Yes, this effect of side by side panels bleeding over into the next panel was first noted in one of the DML threads years ago, by a guy from Mexico. It may have been in Ziggy's old AC thread. He found that playing one panel in close proximity to a second panel that was undriven, caused the second panel to emit sound as if it was part of the first driven panel. Its an interesting effect and seems very dependent on the proximity of the panels to each other...I'm guessing millimeters apart at best, but could be more...Sorry don't have time to look up what thread that was in...
geo
Yes in Azima's widely circulated paper that was printed in Audio magazine he stated: "[FONT=Verdana, Arial, Helvetica, sans-serif][SIZE=-1]But the panel itself operates wholly in resonance of course, which is one of the features of NXT that most concerns audio people raised to regard resonance as anathema. Doesn't all this resonance in the panel colour the sound unacceptably? The surprising answer is no, it doesn't, because of the highly complex nature of the panel's vibration. The impulse response of an NXT panel (Figure 4) displays a long resonant `tail' which would damn any conventional loudspeaker, but in fact the sound quality has extraordinary clarity and transparency confirming the measured flat frequency response (Figure 5). [/SIZE][/FONT]
geo,
Thanks, I did not recall that part of the Azima paper.
It just reinforces my feeling that my own impulse measurement is somehow flawed or different.
Here's the REW impulse response I measured for one version of my DML, and it's typical or virtually every DML measurement I've made, regardless of the panel material, boundary conditions, exciter position, etc:
And for comparison, below is the impulse response of my Alpair 10.3M MLTL. A bit different but still nothing like what Azima or XRK is showing. Is it my speakers? My room? My measurement settings?
Thanks, I did not recall that part of the Azima paper.
It just reinforces my feeling that my own impulse measurement is somehow flawed or different.
Here's the REW impulse response I measured for one version of my DML, and it's typical or virtually every DML measurement I've made, regardless of the panel material, boundary conditions, exciter position, etc:
And for comparison, below is the impulse response of my Alpair 10.3M MLTL. A bit different but still nothing like what Azima or XRK is showing. Is it my speakers? My room? My measurement settings?
Yep, I noticed that too on my re-review of that video. But I think it's worth noting that the tapering profile in the TI video really looks nothing like the Acoustic Black Hole (ABH) profile described in the paper Burnt referred too. So, while I'm far from sure the ABH concept can be (easily) successfully applied to a DML, I'm also not ready to accept that it can't, just because in the TI video they tried something that seemed like a similar concept. The ABH profile is a shape with some very specific characteristics, and requires combination with damping materials applied at the thinned section to be most effectively applied. I frankly suspect it can be successfully applied, but probably requires someone with much deeper understanding than myself!
Eric
Hover mouse pointer into the yellow marker area as first visual below and pick "%" sign, then adjust vertical and horizontal scales to something usefull coverage, BTW the two IR below is accurate the same session but use a different setting for presentation : )
Attachments
Last edited:
Agreed, I think it can be done, but I've "read" that paper 3 times and still don't understand the ABH concept or how its totally implemented...and the math was way beyond my evidently very limited brain capacity :-(
I will give an explanation a try.
The ABH principle is based on the surface waves that produce the sound we hear going through transition as the material thins. The velocity of the wave slows as the material thins and in compensation the wave height increases. If you have the profile right in theory the velocity slows to zero and the waves form standing waves on the edge. If you combine this with an absorption layer this energy is transformed to heat.
One way of thinking of it is like waves on the beach. If you look some distance to the peaks of the waves isn’t very big, they look like fat ripples. As the waves reach the beach, which is a tapering profile, the waves slow and the wave height increases. The waves crash on the beach expending a lot of energy which is not reflected back. If you looked at a beach from a height you would see smooth ripples of waves travelling towards the beach and a lot of energy at the edge of the sea where the waves are rising and then crashing onto the beach itself. A DML doesn’t have an interface between a beach and the rest of the sand where the energy is expended but the absorption material, e.g. a bead of sealant or a layer of soft polymer, can act in the same manner.
