Hi Oliver,
OK i think i understand ...
When you look at the laser holographs from the manger transducer,
you see clearly that it operates in a modal fashion, even at the upper frequencies.
The lowest resonances are even visible in the impedance plot.
Very similar behavior like a rigid panel transducer.
What is the main difference between the bending stiff panels
called "DML" and a transducer like Manger ?
I think it is the coincidence frequency beeing above the
audible band in case of the manger, because the propagation
speed of the bending waves is lower than on a rigid panel.
This allows for high modal density on a small membrane area.
But there are also bending wave transducers in rigid panel style
introduced with coincidence frequency at the upper limit of
hearing.
The membrane not excited in the center may be a further "DML"
feature.
And the Manger having a surround which strongly reduces
reflection ...
But a difference in principle of operation ?
It is merely a difference in dimensioning and materials used:
- where do you want to adjust the Q of your modes ?
- what about modal density and modal overlap ?
For good sound quality, you need modal overlap to be high
enough, so that single modes are not audibly detectable.
The undamped rigid panel style DML which has damping only
via sound radiation is IMO not suited for high quality audio
reproduction. How should it be ?
As i have seen in recent papers the tendency goes towards
increasing the coincidence frequency and/or to account for damping
in those cases, where sound quality plays a role.
Whether this coincides with the initial approach of rigid panel
DML is a wording problem IMO and should not be our problem
in diyAudio at least
...
OK i think i understand ...
When you look at the laser holographs from the manger transducer,
you see clearly that it operates in a modal fashion, even at the upper frequencies.
The lowest resonances are even visible in the impedance plot.
Very similar behavior like a rigid panel transducer.
What is the main difference between the bending stiff panels
called "DML" and a transducer like Manger ?
I think it is the coincidence frequency beeing above the
audible band in case of the manger, because the propagation
speed of the bending waves is lower than on a rigid panel.
This allows for high modal density on a small membrane area.
But there are also bending wave transducers in rigid panel style
introduced with coincidence frequency at the upper limit of
hearing.
The membrane not excited in the center may be a further "DML"
feature.
And the Manger having a surround which strongly reduces
reflection ...
But a difference in principle of operation ?
It is merely a difference in dimensioning and materials used:
- where do you want to adjust the Q of your modes ?
- what about modal density and modal overlap ?
For good sound quality, you need modal overlap to be high
enough, so that single modes are not audibly detectable.
The undamped rigid panel style DML which has damping only
via sound radiation is IMO not suited for high quality audio
reproduction. How should it be ?
As i have seen in recent papers the tendency goes towards
increasing the coincidence frequency and/or to account for damping
in those cases, where sound quality plays a role.
Whether this coincides with the initial approach of rigid panel
DML is a wording problem IMO and should not be our problem
in diyAudio at least
...
Last edited:
I think there is an intermediate transition from what is "modal" behavior
to a "statistical" behavior, characterised by modal density and overlap
being as high, that the modes cannot be identified anymore.
When it comes to energy storage: Which decay time is tolerable ?
If you would state - which i do not believe - that a DML must have
identifyable modes and a long decay, then a DML would be
d e f i n e d by its poor quality.
The NXT developers will probably not be amused from such
a definition.
Is there a decay time limit separating "conventional bending transducers"
from "DML" ?
I think in fact a DML also aims for "hiding" the modal character by
statistical behavior (high modal density and overlap).
Which decay time is tolerable may be a point worth discussing.
But i think there is also an efficiency vs. quality conflict.
A panel lacking any damping might go very loud even when
using a weak and cheap exciter.
Quality increases manufacuring cost ... i guess this observation is
not entirely new and also applies to other kinds of stuff.
Kind Regards
My personal definition of conventional bending wave operation is that the wave doesn't reach the edge.
What I am interested in is the frequency-dependence of damping. How is it in air? And what materials come close?
In the early results from the prototypes i posted the waves are far from
being absorbed before reaching the edges and there is also reflection.
If e.g. in the Manger transducer there was no reflection from the
edges, there would not be visible resonances in the impedance
curve.
