It’s bliss to know nutting, you start impossible things….

Sarel.wagner

AVForums Member
Joined
Jun 15, 2022
Messages
110
Reaction score
69
Location
Centurion
Yes, I am an idiot and know nutting, so now I want to start experimenting with horns and flat panel or DML speakers, not for the same speaker tho. Or maybe I can combine them, no :unsure::LOL: The common factor here is FR drivers (exciter). As far as speaker design goes, I am an ignoramus, never before attempted anything of the sort. Building speaker coffins, maybe about a dozen or so. Subs for various uses, also quite a few. But never anything to do with Horns, TLs, TQWTs or DMLs. Just some background for relevance.

I have various bookshelf speakers by Mission, B&W, Monitor Audio and a set of near mint Yamaha NS690 Mark IIIs, TLS Refrence Sub and Yamaha active studio speakers. In the past had some Tannoys, Bostons, Polk, Jamo, AR etc etc. For headphones I currently own the HP1 Yamaha Orthodynamics restored, AKG, Sony and Tascam TH-02. So a good spread of speakers and sound generators to compare to.

I am in the process of auditioning, in the next few weeks, my first pair of FR speakers. Hopefully, all indications of their pedigree seems to indicate classy, they will work out and I will buy them. Will post about those soon. This is for my Dynaco SCA-35 integrated amp.

All this is because of how my mind works. I just have to understand stuff. Horn speakers, and even more so DMLs, intrigues me. Just because the flat panels workings are completely different to cone type drivers. Speaker drivers with modern lower efficiency and higher power handling vs higher efficiency but lower power handling and full range units is another intriguing trait. The difference between modern 2 way and multi way speakers and their phasing and timing vs FR systems where some of these traits does not exist, is attractive, altho they have different traits. And yes, everything is a compromise of some sort. Both types are compromised.

I do have, and use, REW with a Minidsp calibrated microphone, and my guess is I will use it quite a bit more for the experimentation here. The real world, unlike ideal formulas and mathematical equations, is not ideal, indeed flawed, as are the rooms where we use these speakers. It is fairly common knowledge that the room normally dominate LF response and not the coffin and its driver.

Lots of reading up, and experimentation to follow, nevermind getting all the materials to do the experimentation with. Then the brainstorming and finding creative solutions to crappy response curves, bad materials properties and all the compromises to make the process of creating a new design somewhat acceptable. Maybe all of this turns out to be a bust and wast of time and money. At least have some fun and learn something of this fools errand 🤔🤷‍♂️

Groetnis
PS: do chime in, I need all the help on here please. Also it’s a nice sound board, no pun intended….
 

Sarel.wagner

AVForums Member
Joined
Jun 15, 2022
Messages
110
Reaction score
69
Location
Centurion
:) at some point, I had to learn of these myself, way when NXT I think, brought flat panels to the market. This was about 2000 or 2001.

Groetnis
 

Sarel.wagner

AVForums Member
Joined
Jun 15, 2022
Messages
110
Reaction score
69
Location
Centurion
I have been reading up on DML panels for a number of years. There are some research papers, lots flying over my head, and some that are understandable. There are some YouTube vids on these as well. The volume of work in this area is substantially less, from what I could see, than for tradition piston/cone based systems. So what are the differences then? Lets start with the invention of the dynamic driver.

1874 - Ernst W. Siemens filed the first patent, shorty thereafter Bell also did, for a device that can be considered as the grandfather of the speaker. In 1925 Chester Rice and Edward Kellog at GE defined the speaker principles and the baffle to prevent rear waves from cancelling the front waves. Herman Danger filed his coaxial patent in 1928. In 1929 Kellog filed for the Electrostatic radiator. During 1940 Paul Klipsch filed for the corner Horn Speaker.

Then in 1996 the Verity Group in Britain formed New Transducers Ltd, now the NXT company , to develop the Distributed-Mode Loudspeaker (DML) based on the 1991 patent by Dr Ken Heron of Britain's Defence Evaluation & Research Agency and in 1998 Benwin sold the first DML flat panel loudspeakers. I only heard of these with NXT marketing them later. The patent of Dr heron stemmed from analysis of the vibrations and resonances in airframe panels. This lead to the realisation that a vibrating panel can generate sound.

After that condensed view of the speaker driver history, we see that DMLs are still very new and young, a mere 21 years, barely out of puberty. Ahh and yes, we can see the pimples on them still.... :) The operating principle cannot, I guess, be more different. Traditional cone based speakers make use of rather stiff cone material that will eliminate distortion. They are designed for specific narrow audible bands in an attempt to have the cone (actually the complete electromagnetic engine) behave as optimally as possible in its frequency range to prevent distortion.

