Hello
At this step, I am not sure to fully understand want is your design. I think I can make a synthesis of what was discussed here : a no exciter support situation is acceptable for testing (short term) not for long term use. The weight of the magnet applies a vertical fatigue force on the suspension. the difficulty of the support design it to avoid any constraint on the voice coil. Different solutions were shared here (spline, vertical ribbon...)
Would you share the link to that video?
It is not surprising to get a better LF extension with a support of the exciter. The magnet weight acts as an inertia on the voice coil system. Adding a support is about the same as increasing the magnet weight (from BW point of view). The coil force is no more used to move the panel and the magnet but only the panel.
Christian
Hi Christian.
I think there is more going on in the low frequency region than you mentioned.
I think of an unsupported exciter as a low frequency resonator.
The performance would vary depending on panel material .
But let's take an example of a large EPS panel.
The panel is very light ,but the air pressure on the panel makes it very hard to move in the very low frequencies.
So a spine mounted exciter will struggle to move even a light panel.
But a free floating exciter bounces around like a weight on a spring at very low frequencies, this causes low frequency resonances similar to a helmholt resonator.
Similar to a table tennis bat with a ball attached with a rubber string.
The ball is bouncing back and forwards banging on the panel.
Extending the low frequencies.
I mentioned this many years ago.
What do you think ?
All the best .
Steve
I think there is more going on in the low frequency region than you mentioned.
I think of an unsupported exciter as a low frequency resonator.
The performance would vary depending on panel material .
But let's take an example of a large EPS panel.
The panel is very light ,but the air pressure on the panel makes it very hard to move in the very low frequencies.
So a spine mounted exciter will struggle to move even a light panel.
But a free floating exciter bounces around like a weight on a spring at very low frequencies, this causes low frequency resonances similar to a helmholt resonator.
Similar to a table tennis bat with a ball attached with a rubber string.
The ball is bouncing back and forwards banging on the panel.
Extending the low frequencies.
I mentioned this many years ago.
What do you think ?
All the best .
Steve
I'm looking forward to your sketch and following your build.Although the timber structure is very simple, it doesn't lend itself to easily written description.
I wish I was one that could draw/sketch in 3D. I will try to attempt to make a sketch and post it
( I will post stage by stage photo's as the project proceeds. I'm not a very 'quick build' person )
One thing I suggest is that you consider doing a little prototyping before committing to a particular design. Watching videos is a great place to get ideas, but trying out different panel materials, panel sizes, shapes, and mounting arrangements on your own is a much better teacher.
Concerning the exciter mounting, I think that for your intended design, the lack of a spine to support the exciter won't be any problem. Some exciters will sag over time, if unsupported. But the 32P has a steel spider which should be able to resist that tendency. Also it has a broad mounting flange so it will be less likely to fall off over time. The other thing that can happen with unsupported exciters is that they can bang against the panel at low frequencies. But if you filter out the LF (as you intend) this shouldn't be an issue either.
Concerning the panel size, I wonder if you might be able to get away with a much smaller panel. Panels that are too small won't reach low frequencies, but since you are using a woofer for the low end, a relatively small panel might be enough. IIRC, you are planning on 1200x1200mm. Without knowing how stiff and heavy your panel is, I can't really say if it's truly "large" or not. But it may very well be a lot larger than you need, while not providing any real benefit. A little prototyping would help you know for sure.
Good luck!
Eric
@deanznz and @jaxboy There have been some breakthroughs in Tinnitus treatments FYI
https://www.lenire.com/
A general information blog for sufferers.
https://www.tinnitustalk.com/
https://www.lenire.com/
A general information blog for sufferers.
https://www.tinnitustalk.com/
Hello Steve,
Your are right. Several resonances are at play and coupled one with the other. The problem is we have no tool or design guide lines to predict the result.
Christian
WOW, so many comments & questions - that's great >
Firstly, regarding prototypes, I don't actually have the budget for multiple timber purchases.
This project is for fun and I am confidently expecting good results - especially after the exciting XO stage.
I have the PC spectrum analyzer and I am more than prepared to dig-out my old Yamaha graphic equalizer.
Re. the 5 panel demo video, I made no notes of its title - but it is out-there somewhere on YouTube.
Yes, my panels are large, but I am using two exciters per panel for the following 3 reasons :
Re. magnet weight, we have the wonderful realm of Neodymium to thank for the light weight of the Dayton exciters.
As mentioned, the spider/suspension of the Dayton EX32EP2-4 is actually metal. (no sag problem)
After further Dayton spec. info. my crossover frequency won't be 200Hz , but in the 300Hz region.
If there is some question(s) I failed to address, please let me know.
PS.
I have a 'mental anxiety' regarding drawing - please be patient and I may come up with something.
Firstly, regarding prototypes, I don't actually have the budget for multiple timber purchases.
This project is for fun and I am confidently expecting good results - especially after the exciting XO stage.
I have the PC spectrum analyzer and I am more than prepared to dig-out my old Yamaha graphic equalizer.