With a perfect termination at the edges the waves would look like
propagating beyond the edges and disappear without reflection,
like waves in an infinite large pool, where the form of the panel is
just a picture frame restricting your view. A perfect termination would
simulate an infinitely large panel.
Simply speaking this is why also a rigid panel style DML gets better
the larger it is - i see your point but i feel that the distinction is
rather artificial.
Using your definition even an upsized DML would not be a 'DML' anymore.
Maybe it is the term itself which is not very instructive.
But we are all on thin ice here, because there are no tools or
established quality measures to describe sane limits for the
bahaviour of these kind of speakers, be it
"conventional bending wave" or "DML".
There is still work to do i feel.
Btw: What kind of fish is the Göbel transducer in your personal opinion ?
It woud be a DML up to a lower frequency.
The importance of distinguishing between DML and not-DML is lobing. In the frequency range where one according to my defintion has a true bending wave transducer with multiple exciters one has the same lobing problems as with pistonic speakers.
About the Goebel:
My impression was that it has ultra-high resolution at low levels (even higher than in reality), but gets into trouble at higher levels.
Would you classify it as "DML" or "Conventional Bending Wave Transducer"
according to your definition ?
Assuming extremely high damping of the panel or whatever used as membrane,
a multiactuator panel would at high frequencies in fact show similar behavior
like multiple pistonic tweeters mounted apart, given that the exciters cover
the same frequency range.
But also - and especially - with a low damped panel you have disturbing
interference effects of the bending waves on the panel itself, when exciters
radiating the same frequency are only a few halfwavelenghts
(here: bending wavelengths on the panel) apart. That effect gets even
worse with a "low loss" panel.
This is why it is mostly disregarded to have more than one exciter for the
highs. As i remember you told me once, that the highs of the podium did not
meet your taste ... and this might possibly be one of the reasons.
I assume the "Podium" is a true "DML" according to your classification ?
Made of thick and rigid honecomb core, Shelley Katz states he loves the
material because it "rings like a bell" ...
The art is IMO to adjust damping of membrane and suspension in a way to maintain
bending wave propagation on the membrane to support frequency independent
dispersion but also have a smooth and fast spectral decay.
There is no perfect recipe how to achieve that, many solutions are possible.
A panel and its suspension having extremely low loss is only good for making
a mechanical reverberator (plate reverberator) - as we know such devices
have been in use in the past decades for studio reverberation - but are not
suitable as quality loudspeakers.
I personally do not think it is useful to make a distinction between different
bending wave speaker concepts as if they were different principles.
To me a Podium is a Podium, a Manger is a Manger and a Linaeum is a Linaeum
for example.
There are differing technical solutions applied (and often patented) by each
manufacturer to achieve the best compromise or the "optimum" due to ones
philosophy.
I would summarise those speakers simply as "bending wave speakers" or as
"constant velocity speakers", because the average velocity magnitude across
the membrane area is nearly frequency independent, whereas the conventional
pistonic speaker is mainly a "constant acceleration" speaker.
a multiactuator panel would at high frequencies in fact show similar behavior
like multiple pistonic tweeters mounted apart, given that the exciters cover
the same frequency range.
But also - and especially - with a low damped panel you have disturbing
interference effects of the bending waves on the panel itself, when exciters
radiating the same frequency are only a few halfwavelenghts
(here: bending wavelengths on the panel) apart. That effect gets even
worse with a "low loss" panel.
This is why it is mostly disregarded to have more than one exciter for the
highs. As i remember you told me once, that the highs of the podium did not
meet your taste ... and this might possibly be one of the reasons.
I assume the "Podium" is a true "DML" according to your classification ?
Made of thick and rigid honecomb core, Shelley Katz states he loves the
material because it "rings like a bell" ...
The art is IMO to adjust damping of membrane and suspension in a way to maintain
bending wave propagation on the membrane to support frequency independent
dispersion but also have a smooth and fast spectral decay.
There is no perfect recipe how to achieve that, many solutions are possible.
A panel and its suspension having extremely low loss is only good for making
a mechanical reverberator (plate reverberator) - as we know such devices
have been in use in the past decades for studio reverberation - but are not
suitable as quality loudspeakers.