Below picture shows the ranges and this equates to about 10 Octaves. Each doubling of a frequency is an octave.
photo_2022-06-27 11.18.03.png


The original drivers were full range with not to good frequency response. They also, most likely, suffered from lots of distortion, it's a guess. As these drivers were better understood, and more research done on them, designers realised that stiffer cones, and smaller ones, lessened distortion. This to some degree, plus the desire and availability of more powerful amplifiers, drove the current speaker designs. Then one day recently, from left field, came a new vision.

DML (Distributed Mode Loudspeaker) is a flat vibrating panel really. This makes use of an exciter, the voice coil assembly plus suspension, and is bolted or glued to some structure. Certain structures perform better than others. Unlike traditional driver units, the exciters excite the panel in many frequencies at the same time, many modes of vibration simultaneously. More to follow....

Groetnis
 

Sarel.wagner

AVForums Member
Joined
Jun 15, 2022
Messages
110
Reaction score
69
Location
Centurion
The concept of a frequency versus sensitivity (amplitude) audiogram plot of human hearing was conceived by German physicist Max Wien in 1903. From this time on, we started to understand our own hearing better, listening, aah but that is a different story.... Psycoacoustics: Psychoacoustics combines the study of acoustics and auditory physiology to determine the relationship between a sound's characteristics and the auditory sensation that it provokes. Loudness is the subjective measure of perceived sound intensity.

The separation of our ears, the shape of our heads and the source(s) of the sound play together (no pun) to our enjoyment of critical listening. Also, our ears are not the only way we can detect or enjoy sound. Not to speak of the room we sit in to listen to music and its effects on what we actually hear or percieve. This is because of the influence of the room on the speaker driver and enclosure and on the sound reproduced. We also have the effects of the tonal balance of the sounds or music (relative distribution per octave) and the frequency masking effect or any other effects on the sound or music, on our hearing.

Then we have the Haas effect which explains that when two identical sounds occur within 30 milliseconds of one another (sometimes up to 40ms), we perceive them as a single event or sound. A recording trick used frequently is to duplicate a lead singers voice to a second mono channel, delay one channel slightly (likely just enough to hear them as separate but not so much to be perceived as echo) and then pan the two channels to create a simulated stereo effect from one mono recording. Our sources are no so pure or rather faithful after all. All studios manipulate us, to like what they do more....

Our hearing capabilities is based on perception of the reproduced sound and the specific acoustic environment we listen in. No wonder there are such diverging opinions on SQ and what we may or may not like. The original compression CoDec technology by Fraunhofer Institute, MP3, is a lossy CoDec. A lossy compression codec works on the theory that, as the human ear is already discarding a lot of information in the perception of sound, you might as well simply not encode this redundant information. So most digital music, even some lossless CoDecs, lose some information based on Psychoacoustic principles.

What does all this have to do with what I want to experiment with then? Kinda everything. How we hear, and perceive, music is as much a factor of the room and speakers as Psychoacoustics. Flat panel speakers are new and not yet fully understood me thinks. Just my opinion. I do have lots of questions, and lots more by the day. I also have lots of ideas in my head that I want to try.

I will document those on here.

Groetnis
 

Sarel.wagner

AVForums Member
Joined
Jun 15, 2022
Messages
110
Reaction score
69
Location
Centurion
Lets get real. What are the differences between dynamic cone type drivers and DML panels really?
The traditional speaker driver uses a piston type motion to produce pressure waves we hear as sound. The piston idea stems from the fact that the complete cone must move as one, ie the inner (middle) part connected to the voice coil must move at the same time with the same amount as the outer edge of the cone. If this does not happen, distortion is the result. Stiffness, like with many things 🤔 are vital:D A lot of attention and engineering is focused on resolving cone issues.

FR drivers are different. Their cones are designed and engineered to be less stiff to allow the higher frequencies to be produced. The cone must be stiff enough for the piston movement to produce lower octaves, but flex enough to produce the higher octaves and not start ringing.

With dynamic speaker drivers, the front and rear of the cone produces pressure waves and thus sound. The front and rear waves are out of phase and this is destructive in nature if not kept separate, hence the baffle or enclosure. (No, infinite baffle is a misnomer, it will have to be of infinite size, no really. Engineering, not marketing.) Ergo, half the energy is wasted and we attempt to keep em inside the coffin, ports are an attempt to harvest some of that energy and give it back to the room. There are also other reasons ports are used but none the less, ports always adds energy back to the room.

