Good to know.
Yes. 0.5mm skins and 2mm foam.
Based on my experiments so far I would go with even thinner skins if I could find them at this kind of price.
I found a surplus sheet of aramid honeycomb panel on Ebay large enough to cut into 2 speakers. I can't wait to experiment with it. The seller didn't know the skin thickness but it looks really thin in the pictures.
Sorry, I meant to link to their 50w version:
https://www.alibaba.com/product-det...l?spm=a2700.shop_plgr.41413.40.2bbb263b9hJmMP
But I noticed now that it says 32mm voice coil as well, and that it now on the product page for the one I bought doesn't mention voice coil diameter. Quite sure it used to say 50mm.
Indeed the FR looks better on the 30HESF. Also thinking about in what configuration I can use them. It seems like many that are happy with multiple exciters, including Tectonic, use 4 exciters and rarely 2. I recall them going into that in one of their videos, how 3 or more exciters can work together when placed close around the same centre. Will I be able to put 4x 30HESF per plate, or will I overload it?
Interesting DML speaker from billionsound:
https://billionsound.en.alibaba.com...?spm=a2700.shop_index.111713.4.913b22d70pY4dk
Claims 95dB sensitivity, which is impressive if true.
Looks like it is using aluminium cardboard sandwich plate. I was recommended that from the manufacturer of my exciter as well, apparently that is what they test with.
https://billionsound.en.alibaba.com...?spm=a2700.shop_index.111713.4.913b22d70pY4dk
Claims 95dB sensitivity, which is impressive if true.
Looks like it is using aluminium cardboard sandwich plate. I was recommended that from the manufacturer of my exciter as well, apparently that is what they test with.
The response curve looks similar to what I measured with that exciter.
I think you are right about thinner CF skins. When I make my next CF/balsa panels I'll probably use this, which finishes out to 4 mils (0.12 mm), which is about half the thickness of the CF skins on my recent panels.
https://compositeenvisions.com/prod...n-weave-12k-50-127cm-2-65oz-90gsm-spread-tow/
According to my estimates, the thinner carbon should provide greater efficiency, and allow the same frequency range from a smaller panel, when compared to one with thicker skins.
I'll be interested to hear about your aramid honeycomb panels when they arrive.
Eric
Hello DML'ers,
I moved my office into a room where I can only have speakers on the ceiling. After some research I decided to give DML speakers a try. I successfully build a set of exotic Cornu Spiral speakers before and have some experience with tools like AJHorn.
Unfortunately, I made a mistake and ended up with a dilemma that I hope you can help me solve.
I already built one 10 inch flat sub with its own plate amp for the ceiling and I am merely done with the panel speakers.
They consist of many compromises and took quite some time to source.
sheets : 3.2x4 ft 5mm RevolutionPly
framing: 3.2x4 ft MDF/FIR sandwich frame
suspension: 1x1 inch soft/spongy foam that is sandwiched between frame and panel with the help of spray glue.
Now, I thought that I am smart and I put guitar bracing similar to the Yamaha FGX guitar tops onto the panels' backside (which I already partially tuned). However the weight of these panels is now not just a few grams anymore and I have doubt that one exciter will be able to produce enough volume.
Luckily, I planned for a multi panel system and I got 6 exciters.
2 x DAEX25SHF-4 4 Ohm 20 Watts
2 x DAEX30HESF-4 4 Ohm 40 Watts
2 x DAEX32EP-4 4 Ohm 40Watts
So, then I started to carve/remove plywood at the spots where I want the exciters to sit and filled part of these voids with polyurethane. These voids have a cone shape with the tip close to the surface layer of the actual front side of the panel.
And now the dilemma and the question.
All is prepared for final exciter placement and wiring and I just NOW notice that I can only wire ...
... 3 exciters in series adding 12 Ohm load to my tube amp.
... 2 exciters in series and 1 exciter with a 4 Ohm resistor added, which will load the amp with 4 Ohm.
I have no idea how the 20Watts / 40Watts mismatch will affect things. And, I also wonder if I could compensate with a 8Ohm instead of a 4Ohm resistor.
Like
---- 4 ---- 4 ---
---| | ---
--- 4 ---- 8 ---
Or, shall I forget about using 3 exciters and either get a 4th one (but which?) or just go with 2 exciters (which would be a bummer in respect to the polyurethane spot that I already made)?
