from the very tip of the corners in fact the shape usually cuts off the tip of the corners.

[Unbiasedsound]

At least it gives me some comfort that you agree that research can build on past research, otherwise, I respectfully disagree with you on just about everything else.

With an end to school in sight and, fingers crossed, a job change, I hope to finally have enough time to attend a few diy meets. I'd love to hear your $60 speakers, that cost $2k and 4 years to build, that can't be beat for $60, and sound as good as any $2k speaker. I'll even buy the drinks.

bradley.s You might find this youtube vid interesting. If you are placing multiple exciters on a panel with this many nodes? where do you place the additional exciters so that they act constructively instead of destructively? IIRC, that paper did mention something about the damping of the panel materiel and a minimal separation distance between exciters. If only we had some modeling software ...

What do you specifically disagree with me on? During my 4 years on many different forums I see people post measurements that leads nowhere and most usually give up since they don't know how to make there panels sound better. All the measurements in the world wont tell you how to make ones panels sound better. If it did everyone would have great sounding panels by now and or way better sounding panels them my own Bastats but they don't why? Its funny how you will rely on some modeling software but ignore techniques used by successful DML designs.

I most likely will never attend any DIY meets due to my location. BUT you are welcome to come visit and listen to my panels anytime. Although I wont show you or tell you how I built them. LOL

[Kornbread]

Research builds upon research proving once again that it is a theory. Theories can always change while FACTS remain the same forever.

Even though P.E. provides measurements they don't tell you much about said product then the ACTUAL CUSTOMER REVIEWS, which is the reason P.E. even has customer reviews.

No, wrong analogy. EXPERIMENTATION cost around $2K and 4 years of time.... My Bastat panels cost 60.00 to build but has the sound quality of around 2k speakers. Which means if I market my speakers at 2K each and sell just 2 pairs I would get back my 2k loss in experiments.

The reason I asked you that is because NOT MANY people have heard the higher quality commercialized DML speakers because they are rare. Not many have heard Podiums, not many have heard Bertagni's etc....A lot of people on the other hand have heard the more POPULAR conventional cone type of speakers. I got the hi fidelity fever ever since I got my first kenwood rack stereo system set in the 70's. I was obsessed with audio gear but on a working mans beer budget. In my life time I have purchased over 100 pairs of speakers and 20 different subs most of them from second hand. One of my favorite past times is to visit all the local audio stores and just listen to them demo speakers and other audio gear so I've heard in person many many different speakers in person regardless if Ive owned them or not.. Ive heard Maggies, Quads, Carver Amazing , Apogee's, Vandersteens, Martin logan, Kefs,focal, Dynaudio, Mirage, Def Tech, Totem, Polks JBL, Klipsch, tannoys etc etc etc.

For a failure to become a success one must improve on the failure or they will just fail again. After improving on a failure its called a success. You can either repeat the failed attempt or repeat the improved attempt that succeeded. LOL

At least it gives me some comfort that you agree that research can build on past research, otherwise, I respectfully disagree with you on just about everything else.

With an end to school in sight and, fingers crossed, a job change, I hope to finally have enough time to attend a few diy meets. I'd love to hear your $60 speakers, that cost $2k and 4 years to build, that can't be beat for $60, and sound as good as any $2k speaker. I'll even buy the drinks.

bradley.s You might find this youtube vid interesting. If you are placing multiple exciters on a panel with this many nodes? where do you place the additional exciters so that they act constructively instead of destructively? IIRC, that paper did mention something about the damping of the panel materiel and a minimal separation distance between exciters. If only we had some modeling software ...

[Unbiasedsound]

In general, research papers are academic fact based articles as void as possible of bias and conjecture which could serve to invalidate the findings. Research often builds upon research.

As mentioned, I do not sense measurements tell everything about how a product sounds, just the vast majority. This is a good thing about PE, they go to the trouble of including accurate measurements for many of their products, even if it entails taking their own.

By your own admission, a $60 dml has cost $2k and 4 years of time. $2k is a lot of coin and could build a killer system.

Unbiasedsound asked a question of me pertaining to listening to certain brands of dml panels. A reversal of questioning is fair. How many, and I use this phrase loosely, 'high-end' and well set-up speaker systems have you sat down and listened to? What brands, models, where, etc. We have no baseline to base your frame of reference when you say " ... can sound like speakers costing around $2k".