These drivers are point sources for sound, ie the sound is radiated from a point and has directionality. Yes, it radiates in a spherical manner, all sound does except where phase arrays are involved. With phase arrays we can steer sound in a direction and off axis (without physically moving anything), the sound volume (intensity) is reduced dramatically in comparison to the on axis. This however, is for another day.

Due to the fact that traditional speakers are point sources and have directionality, there is a sweet spot for critical listening, the triangle, and toe in. The off axis sound gets reflected by the room, normally all 6 surfaces and any objects in the room collectively does this. It's a complex pressure distribution and creates multiple high and low pressure points. These reflections may be additive or destructive, depending on frequency and the room reflections. The direct sounds are desirable, reflected sounds are less so. This is oft referred to as the room sound. Some rooms sound dead (fewer reflections or due to the Haas effect) or alive (more reflections) or muddy (timing of reflections vs direct sounds again Haas effect and others). Most of these traits are stemming from the fact that the sound generator is a point source.

Lets compare traditional speakers to musical instruments and the human voice.
Our voices can be considered as somewhat of a point source, or an Aerophone. Aerophones are instruments like a flute, trumpet and others that vibrate the air, just like our voices. However, Aerophones are not point sources, indeed most any musical instrument is like this. String and percussion and most other types of instruments are also not point sources. They have many point sources, but these are distributed over a larger area. Strings, for example, emit sound from it’s full length, except for the nodal points themselves. This also means there are no out of phase waves generated for the most part. Then we have Timber, this is the harmonics or overtones of an instrument or a voice. The quality given to a sound by its overtones: such as the resonance by which the ear recognises and identifies a voiced speech sound or the quality of tone distinctive of a particular singing voice or musical instrument.

Consider this, the Blue-man Group drumming with them paint drums. Clearly visible is the vibrations across the whole area making the paint droplets fly into the air at varying velocity and distances. It shows the vibrations spread across all of the surface.

ACA501DD-45ED-460C-B76E-6160F48C25A2.jpeg D8A422CE-715E-4919-8BC6-383D07ADDB76.jpeg

Flat panel speakers or DMLs (Dynamic Mode Loudspeakers)
These devices are not point sources. Very importantly, they radiate sound all over the panel, front and rear, at the same time and in phase. Almost no radiation from the sides. No coffin box required. Currently, depending on materials, they may not be as loud as traditional speakers, but they can be. They are very simplistic devices, but may require some EQ or DSP manipulation to get the best out of em. The basic speaker including exciter is cheap as chips, or could be if no exotic materials are used for the panels or frames.

Because these DMLs are not point sources, they can be used in multiples if required, with no phasing or cancellation effects. They are typically thin, no coffin, and can be placed close to side walls and relatively close to back walls. Well that is what the research indicates, we shall see. They are an easy load to drive, have very good coupling to the air and good energy transfer from electrical to sound.

A few points to consider. They have less base than desired, normally. This is panel size dependent, and also panel material influenced. There may be tonal balance issues, same dependencies. The panel, due to material, sizes and treatments applied, exciter placement and panel thickness as well as inherent properties like flex and dumping, may make them perform better or worse than expected.

More to follow, groetnis
 

Sarel.wagner

AVForums Member
Joined
Jun 15, 2022
Messages
110
Reaction score
69
Location
Centurion
So enough already, when does me starts? Wellll, stock really, rather the lack thereof. Awaiting the exciters, on order with as yet unknown arrival time. Do have XPS foam, do have the wrong plywood, need Baltic Birch ply 6mm, or other zero void ply. Do have Aluminium, perspex and need to get the composite stuff. But first I have to learn about honeycomb and working with composites. Do have a vacuum pump but need to get vacuum bags to use for a press.

Currently busy designing and working on the panels sizes themselves to try and figure out what makes sense to me. (Well really just speculating and doodling mostly) Everyone and their dog (me including my Rottie) does something else, this is both good and bad. I will look at this as good :) Due to internal reflections, most people agree square ain't no good. Lotsa people are referring to the 2/5 vs 3/5 ratios. This could be used for the panel as well as placement of the transducer or exciter if you will. So design ideas and then drawing up some stuff to visualise things. Lots of messing around is seen in the future....

This bigger is better syndrome thing, ai... Bigger is mo bestest, since size vs frequency response is a thing here. To get to a 44Hz audible target we need a 1 square meter panel surface, this is however starting to fall off a cliff at 166Hz, so no bueno. 1/2 a square meter panel is dead below 212Hz and barely audible at 150Hz :eek: Conversely a 2.8 square panel starts dropping off at 88Hz but dang, that is a full ply or ally panel size.... To be realistic, a 1.3 square meter panel will be audible, barely, at 20Hz and start falling off a cliff at 150Hz.