I personally do not think it is useful to make a distinction between different
bending wave speaker concepts as if they were different principles.
To me a Podium is a Podium, a Manger is a Manger and a Linaeum is a Linaeum
for example.
There are differing technical solutions applied (and often patented) by each
manufacturer to achieve the best compromise or the "optimum" due to ones
philosophy.
I would summarise those speakers simply as "bending wave speakers" or as
"constant velocity speakers", because the average velocity magnitude across
the membrane area is nearly frequency independent, whereas the conventional
pistonic speaker is mainly a "constant acceleration" speaker.
Last edited:
Hi guys,
I'm sort of back, been very busy with work related stuff.
I took a look again on the Audiocircle forum and appearantly I ticked off a guy named "zygadr" (at least I think it was me). He feels that there are people (like me) questioning his results and ask the same questions over and over again. He refers to diyaudio around the date that I started my replies.
I think the same questions keep popping up due to little "evidence". If I state that a certain material is THE BEST THERE IS, OMG I CAN'T BELIEVE MY EARS!! Then people like me are interested in reproducing this sound and try to get into the details to how this sound could be so good/bad.
Please post pictures and detailed information of the build and preferably with measurements. If these lack, someone else can never rebuild them.
People tend to forget that a sound which sounds great to one, may sound awfull to the other.
Let's not forget that I am interested in NXT tech. and do not think they are toys, but I question certain claims and would like to see them verified by measurements (and I don't care about the straightness of a graph, a graph is a handy tool).
So, zygadr if I ticked you off, sorry mate
I'm sort of back, been very busy with work related stuff.
I took a look again on the Audiocircle forum and appearantly I ticked off a guy named "zygadr" (at least I think it was me). He feels that there are people (like me) questioning his results and ask the same questions over and over again. He refers to diyaudio around the date that I started my replies.
I think the same questions keep popping up due to little "evidence". If I state that a certain material is THE BEST THERE IS, OMG I CAN'T BELIEVE MY EARS!!
Then people like me are interested in reproducing this sound and try to get into the details to how this sound could be so good/bad.Please post pictures and detailed information of the build and preferably with measurements. If these lack, someone else can never rebuild them.
People tend to forget that a sound which sounds great to one, may sound awfull to the other.
Let's not forget that I am interested in NXT tech. and do not think they are toys, but I question certain claims and would like to see them verified by measurements (and I don't care about the straightness of a graph, a graph is a handy tool).
So, zygadr if I ticked you off, sorry mate
What are those claims you question particularly ?
What feels a a little strange to me is that NXT has managed somehow,
that all bending wave related technology is widely associated with
their name. Be it "flexurally rigid" panels or not.
Teragaki Lab
Just to have some pictures from another part of world.
Lets wait until someone calls the constructions of Mr.
Teragaki "NXT Panels" ...
Just to have some pictures from another part of world.
Lets wait until someone calls the constructions of Mr.
Teragaki "NXT Panels" ...
... in his age he might be able to give it a smile.
What is called "NXT Technology" arises from testing materials
like honeycomb for the purpose of making aircraft cabins.
Exciters were applied for sound measurement puposes.
A loudspeaker evolved accidently.
The principle was investigated and improved.
We should not forget that early violin makers did similar things
some hundred years ago and some loudspeaker manufacturers
did it many decades ago.
De facto any conventional fullrange speaker falls into the category
"bending wave transducer" when leaving its individual frequency range
of pistonic operation.
And as in instrument making there are even in conventional
loudspeaker driver design discussions about "the best" material to choose.
If we cannot specify the material properties we want, the material discussion
runs in circles.
To me the greatest chance in bending wave transducers is to establish a
new paradigm in working with the material instead of against it.
As it is obvious, that a lightweight perfectly rigid cone of sufficiently
large area, which covers the whole band of human hearing with pistonic
movement is impossible for many reasons. With respect to the frequency
dependent polar dispersion it is not even desirable.