Any old profile does not work, it has to be a precise power law curve to get the effect working.
The ABH principle is based on the surface waves that produce the sound we hear going through transition as the material thins. The velocity of the wave slows as the material thins and in compensation the wave height increases. If you have the profile right in theory the velocity slows to zero and the waves form standing waves on the edge. If you combine this with an absorption layer this energy is transformed to heat.
One way of thinking of it is like waves on the beach. If you look some distance to the peaks of the waves isn’t very big, they look like fat ripples. As the waves reach the beach, which is a tapering profile, the waves slow and the wave height increases. The waves crash on the beach expending a lot of energy which is not reflected back. If you looked at a beach from a height you would see smooth ripples of waves travelling towards the beach and a lot of energy at the edge of the sea where the waves are rising and then crashing onto the beach itself. A DML doesn’t have an interface between a beach and the rest of the sand where the energy is expended but the absorption material, e.g. a bead of sealant or a layer of soft polymer, can act in the same manner.
Any old profile does not work, it has to be a precise power law curve to get the effect working.
Thank you Geo, that sounds VERY interesting. I have two new XPS panels arriving soon so I will experiment and feed back.
P.S. is anyone else getting funny looks from their family when you ask for XPS panels for Christmas?
That is basically like XRK's foam core enclosure. Most (hard) wood enclosures mostly reflect sound. In foam core enclosure instead of reflecting sound it absorbs it and resonates. This resonance can produce more bass, for some reason it reminds me of a passive radiator, maybe similar characteristics not sure but it does work from my experience.
I've not been on the forum for a while, but about 10 days ago, I had a group of friends round for one of our regular music & hifi sessions. My 3mm poplar ply DML over 15 inch OB speakers got their first exposure to a wider audience.
To cut a long story short, the speakers were a big hit. Comments like 'open' and 'timed well' along with other desirable aspects such as sound stage and tonal balance also received positive comments.
The overall setup wasn't expensive. Raspberry pi with Hifiberry SPDIF HAT feeding Khadas tone board DAC and then on to a £30 Chinese valve preamp, in turn feeding a miniDSP which then output to a pair of TDA7498E boards in dual mono setup in turn driving the speakers. All cabling DIY solid core except the SPDIF cable.
In previous testing, it was the solid core speaker cables (runs of different gauge copper magnet wire in unshrunk protective heatshrink) that made a significant difference for the better. I've always had good experiences with solid core cables regardless of the rest of the setup.
Given that no part of the system had much more than 30 hours on the clock, I'd say this was a very decent outcome.
One aspect that I need to address is a slight hardness somewhere in the upper midrange / lower treble. It could be room acoustics, exciters not yet bedded in, valves with insufficient time or any one of a number of possibilities. However, I have no plans to go back to either conventional box speakers, or OB speakers with all cone drivers.
One of the above friends will be coming round sometime in the next couple of weeks to try out his very expensive Townshend speaker isolators on them and also a very decent CD player to see if the Pi or DAC is a weak link in the chain.
The speakers were also regarded favourably in comparison with electrostatics and it was felt that they were in Magnepan territory in overall character. Because I use a 15 inch bass driver and cross over at 320hz with an LR48 crossover, the exciter is relieved of a lot of power handling requirements, meaning that more than acceptable volumes are possible from just a single exciter per panel (I use the Dayton Audio 24w unit). Treble seems well extended and there would appear to be no need to mess around with additional tweeters or supertweeters. I have measured good bass output to 30hz.
So, overall, it has to be my successful audio project yet both in terms of sound quality and value for money. Some rather more aesthetically pleasing woodwork might be on the cards in a year or two!
To cut a long story short, the speakers were a big hit. Comments like 'open' and 'timed well' along with other desirable aspects such as sound stage and tonal balance also received positive comments.