Ideally, we want to get to at least the Mid base Octave at around 60Hz..... Hmmm wonder how? Maybe thinner panel material, certainly not a bigger panel no. [email protected] Less stiff or more stifferer panels is also an option. This is, I guess, where them composites get attractive, and mo expensive too.... Thinner but more stiffer, carbon composite honeycomb anybody? But first the basics. Manageable panel sizes and thickness to start off with, get some empirical data and do careful measurements. Go from there, time and paper is patient, me, not so much.


At least got some instruments to do measurements with, so that should be good, no.
Groetnis
 

Sarel.wagner

AVForums Member
Joined
Jun 15, 2022
Messages
110
Reaction score
69
Location
Centurion
Around the year 1500:
Leonardo Da Vinci wrote “an average human looks without seeing, listens without hearing, touches without feeling, eats without tasting, moves without physical awareness, inhales without awareness of odor or fragrance, and talks without thinking.” I can only hope that last bit is not me. Wait but why?

The start of Psychoacoustics:
Aristotle (around 350 BC) was the first to suggest that sound is carried by air movement. But Leonardo De Vinci (around 1500) was likely the first to realize that such movement was probably in the form of waves. Galileo Galilei, 100 years later, scraped a chisel across a brass plate, producing a screechy pitch. Galileo calculated that the spacing of the grooves caused by the chisel was related to the perceived pitch of the screech. However, it wasn’t until the 17th century that the relationship between vibratory frequency and pitch was confirmed. Robert Hooke (1635-1703) made a wheel with small teeth sticking out from the edge at equal intervals. As the wheel rotated on an axle and the teeth pressed on a card, a sound was produced when the card vibrated. The pitch of the sound rose as the wheel’s rotational speed increased.

And here we are today, trying to figure out vibrations to make sound. As someone once said: They keep on stealing my inventions more than a 100 years ago! :ROFLMAO:

Groetnis
 
Last edited:

Ingvar Ahlberg

AVForums Super Veteran
Joined
Jan 16, 2014
Messages
1,899
Reaction score
693
Location
The Swedish westcoast
Flat panel speakers or DMLs (Dynamic Mode Loudspeakers)
These devices are not point sources. Very importantly, they radiate sound all over the panel, front and rear, at the same time and in phase.

Sorry byt that is not correct, almost all types of planar speakers, from the realy lousy NXT versions from the late 60:s to and including all good electrostatics and magnetostatics they all radiate both forward and backwards but out of phase, as far as I know the only speaker driver type that radiates both ways but in phase is air motion transformers and ESS/Heil laboratories actually made an attempt to a "full range" version in the 70:s, it was said to work but at 800 kg each it didn´t make it to any exhibitons, as far as I know.

A plane wave speaker is only about frontal radiating size, nothing else, it´s simple to build a plane wave source out of normal dynamic drivers, just use as many as You need in a panel large enough, in a "normal" listening room in domestic circumstances the Audio Tronic Megatrend speakers worked as a plane wave source at least up to 10 meter listening distance, modern high quality cinema installations (not talking about HT) do this in the entire auditorium, using only dynamic speaker drivers.

Ingvar
 

Ingvar Ahlberg

AVForums Super Veteran
Joined
Jan 16, 2014
Messages
1,899
Reaction score
693
Location
The Swedish westcoast
But perhaps the Teragaki designed speakers from the late 70:s early 80:s did what You are aming at, reasonably stiff but light wood panels mounted with mechanical tension (i.e they were bent so that the panels were at a high state of mechanical tension) and excited from a single point bya a type of voice coil, the vibration pattern/radiation should reasonably be in phase both forward and bacwards.

They were regarded as extremely good by the few listeners that wrote about them, of the very limited quantity produced none left Japan as far as I know.

Ingvar
 

Smitty

New member
Joined
Dec 22, 2021
Messages
34
Reaction score
9
Location
Cape Town
Have you watched the Tech Ingredients videos on DML / flat panel speakers?

IIRC they have built and tested a lot of what you are thinking about, for example, size, material, exciter placement, ways of suspending the panels.

 

Sarel.wagner

AVForums Member
Joined
Jun 15, 2022
Messages
110
Reaction score
69
Location
Centurion
@Ingvar Ahlberg Thank you for chipping in. Your experience over many years with speaker is invaluable. Maybe an elaboration on what I actually tried to explain is in order as the previous was way to terse. I will dig out the actual research used and reference those later.