Thank you for any ideas, experience and "do not do that"'s
Sebastian
I moved my office into a room where I can only have speakers on the ceiling. After some research I decided to give DML speakers a try. I successfully build a set of exotic Cornu Spiral speakers before and have some experience with tools like AJHorn.
Unfortunately, I made a mistake and ended up with a dilemma that I hope you can help me solve.
I already built one 10 inch flat sub with its own plate amp for the ceiling and I am merely done with the panel speakers.
They consist of many compromises and took quite some time to source.
sheets : 3.2x4 ft 5mm RevolutionPly
framing: 3.2x4 ft MDF/FIR sandwich frame
suspension: 1x1 inch soft/spongy foam that is sandwiched between frame and panel with the help of spray glue.
Now, I thought that I am smart and I put guitar bracing similar to the Yamaha FGX guitar tops onto the panels' backside (which I already partially tuned). However the weight of these panels is now not just a few grams anymore and I have doubt that one exciter will be able to produce enough volume.
Luckily, I planned for a multi panel system and I got 6 exciters.
2 x DAEX25SHF-4 4 Ohm 20 Watts
2 x DAEX30HESF-4 4 Ohm 40 Watts
2 x DAEX32EP-4 4 Ohm 40Watts
So, then I started to carve/remove plywood at the spots where I want the exciters to sit and filled part of these voids with polyurethane. These voids have a cone shape with the tip close to the surface layer of the actual front side of the panel.
And now the dilemma and the question.
All is prepared for final exciter placement and wiring and I just NOW notice that I can only wire ...
... 3 exciters in series adding 12 Ohm load to my tube amp.
... 2 exciters in series and 1 exciter with a 4 Ohm resistor added, which will load the amp with 4 Ohm.
I have no idea how the 20Watts / 40Watts mismatch will affect things. And, I also wonder if I could compensate with a 8Ohm instead of a 4Ohm resistor.
Like
---- 4 ---- 4 ---
---| | ---
--- 4 ---- 8 ---
Or, shall I forget about using 3 exciters and either get a 4th one (but which?) or just go with 2 exciters (which would be a bummer in respect to the polyurethane spot that I already made)?
Thank you for any ideas, experience and "do not do that"'s
Sebastian
Thanks for you patience in explaining! I got the impression that it shouldn't matter so much which exciter you use, but seems like the inductance indeed vary greatly.
The Dayton 25FHE might have a dip around 12k according to Daytons graph, but it does have less difference between peak and 20k, which matches the lower inductance. Since you don't have the same dip, I don't how much one should read into the details of the FR since there are so many factors involved...maybe the impedance graph actually tells more about what response to expect?
No problem. Try explaining is also learning.
Here is how I see this DML system:
- the exciter is an electrodynamic motor based on Laplace force : the current thanks to the magnetic field gives a force to the membrane. For a DML we expect a magnitude of the force constant with the frequency. As the force is directly linked to the current, any reduction in the current or any other use of the force will reduce what is apply to the panel. The voice coil inductance is (always from my understanding at the moment) an element that reduced the current. I suspected the mass of the voice coil to be also a limitation factor but I have no proof of that. So the exciter comportment should something quite smooth versus frequency
- the panel makes the other part of the story based on its modal behavior... and here the effect from one material to an other are so strong that I don't think we can easily use the information of an FR (but our own). Perhaps if one day, some standard panel will be defined...
Let me try my question about what you all have learned about making great dmls a different way. Of all the dmls you have actually built what are the characteristics of the one you listen to most often? By characteristics I'm meaning panel material and size, panel treatments, exciters, exciter placement, etc., etc. I don't mean to be a nuisance with this question but I strongly suspect that there are a lot of people that would be very interested in the responses. If your best dml so far is in the history spreadsheet you might mention that. What do you particulary like about your best on so far?
I agree with Christian. I don't think the Dayton FR graphs tell you anything meaningful. I suspect they are included because people expect to see it, but not because it's actually indicative of anything meaningful to you and I. The impedance curve may be meaningful to some, however.
Eric
Hello Getgoin
I will see in which way I will be able to answer more in detail to your questions... I understand you question which the one I had in last December when I started joining this thread. It is really a big difficulty from this thread to have the description of what is in use, what are the best designs, the errors... I try to record in the file the answer that are kindly given.
So, short summary of my experience :
- phase 1 : documentation. Reading the threads (fully!) about DML, academic papers, some patents...