The overall setup wasn't expensive. Raspberry pi with Hifiberry SPDIF HAT feeding Khadas tone board DAC and then on to a £30 Chinese valve preamp, in turn feeding a miniDSP which then output to a pair of TDA7498E boards in dual mono setup in turn driving the speakers. All cabling DIY solid core except the SPDIF cable.
In previous testing, it was the solid core speaker cables (runs of different gauge copper magnet wire in unshrunk protective heatshrink) that made a significant difference for the better. I've always had good experiences with solid core cables regardless of the rest of the setup.
Given that no part of the system had much more than 30 hours on the clock, I'd say this was a very decent outcome.
One aspect that I need to address is a slight hardness somewhere in the upper midrange / lower treble. It could be room acoustics, exciters not yet bedded in, valves with insufficient time or any one of a number of possibilities. However, I have no plans to go back to either conventional box speakers, or OB speakers with all cone drivers.
One of the above friends will be coming round sometime in the next couple of weeks to try out his very expensive Townshend speaker isolators on them and also a very decent CD player to see if the Pi or DAC is a weak link in the chain.
The speakers were also regarded favourably in comparison with electrostatics and it was felt that they were in Magnepan territory in overall character. Because I use a 15 inch bass driver and cross over at 320hz with an LR48 crossover, the exciter is relieved of a lot of power handling requirements, meaning that more than acceptable volumes are possible from just a single exciter per panel (I use the Dayton Audio 24w unit). Treble seems well extended and there would appear to be no need to mess around with additional tweeters or supertweeters. I have measured good bass output to 30hz.
So, overall, it has to be my successful audio project yet both in terms of sound quality and value for money. Some rather more aesthetically pleasing woodwork might be on the cards in a year or two!
That is good to hear M'lord 
I have also abandoned my ESL's. I loved them, and still do, but they are inferior to DML's in many areas, attack, soundstage, timbral accuracy. Any percussive instrument like piano, drum, cymbal, sound soft and unconvincing after DML's.
I also find one exciter on a panel will go plenty loud without strain if you cross them over to a sub. Glad you got such a positive reception to your panels
I have also abandoned my ESL's. I loved them, and still do, but they are inferior to DML's in many areas, attack, soundstage, timbral accuracy. Any percussive instrument like piano, drum, cymbal, sound soft and unconvincing after DML's.I also find one exciter on a panel will go plenty loud without strain if you cross them over to a sub. Glad you got such a positive reception to your panels
Mixing Exciters
The DML's I've built so far have used two exciters per panel. But I've been playing around with putting 3 to 6 exciters on the same panel. Using lots of exciters really seems to smooth the frequency response. Further, I've been mixing different model exciters in the same panel to further smooth the response, as different exciters do better in different ranges. For example, the Tectonic TEAX32C30-4/B gives good frequency response, while the Dayton Audio DEAX32U-4 provides better low frequency response. At least that's how they work on my plywood panels.
Are there any issues I should be aware of with respect to mixing different exciters? In particular with respect to damaging my amp or the exciters? Is there any reason not to, say, wire two different models of exciters together in series? Or in parallel?
Is there any problem with mixing exciters with different power ratings? So far, I'm using only 4 ohm exciters in the 20-30 Watts range. But I have some 4 ohm 40 watt exciters too, that I could try mixing in (but have not yet).
What if say I wanted to use 4 exciters on a panel, 2 of model A, and 2 of model B on the same panel, with two series pairs combined in parallel? Should the pairs in series be the same mode (i.e. A/A and B/B) Or should the series pairs be mixed (i.e. both A/B)? Does it matter at all?
Are there any particular ways of wiring multiple different exciters together that I should avoid? I know enough not to put too many in parallel and reduce the overall resistance below 4 ohms. But what about using arrangements with 12 or 16 ohms resistance (i.e. 3 or 4 exciters all in series). Is there any danger in that? Any other things I should avoid?