Re. the 5 panel demo video, I made no notes of its title - but it is out-there somewhere on YouTube.
Yes, my panels are large, but I am using two exciters per panel for the following 3 reasons :
- 2 exciters help overcome the inertia of the panel - that is actually light weight.
- 2 exciters create a nice 8 ohm load, plus double the power handling capacity.
- Because the 2 exciters are in series, this will reduce the 'panel volume level' to better match the woofer's efficiency.
Re. magnet weight, we have the wonderful realm of Neodymium to thank for the light weight of the Dayton exciters.
As mentioned, the spider/suspension of the Dayton EX32EP2-4 is actually metal. (no sag problem)
After further Dayton spec. info. my crossover frequency won't be 200Hz , but in the 300Hz region.
If there is some question(s) I failed to address, please let me know.
PS.
I have a 'mental anxiety' regarding drawing - please be patient and I may come up with something.
Unfortunately Steve, you generate tinnitus independently of your hearing ability, so you'll always have it....☹️
Also lucky you didn't try my dome, as you wouldn't have been able to hear it anyway 😵💫
It's a coupled mass system Christian, and the support problem can be easily and successfully overcome by hangers as documented previously
Eucy
This was from me, jaxboy. I was really shocked at how the freq top end of my hearing seems to be so much lower, as it sounds to me to be the same as it was, but is 3kHz lower than it was last year. I used the same freq generator, too. I just retested it and my hearing tops at 4740Hz now. The tinnitus freq seems to be at about the same point relative to my absolute hearing. My low freq is 41 Hz, which is great, as I love bass. As to DMLs, this means that I can play my whole freq range on them.
A few more comments re. my large DML panels >
At the outset, I decided that if I was going to hear this 'wide-open-expansive' kind of sound,
I would go 'all the way' with large panels. The large size panels actually serve more than one function.
The large DML panels are just above the (open baffle) woofer height, and this will actually create
a 'phantom baffle' augmenting the low frequency output.
This also means that the DML panel will actually acoustically couple with sound pressure waves of the woofer.
The panel will vibrate at low frequencies, but in a passive (not active) way. (again the bottom of panel requires suspension)
This woofer/panel interaction will get 'sorted-out' in 2 or 3 ways.
( it's a bit like - 'low feedback tube amp fans' that are more than happy to have 5% THD for the sound they love )
At the outset, I decided that if I was going to hear this 'wide-open-expansive' kind of sound,
I would go 'all the way' with large panels. The large size panels actually serve more than one function.
The large DML panels are just above the (open baffle) woofer height, and this will actually create
a 'phantom baffle' augmenting the low frequency output.
This also means that the DML panel will actually acoustically couple with sound pressure waves of the woofer.
The panel will vibrate at low frequencies, but in a passive (not active) way. (again the bottom of panel requires suspension)
This woofer/panel interaction will get 'sorted-out' in 2 or 3 ways.
- The phase characteristics of the crossover circuit.
- The phase orientation of the drivers.
- If needed, active equalization.
( it's a bit like - 'low feedback tube amp fans' that are more than happy to have 5% THD for the sound they love )
Last edited:
To all the DIYers around,
I would like to recommend reading the paper from Charlie.
The paper coverage is wider than open baffle design with standard loudspeakers as it explains the benefit of an omnidirectional dipole in a room, how to deal with the power response of an omnidirectional system, some inputs for the implementation of EQ and crossover in a Linux system.
Thank you for the paper @CharlieLaub , thank you for pointing to it @deanznz
Sometimes I relate it to a better option/outcome that presents itself.This is an example of a strange phenomenon I call > Mark's Luck
Is there a specific distance from the wall a hanging DML needs to be? I have read some generic statements about being away from the wall, but it would seem that there would be a specific distance. If you go for a box/frame with the weatherstripping instead, do you lose the wide soundstage?
Sometimes I relate it to a better option/outcome that presents itself.
Hey there @deanznz >
It would seem you know something about serendipity.
The forced 'timber cutting pause' gave me extra time to contemplate the sheer size of my build plan.
I already have two gigantic speakers in my lounge room, and my new plans were just TOO BIG.
I can now re-scale the project and get improved transient response as a result.
The project will continue in financial time, and I'll post photos.
CHEERS to 'the wise one'
Hey there @deanznz >
It would seem you know something about serendipity.
The forced 'timber cutting pause' gave me extra time to contemplate the sheer size of my build plan.
I already have two gigantic speakers in my lounge room, and my new plans were just TOO BIG.
I can now re-scale the project and get improved transient response as a result.
The project will continue in financial time, and I'll post photos.
CHEERS to 'the wise one'
Hello
DML belongs to the family of the dipoles with a rear wave almost of level and FR than the front one.
I have read the same general statement (is possible 1m, more than 30cm...) but nothing about a specific distance.
I have just read about a dipole made with 2 BMR (one on each direction in opposite phase) the level of the rear one was reduced by 2.5dB for a better image.
We could think the distance is a matter of minimal attenuation or delay (or both) of the rear wave.