To face that reality and acknowledge, that even latest and best conventional
fullrange drivers have to give up the pistonic paradigm ( Is it a dogma in fact ?)
is a great step forward.
Over there in the ESL threads, it is also stated that a large membrane
driven by a uniformly spread force across its area does not mean it is
moving uniformly over the whole area.
Bending waves where you look. Having accepted that, we should not
expect to find the one and only material, because there are too
many degrees of freedom in construction.
What is called "NXT Technology" arises from testing materials
like honeycomb for the purpose of making aircraft cabins.
Exciters were applied for sound measurement puposes.
A loudspeaker evolved accidently.
The principle was investigated and improved.
We should not forget that early violin makers did similar things
some hundred years ago and some loudspeaker manufacturers
did it many decades ago.
De facto any conventional fullrange speaker falls into the category
"bending wave transducer" when leaving its individual frequency range
of pistonic operation.
And as in instrument making there are even in conventional
loudspeaker driver design discussions about "the best" material to choose.
If we cannot specify the material properties we want, the material discussion
runs in circles.
To me the greatest chance in bending wave transducers is to establish a
new paradigm in working with the material instead of against it.
As it is obvious, that a lightweight perfectly rigid cone of sufficiently
large area, which covers the whole band of human hearing with pistonic
movement is impossible for many reasons. With respect to the frequency
dependent polar dispersion it is not even desirable.
To face that reality and acknowledge, that even latest and best conventional
fullrange drivers have to give up the pistonic paradigm ( Is it a dogma in fact ?)
is a great step forward.
Over there in the ESL threads, it is also stated that a large membrane
driven by a uniformly spread force across its area does not mean it is
moving uniformly over the whole area.
Bending waves where you look. Having accepted that, we should not
expect to find the one and only material, because there are too
many degrees of freedom in construction.
Last edited:
Sound Advance ST2416S-3D7 2x2 SpeakerTile Optima/Ultima Pr | Parts-Express.com
seems an nxt implementation, nominal specs are very good considering panel size 89db, -4 @ 60hz
seems an nxt implementation, nominal specs are very good considering panel size 89db, -4 @ 60hz
Hi!
I discovered NXT drivers a little while ago having taken an "inflatable" speaker apart. I took the 4 inch square panels out and used them as horn drivers for a pair of horns I made with a son and his friend for fun one weekend . . . We wanted something really indestructable! They worked!
Recently I wanted a wide non directional sound to come from a large surface area on a wall to provide reverberation in an organ installation:
YouTube - Alfred Hollins Concert Overture in C minor (1899) - Jeremy Filsell
YouTube - Hugh Potton plays Reubke Sonata 94th Psalm PART 1
It struck me that the inflatable speaker transducers were just the thing so took another couple apart. I put them onto panels of transparent Correlux - a pair of 6mm 2ft x 4ft and another pair, one on a 2ft square double skin 10mm board and another on a triple wall 1ft x 5ft panel about 12 mm thick.
Results were astounding. Supported by minimum mounting points the sound was incredible. However, as soon as they go near a wall, bass disappears.
At present they are hung by a pair of strings on the top corners, and lifted up so that at the top they are about 9 inches from the wall, being attached to the wall at the bottom glued to a couple of felt pads, more resting on the top of the pad than stuck to it with any surface area. The result is astounding. People cannot believe that they produce the sound heard . . . And the sound is very immediate - something akin to what I'd expect an electrostatic to sound like.
The thinner panels are a bit more mellow and the thicker ones give more treble.
What happened to Ziggy? Why did his account get disabled?
Best wishes
David P
I discovered NXT drivers a little while ago having taken an "inflatable" speaker apart. I took the 4 inch square panels out and used them as horn drivers for a pair of horns I made with a son and his friend for fun one weekend . . . We wanted something really indestructable! They worked!
Recently I wanted a wide non directional sound to come from a large surface area on a wall to provide reverberation in an organ installation:
YouTube - Alfred Hollins Concert Overture in C minor (1899) - Jeremy Filsell
YouTube - Hugh Potton plays Reubke Sonata 94th Psalm PART 1
It struck me that the inflatable speaker transducers were just the thing so took another couple apart. I put them onto panels of transparent Correlux - a pair of 6mm 2ft x 4ft and another pair, one on a 2ft square double skin 10mm board and another on a triple wall 1ft x 5ft panel about 12 mm thick.