Eric
The DML's I've built so far have used two exciters per panel. But I've been playing around with putting 3 to 6 exciters on the same panel. Using lots of exciters really seems to smooth the frequency response. Further, I've been mixing different model exciters in the same panel to further smooth the response, as different exciters do better in different ranges. For example, the Tectonic TEAX32C30-4/B gives good frequency response, while the Dayton Audio DEAX32U-4 provides better low frequency response. At least that's how they work on my plywood panels.
Are there any issues I should be aware of with respect to mixing different exciters? In particular with respect to damaging my amp or the exciters? Is there any reason not to, say, wire two different models of exciters together in series? Or in parallel?
Is there any problem with mixing exciters with different power ratings? So far, I'm using only 4 ohm exciters in the 20-30 Watts range. But I have some 4 ohm 40 watt exciters too, that I could try mixing in (but have not yet).
What if say I wanted to use 4 exciters on a panel, 2 of model A, and 2 of model B on the same panel, with two series pairs combined in parallel? Should the pairs in series be the same mode (i.e. A/A and B/B) Or should the series pairs be mixed (i.e. both A/B)? Does it matter at all?
Are there any particular ways of wiring multiple different exciters together that I should avoid? I know enough not to put too many in parallel and reduce the overall resistance below 4 ohms. But what about using arrangements with 12 or 16 ohms resistance (i.e. 3 or 4 exciters all in series). Is there any danger in that? Any other things I should avoid?
Eric
VELERIC
I have built many panels with many combinations of drivers from 1 up to 8 per side and even one big single panel (7 square meters) in a left-center-right arrangement, and as long as you stay above the recommended impedance (4ohms) for your amplifier you should never have any problem.
if you go above 8ohms the available power will be limited, but that won't damage anything.
Though you should keep a reasonable balance in the power distribution among your drivers. 2x4 ohms in serie could be put in parallele to one single 8ohms of a different flavour, result will be 4ohms...
Different power ratings are not a problem but they may have different level output for the same wattage, and that could be disturbing.
the way your wire the drivers in serie-parallele or the other way is not important, but if you try with something like 32 drivers or more, the lenght or wire may come in play.
At the end of the journey, you will discover by yourself that one single good driver, well placed on a good panel, is all what you need
happy 2020 to all the people
I have built many panels with many combinations of drivers from 1 up to 8 per side and even one big single panel (7 square meters) in a left-center-right arrangement, and as long as you stay above the recommended impedance (4ohms) for your amplifier you should never have any problem.
if you go above 8ohms the available power will be limited, but that won't damage anything.
Though you should keep a reasonable balance in the power distribution among your drivers. 2x4 ohms in serie could be put in parallele to one single 8ohms of a different flavour, result will be 4ohms...
Different power ratings are not a problem but they may have different level output for the same wattage, and that could be disturbing.
the way your wire the drivers in serie-parallele or the other way is not important, but if you try with something like 32 drivers or more, the lenght or wire may come in play.
At the end of the journey, you will discover by yourself that one single good driver, well placed on a good panel, is all what you need
happy 2020 to all the people
Exciter options
Looking at exciter options I also noticed that Mouser and DigiKey stock these PUI Audio exciters, as well as Visaton exciters. RS can supply Visaton and Tectonic. These PUI exciters are mostly very cheap (similar to Tectonic), but I have not seen a single post or mention where someone used one of these? Maybe they don't sound great and the universal rule - you get what you pay for - applies here? Yet, the Tectonics are popular despite being very cheap. Visaton exciters on the other hand are mostly quite expensive and also seldom get mentioned. Should I just stick to sourcing the Dayton (middle of the road price wise) exciters?
This is great news for newbies like me who don’t feel like testing different panel materials, exciter and size combinations. Question is still – which exciter is that one single good driver and what is a good panel (material and size) to build a really nice sounding DML? I, know, it is similar to a question like – what is the best sounding FR driver? Too many other variables I guess. I'm just not sure where to begin yet.