My understanding is the quality of the imaging comes first from the omnidirectional dispersion of the DML (see also for that BMR or low directivity open baffles).
I am not sure to fully understand "box/fram with a weatherstripping".
My panels as many others use a "weatherstripping" on a frame to offer a mechanical structure and a suspension to the membrane but the back remains fully open. In such design, it remains a dipole, the rear wave is free to expand.
the exchanges here about the frame/weatherstripping versus a free edge panel are more about the smoothness of the FR. Let see what will be the other answers to your questions.
There are papers suggesting to build a flat closed box as for a cone speaker in a closed box but I don't remember it was experimented here.
Christian
By "box/fram(e) with weatherstripping" I was referring to some projects I looked at - don't remember exactly where - that used the DML panel by mounting on or even recessed into a wall. There was some effort to dampen reflections behind the DML panel and on the frame. I wasn't sure whether or not that defeats the advantages of the DML in the first place.
Regarding placement of the panel - after some reading it seems that a panel would need to be placed 11 feet from anything that would reflect the sound, or attenuated -12db, or some combo of the 2 to have the ideal sound. However, some uneducated mathing seems to indicate that reflections from the floor would be mostly within 1-5ms give a 3-4 foot high sound source and the same listening height. So it would seem to me that in a room that is 12x15 or 15x20 with an 8' ceiling, you would have tons of reflections sub 10ms even if you were able to 100% contain the back wave.
If the back wave is not out of phase with the front wave (imprecise I know, as they would really be the same wave) then why would bass be cancelled more than treble?
https://www.audio-forums.com/articl...paciousness-and-the-initial-time-delay-gap.9/
Regarding placement of the panel - after some reading it seems that a panel would need to be placed 11 feet from anything that would reflect the sound, or attenuated -12db, or some combo of the 2 to have the ideal sound. However, some uneducated mathing seems to indicate that reflections from the floor would be mostly within 1-5ms give a 3-4 foot high sound source and the same listening height. So it would seem to me that in a room that is 12x15 or 15x20 with an 8' ceiling, you would have tons of reflections sub 10ms even if you were able to 100% contain the back wave.
If the back wave is not out of phase with the front wave (imprecise I know, as they would really be the same wave) then why would bass be cancelled more than treble?
https://www.audio-forums.com/articl...paciousness-and-the-initial-time-delay-gap.9/
ajh.
I posted some videos on YouTube of various size and types of panels.
To give an idea of what can be achieved with a dml panel.
These are just a few.
Use headphones only, for best sounding results.
If mounting near to a wall I would recommend the mounting in the rather rushed drawing.
As long as you do not go too close with the wall and panel in parallel.
Steve.
I posted some videos on YouTube of various size and types of panels.
To give an idea of what can be achieved with a dml panel.
These are just a few.
Use headphones only, for best sounding results.
If mounting near to a wall I would recommend the mounting in the rather rushed drawing.
As long as you do not go too close with the wall and panel in parallel.
Steve.
Attachments
Just a handy hint for DML builders >
If you are going to use "weatherstripping" for panel suspension, don't use 'fluffy foam type',
it will change, deteriorate, and rot. Butyl rubber products are the way to go.
In Australia I found an amazing range of products at Bunnings. (the U.S. would probably be Home Depo)
I would recommend not using 'overly tight' suspension.
If you are going to use "weatherstripping" for panel suspension, don't use 'fluffy foam type',
it will change, deteriorate, and rot. Butyl rubber products are the way to go.
In Australia I found an amazing range of products at Bunnings. (the U.S. would probably be Home Depo)
I would recommend not using 'overly tight' suspension.
OK, clear. Having a loudspeaker flush with the wall is recommended to avoid the diffraction effect by the edges of the box. In the case of DML, I don't have in mind studies about that. As Steve said, the worst situation is the rear wall parallel to the membrane. What was not mentioned up to now is the consequence might depend on the membrane material. An "heavy" material like wood might be less sensitive than a light one.
11 feet is not a "home situation" sure. Yes, reflections much before 5ms is what we have.
About 10 years ago, I built a pair of line array (16 8cm full ranges in a narrow and tall closed box). Before that I had 3 way classical loudspeakers. About 3 years ago, I started with DML. In the same rooms (I tested in 2 locations), the sound of the DML is more pleasant to me and this with no great care about their position. The sound stage of the line array is similar to headphones with a sound "in your head". With the DML, the sound stage is in front of you with a low sensitivity to your own position. The sound is not located in the loudspeakers as it happened with my 3 ways. I haven't tested other types of dipole but a figure of 8 dispersion seems a key point. An other characteristic of DML is by their dimensions that are generally larger than a standard loudspeaker, they are self-baffled (the membrane is the baffle) with dimensions such the dipole roll-off occurs in the hundreds of hertz. So no crossover in the kHz range, not too difficult to add a woofer.
In a dipole, the back wave is out of phase which leads to the low frequency roll-off. What is shown in the frequency responses of the article you linked is an in-phase delayed signal