Results were astounding. Supported by minimum mounting points the sound was incredible. However, as soon as they go near a wall, bass disappears.
At present they are hung by a pair of strings on the top corners, and lifted up so that at the top they are about 9 inches from the wall, being attached to the wall at the bottom glued to a couple of felt pads, more resting on the top of the pad than stuck to it with any surface area. The result is astounding. People cannot believe that they produce the sound heard . . . And the sound is very immediate - something akin to what I'd expect an electrostatic to sound like.
The thinner panels are a bit more mellow and the thicker ones give more treble.
What happened to Ziggy? Why did his account get disabled?
Best wishes
David P
Hi!
I'm posting a footnote here for completeness . . . Whilst at school many years ago, a friend brought in a panel exciter which was rather interesting - and it looks as though they are still made:
Invisible Speakers: Official Website for the Rolen-Star Audio Transducers
Best wishes
David P
I'm posting a footnote here for completeness . . . Whilst at school many years ago, a friend brought in a panel exciter which was rather interesting - and it looks as though they are still made:
Invisible Speakers: Official Website for the Rolen-Star Audio Transducers
Best wishes
David P
LineArray, great information, I've enjoyed reading your other posts as well.
David P, I've spoken over the phone with several transducer suppliers. They seem to discourage any attempt at making a real speaker out of their sound makers. I'm not sure why other than to discourage possible disappointment and failure associated with their products.
Hi!
Well despite such discouragement, go ahead and do it. There's Visaton and Parts Express . . . and some of those units are on special offer. Just as the units are designed for multiple mode, then use multiple panels and the idiosyncracies of one size or thickness of panel will fit in to fill the gaps in another.
A friend called in at the house today with some friends and _insisted_ I let them hear my panels . . . and were appropriately impressed.
So don't get deterred!
Best wishes
David P
A commercial application...
I didn't go through and read the whole thread (sorry, lazy)... but I wanted to throw this into the "mix". I "invented" and patented (don't get me started on the word "inventor" or "patents" as they both turn my stomach) a "piezo based foam speaker". It's the core IP for the "singing balloon" made by Anagram Intl. (under "Sing-A-Tune). Years ago I was tasked with putting a piezo element on a foil balloon. If the balloon is fully inflated (or better still over-inflated) and a piezo is attached to the surface you can get some really great sound from it (relatively speaking of course and depending on the source). The problem comes in when the balloon begins to deflate (exacerbated by most stores over-inflating in the first place which stretches the film, I digress). When it deflates the sound goes away completely. This wouldn't make a commercial product so something had to be attached to piezo so that it made sound when deflated. It is well known that when a piezo is attached to a flat section of foam (think "musical greeting cards") it amplifies the sound. However, it only works for a set of frequencies. Given that it's a greeting card playing music, no one really cares. I thought, lets make a "speaker cone" replacing the magnetic assembly with the piezo. This sounded much better. However the film still had to go through the "conversion process"; this involves a number of rollers, heated "heads" etc. so a strict "cone" wouldn't do (it would get crushed). So, I made a "collapsible" speaker cone. It's a series of concentric rings with a place for the piezo element to go in the center. During conversion it goes through the rollers and collapses down and then back to it's original form. When I had it tested (in the 6" outside diameter used on the balloon) it performed virtually the same as a 2 1/4" (toy as we call them) speaker. Not bad for something that weighs less than 10 grams. Bass response can be increased considerably by increasing the size of the diaphram (which increses the weight adding more tilt to the balloon and affects floatability, a no-no). Unfortunately someone else took over the project a few years ago and the sound quality has gone all to hell (I won't get into that)...