The internet is at your disposal. Just a modicum of research should yield many products, some life saving, that were initial failures that became successes.

Research builds upon research proving once again that it is a theory. Theories can always change while FACTS remain the same forever.

Even though P.E. provides measurements they don't tell you much about said product then the ACTUAL CUSTOMER REVIEWS, which is the reason P.E. even has customer reviews.

No, wrong analogy. EXPERIMENTATION cost around $2K and 4 years of time.... My Bastat panels cost 60.00 to build but has the sound quality of around 2k speakers. Which means if I market my speakers at 2K each and sell just 2 pairs I would get back my 2k loss in experiments.

The reason I asked you that is because NOT MANY people have heard the higher quality commercialized DML speakers because they are rare. Not many have heard Podiums, not many have heard Bertagni's etc....A lot of people on the other hand have heard the more POPULAR conventional cone type of speakers. I got the hi fidelity fever ever since I got my first kenwood rack stereo system set in the 70's. I was obsessed with audio gear but on a working mans beer budget. In my life time I have purchased over 100 pairs of speakers and 20 different subs most of them from second hand. One of my favorite past times is to visit all the local audio stores and just listen to them demo speakers and other audio gear so I've heard in person many many different speakers in person regardless if Ive owned them or not.. Ive heard Maggies, Quads, Carver Amazing , Apogee's, Vandersteens, Martin logan, Kefs,focal, Dynaudio, Mirage, Def Tech, Totem, Polks JBL, Klipsch, tannoys etc etc etc.

For a failure to become a success one must improve on the failure or they will just fail again. After improving on a failure its called a success. You can either repeat the failed attempt or repeat the improved attempt that succeeded. LOL

[Kornbread]

Wrong its not fact based SMDH its THEORY based.

Greater things have arisen from success.

I beg to differ. Measurements don't tell you fawk all about how a product actually sounds....A persons own experience DOES tell a lot more of how a product actually sounds. P.E. customer reviews tell a lot of said products sound. It a ton of people said it sounds good then more people buy the product , if a ton of people say it sounds bad not many buy that product which is all based on customers feed back.

All the measurements in the world wont tell you how to make your DML panels sound better. What will make your panels sound better is proven SUCCESSFUL techniques.

I am not sure if it was you or someone else that said that one of the reasons they stopped experimenting with DML panels is because all there DIY conventional cone speakers sounded better then DML panels, proving once again that person aint never heard a good sounding DML panel...… My bastats DO sound as good as some conventional cone speakers or better. Also I am not saying they are the best speakers in the world but if you want to compare best sound quality for the $$$ it cant be beat as my bastats cost less then $60.00 to build a pair (sub not included but required) and can sound like speakers costing around 2k IMO.

In general, research papers are academic fact based articles as void as possible of bias and conjecture which could serve to invalidate the findings. Research often builds upon research.

As mentioned, I do not sense measurements tell everything about how a product sounds, just the vast majority. This is a good thing about PE, they go to the trouble of including accurate measurements for many of their products, even if it entails taking their own.

By your own admission, a $60 dml has cost $2k and 4 years of time. $2k is a lot of coin and could build a killer system.

Unbiasedsound asked a question of me pertaining to listening to certain brands of dml panels. A reversal of questioning is fair. How many, and I use this phrase loosely, 'high-end' and well set-up speaker systems have you sat down and listened to? What brands, models, where, etc. We have no baseline to base your frame of reference when you say " ... can sound like speakers costing around $2k".

The internet is at your disposal. Just a modicum of research should yield many products, some life saving, that were initial failures that became successes.

[Unbiasedsound]

The research paper referenced is "a technical review of the development of a special type of planar loudspeaker array for wave field synthesis, known as multiactuator panel". It is not add hype for some new speaker manufacturer. It is a fact based research paper that simply lays groundwork for others to build upon if they choose to do so. I for one am glad it, and similar research, has been freely shared and encourage others to continue the vibrating panel experiment.

Great things have arisen from failure.