In a word, when a panel is exited, there are many points (modes) of excitation at the same time for a frequency, and all frequencies will have their modes present at the same time. Each point of excitation will produce a pressure wave so that many pressure waves will be generated all over the panel This is diffused radiation, correct no?

All the modes for a specific frequency will be in phase on the front side of the panel. At the same time all the same modes or point on the back side of the panel will also be in phase with each other. This was what I wanted to convey.

You are correct in that the front and rear are out of phase (they have to be since the front and rear are directly coupled to each other) and I am aware that I did not make that clear. Also there are almost no radiation out the sides of the panels at the boundary surfaces.

I hope I got it right this time Ingvar…. 🤦🏻‍♂️🤔 Will look out for the Teragaki design and hopefully find some info, never heard of that before, no wonder.

@Smitty I have seen some of that yes, did not pay much attention to most youtube stuff, was focused on the fundamentals and research papers more than anything else, to try and understand what the principles are first.

Groetnis
 

Sarel.wagner

AVForums Member
Joined
Jun 15, 2022
Messages
110
Reaction score
69
Location
Centurion
258E1E57-B015-463B-9076-F5C7AB5D7FDA.png

This is a screen grab from a thesis here:Thesis link A conventional speaker was compared to a DML and measurements taken every 5 degrees at 1m distance. The beaming of the conventional speaker can clearly be seen and the diffuse nature ( and rear radiation) of the DML

Groetnis
 

Ingvar Ahlberg

AVForums Super Veteran
Joined
Jan 16, 2014
Messages
1,899
Reaction score
693
Location
The Swedish westcoast
Hi Sarel, it was not meant as a poke in the eye, the thing about phase, probably with very good design You can approach back and front radiation in phase byt probably only in limited bandwith, an AMT is totally in phase back/front over the entire bandwith, that was my only point.

But what You are doing is interseting and exciting, and do check out Teragaki, one of those strange Japanese geniouses&/secrets:

TERRA-SP3000 of Teragaki-Labo―寺垣武が開発する寺垣スピーカー「TERRAシリーズ」を製造・販売

Not much info to find that is not in Japanese but there was a few interesting articles and interviews in Swedish magazines in the early 80:s, will try to find, should also be some French write up´s on this man, most of of it covering his magnificent torntable thou.

Ingv
 

Sarel.wagner

AVForums Member
Joined
Jun 15, 2022
Messages
110
Reaction score
69
Location
Centurion
My original understanding of how this DML thing works was not very good. It took some time to find enough good info to visualise the workings.

There seems to be agreement in general, that this picture represents how the panel vibrations look and work. This was a single low frequency. Since there are two modal nodes in blue and one of opposite magnitude in red, there will be some cancellation as they are out of phase, this leads to lower sound output. Below is exaggerated for effect. Panel moves a few millionths of a mm, up to about 20 or 30 millionths.
Screenshot 2022-06-29 at 12.02.48.png


This animation is of a different panel and clearly illustrates the flexing and patterns of modes.
dml-displacement-animation.gif


Now we can at least visualise what happens when the panel is excited.

Groetnis
 

Sarel.wagner

AVForums Member
Joined
Jun 15, 2022
Messages
110
Reaction score
69
Location
Centurion
I have found some fancy shmancy analytics as well as predictive modelling on flat panels, the above images are from some of those. There are many commercial tools used to do this. From FEA to do modelling and predictions on flat panels, to Laser vibrometers and sound analytics to FFT analytics and many more besides.

Anybody got one of them Laser vibrometers I can borrow Please? 😬

I have got a calibrated microphone and REW to do audio tests and am in the process of getting an audio component testing system for other performance analytics. One other thing from a tools perspective, is DSP. Since flat panels are not fully understood yet, and there are lots more to discover, tools are required to help understand panels better. Ears are also tools. 👂 Tools to make anything I have aplenty.

Panel response currently still is one of the biggest issues. Tonal balance is kind of not ideal, depending on what it is made out of mostly, but not entirely. This depends on a number of factors, some of them being;
  • Panel material
  • Panel size
  • Panel aspect ratio
  • Panel mounting
  • Panel thickness
  • Panel Material Isotropy or the lack thereof
  • Panel material density
  • Panel material stiffness
  • Exciter type
  • Exciter response curves
  • Exciter placement position
  • Exciter mounting
  • Lots of thing I do not know yet
  • Lots of things to be discovered yet
There is, in the case of flat panels, at least in the lower octave response, a correlation between size or surface area rather and frequency response. There may be more to it than just the size, stiffness, thickness and other may also play a lesser or bigger role in this.

Groetnis
 
Top