The other consideration I had is using the piezo element (also called a "bender"). These are *typically* high voltage devices; 40Vp-p is not uncommon. The sound IC's max out at 6V so I used a PWM scheme to drive the piezo (common now) to get 12Vp-p across the piezo. I had a design to supply 40Vp-p but the added costs made it prohibitive (when you make 35 million of something pennies add up fast). I used a "double sided" piezo as well; these have crystals on both sides of the brass disk and perform better than single sided in this application.
I spoke with NXT recently about using their speakers for a guitar amp design. It was very odd talking with these guys; I got the weird impression that they didn't really want to sell their stuff (or something). They were always talking about the limitations of the system and even when I said I could design the system around the limitations they were kind of stand-offish (you had to be there). I know of one guitar amp that is using their stuff (so they say). The "other" transducer they have is nothing more than a speaker "coil" with no spider cage or cone. I haven't tried one but Parts Express has them really cheap. I worked with piezos for about a year doing all sorts of "sound" experiments with them. You can do some pretty cool things with them...
I didn't go through and read the whole thread (sorry, lazy)... but I wanted to throw this into the "mix". I "invented" and patented (don't get me started on the word "inventor" or "patents" as they both turn my stomach) a "piezo based foam speaker". It's the core IP for the "singing balloon" made by Anagram Intl. (under "Sing-A-Tune). Years ago I was tasked with putting a piezo element on a foil balloon. If the balloon is fully inflated (or better still over-inflated) and a piezo is attached to the surface you can get some really great sound from it (relatively speaking of course and depending on the source). The problem comes in when the balloon begins to deflate (exacerbated by most stores over-inflating in the first place which stretches the film, I digress). When it deflates the sound goes away completely. This wouldn't make a commercial product so something had to be attached to piezo so that it made sound when deflated. It is well known that when a piezo is attached to a flat section of foam (think "musical greeting cards") it amplifies the sound. However, it only works for a set of frequencies. Given that it's a greeting card playing music, no one really cares. I thought, lets make a "speaker cone" replacing the magnetic assembly with the piezo. This sounded much better. However the film still had to go through the "conversion process"; this involves a number of rollers, heated "heads" etc. so a strict "cone" wouldn't do (it would get crushed). So, I made a "collapsible" speaker cone. It's a series of concentric rings with a place for the piezo element to go in the center. During conversion it goes through the rollers and collapses down and then back to it's original form. When I had it tested (in the 6" outside diameter used on the balloon) it performed virtually the same as a 2 1/4" (toy as we call them) speaker. Not bad for something that weighs less than 10 grams. Bass response can be increased considerably by increasing the size of the diaphram (which increses the weight adding more tilt to the balloon and affects floatability, a no-no). Unfortunately someone else took over the project a few years ago and the sound quality has gone all to hell (I won't get into that)...
The other consideration I had is using the piezo element (also called a "bender"). These are *typically* high voltage devices; 40Vp-p is not uncommon. The sound IC's max out at 6V so I used a PWM scheme to drive the piezo (common now) to get 12Vp-p across the piezo. I had a design to supply 40Vp-p but the added costs made it prohibitive (when you make 35 million of something pennies add up fast). I used a "double sided" piezo as well; these have crystals on both sides of the brass disk and perform better than single sided in this application.
I spoke with NXT recently about using their speakers for a guitar amp design. It was very odd talking with these guys; I got the weird impression that they didn't really want to sell their stuff (or something). They were always talking about the limitations of the system and even when I said I could design the system around the limitations they were kind of stand-offish (you had to be there). I know of one guitar amp that is using their stuff (so they say). The "other" transducer they have is nothing more than a speaker "coil" with no spider cage or cone. I haven't tried one but Parts Express has them really cheap. I worked with piezos for about a year doing all sorts of "sound" experiments with them. You can do some pretty cool things with them...
Don't mean to hijack the thread, but perhaps I can find someone interested in piezos to help me. I want to possibly use a piezo as a supertweeter in an open baffle atop a full range speaker. The full range is 8 ohm, and I'd like to keep the total impedence at about 8 ohm. I'd also like to roll off the bottom end from the tweeter starting at about 8 to 10 KHz and have some ability to adjust the output to match it optimally to the full range. Any advice on a "crossover" that would allow me to do this?
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.