On to sound quality. I am one of those that believes there are some things we cannot measure but can hear, but I also feel 95ish'% of what we can hear can be measured. Most people on this forum want to see measurements, not hear someone gush over how there latest creation is so much more betterer than anything else everer. That tells us very little about how a product actually sounds.

Repeatable, accurate, and real measurements please. Everyone has a laptop, calibrated mics are not that expensive and there's some very accurate low-cost/free software (REW is one of many) floating around the net.

Wrong its not fact based SMDH its THEORY based.

Greater things have arisen from success.

I beg to differ. Measurements don't tell you fawk all about how a product actually sounds....A persons own experience DOES tell a lot more of how a product actually sounds. P.E. customer reviews tell a lot of said products sound. It a ton of people said it sounds good then more people buy the product , if a ton of people say it sounds bad not many buy that product which is all based on customers feed back.

All the measurements in the world wont tell you how to make your DML panels sound better. What will make your panels sound better is proven SUCCESSFUL techniques.

I am not sure if it was you or someone else that said that one of the reasons they stopped experimenting with DML panels is because all there DIY conventional cone speakers sounded better then DML panels, proving once again that person aint never heard a good sounding DML panel...… My bastats DO sound as good as some conventional cone speakers or better. Also I am not saying they are the best speakers in the world but if you want to compare best sound quality for the $$$ it cant be beat as my bastats cost less then $60.00 to build a pair (sub not included but required) and can sound like speakers costing around 2k IMO.

[bradley.s]

There's a 2016 paper titled, "Optimized Driver Placement for Array-Driven Flat-Panel Loudspeakers" by David A. ANDERSON, Michael C. HEILEMANN, Mark F. BOCKO, but it's located at an unsecure link.

It's also too large to upload as an attachment.

"The recently demonstrated ‘modal crossover network’ method for flat panel loudspeaker tuning employs an array of force drivers to selectively excite one or more panel bending modes from a spectrum of panel bending modes. A regularly spaced grid of drivers is a logical configuration for a two-dimensional driver array, and although this can be effective for exciting multiple panel modes it will not necessarily exhibit strong coupling to all of the modes within a given band of frequencies. In this paper a method is described to find optimal force driver array layouts to enable control of all the panel bending modes within a given frequency band. The optimization is carried out both for dynamic force actuators, treated as point forces, and for piezoelectric patch actuators. The optimized array layouts achieve similar maximum mode coupling efficiencies in comparison with regularly spaced driver arrays; however, in the optimized arrays all of the modes within a specified frequency band may be independently addressed, which is important for achieving a desired loudspeaker frequency response. Experiments on flat panel loudspeakers with optimized force actuator array layouts show that each of the panel modes within a selected frequency band may be addressed independently and that the inter-modal crosstalk is typically −30 dB or less with non-ideal drivers."

"6. Conclusion A method was described for finding the optimal force actuator locations for controlling a selected set of bending modes of a flat panel. In the applications of specific interest in this paper (flat panel loudspeakers) it is important to be able to exert independent control over each of the modes in a low-pass frequency band. The geometry of the modes in such a set may vary greatly, which implies that the layout of the actuator array that would be most effective at driving each mode is unique, since a mode is driven most effectively by actuators near a mode’s antinodes. An optimization method was employed to determine the actuator locations that are able to drive every mode in a selected set as equally as possible and to spread the work evenly among the force actuators in the array. Simulations show that either point-force actuators or piezoelectric bending actuators have global optima that give large coupling efficiency between the actuators and modes."

[bradley.s]

Here's a different paper titled, "Optimized Driver Placement for Array-Driven Flat-Panel Loudspeakers"

http://www.czasopisma.pan.pl/Content/101307/PDF/aoa-2017-0010.pdf

[geosand]

Well I think its good to see that the tech is still being researched...The science (computational modelling), the measurements and the build techniques are all needed to progress in the design. Here's the latest on Optimal exciter placement: http://www.aes.org/e-lib/browse.cfm?elib=20277 ... too bad its only behind a pay wall. I wonder how the usual 2/5|3/5 rule matches their computational methods...

[Unbiasedsound]

True, the paper didn't discuss that. On the other hand, both Tectonic and Billionsound stabilize the exciter. Consequently, I lean toward stabilizing the exciters.

One of the main reasons I believe it was a failure is because it didn't use a spline to stabilize the exciter.