- phase 2 : building (see after)
- phase 3 : joining this thread, making "basic tests" to try to have a better understanding and to define what could be the next designs. We'll see if I can come later on that.
- XPS 20mm (home insulation material) 60x125cm, PVA coated + white acrylic on the front face, exciter just glue on the panel off center. The tickness was reduced at the exciter axis to about 10mm with a dome shape to reduce the HF. Good enough to be encouraged to continue but not good enough for a daily use without correction like EQ
- XPS 9mm (depron?) 60x80cm PVA coated, exciter at the center : total failure. No bass, no treble. I am still trying to understand why...
- Poplar plywood 3mm 45x120cm on a frame with a foam suspension and a spine. See the history file for link to posts. Excellent! They are currently playing... even if it is with the TV set. They give us a sound I have never got from other speakers. The sound stage, the atmosphere, the natural. The speed of the transients is also incredible. There is really a magic in the DML sound. No need of EQ even if I think some would help. The bass is of course weak but it is totally usable in a full range in a first step (first step that might last, I have no subwoofer at hand to go with). The sound is so that you forget the speaker not knowing where the sound comes. They are just like 2 pieces of furniture (door with no handle?). They are just laying on my previous 3 ways column. So probably not in the best conditions. but any way I can't offer to my loudspeakers more than 30cm from a wall. I think the design of the spine has to be improved by removing the 10x10cm square piece I have used to maintain the exciter (it is in fact a kind of horn on the rear wave).
- Canvas panel 41x33cm. Extraordinary week-end DIY. Easy to do. Very good result. Almost full range. Not so good as the poplar plywood models but they seat in the living room even if they are not plugged for now. I understand afterward that I was a bit aside the original design which is based on a light plywood pad where I used real balsa so my pad is flexible. I can't say what is the consequence.
Plywood and canvas offer also the possibility of an acceptable looking, style... this is at least true for me at home. If I can test panels with the salmon color of XPS, some better finishing is expected for a long time.
A plywood panel with a frame or the canvas have also the advantage of a minimum mechanical resistance to be in a living room. So as you can see, my criteria go further than the pure sound reproduction.
All of 4 run with Dayton Audio exciters DAEX25FHE-4.
Again, this is my current experience and opinion. I don't know people around me involved in such design so I could compare.
From their principles, DML are a completely different paradigm than the classical pistonic loudspeaker.
Christian
Hello Steve
Yes it works. Just unzip it and play. I already embedded some files here in zip. No technical problem.
Christian
No idea from me... For a main pistonic loudspeaker, adding a serie resistor change the bass response... For a DML, I don't know. Strange idea as it is a kind of waste of energy.
Why not 2 or 4 exciters?
Some drawings or pictures might help to understand your design (carve/remove the plywood? bracing? polyurethane?) Not standard!
Hello homeswinghome,
thank you for your reply, you just got ahead of me. I do not have access to the panels this week, but this is from a previous stage. I hope the converted file works here in the forum.
The idea of the bracing is to ...
... prevent them from sagging when they hang horizontal on the slightly sloped ceiling.
... the ability to tune them if particular frequency ranges are totally off.
Right now these strips are significantly thinner.
You can see the three tear drop shaped spots that I "carved" out of the 2 1/2 of the 3 layers in a conical profile. The idea of such insert was to have a harder surface that the exciter is attached to, hoping for more output in the higher registers like some acrylic panels' SPLs show. The idea of the tear drop shape was to create not just one resonance/peak frequency above 10kHz. This pic is in the middle of the filling process, which took quite some time and many layers to get this all filled up.
One can say, I just took all tweaks that I could find and applied them to these panels.... I have a spine in the working, but that won't be finished before the exciters sit.
Now, we not 2 or 4?
Well, I am not sure if 2 are enough for a 3.2x4 ft panel like this. And, 4? Well, which one might make sense? I am a little worried that I'll end up with way too powerful mids ...
I read in this thread that people try to avoid to re-attach exciters because the 3M's lifespan is limited this way. Therefore I only held the 3 exciter and that is not really representative for the final outcome. While the sound was great and changed a lot depending on the exciter positioning, I still fear that the output will not be suitable for one of the reggae or screaming rock days.
As a comparison my Cornu spiral speakers have on 4" 75Watt full range drivers each and the manage to shake the house.
Cheers!
Sebastian
Attachments
Last edited:
I see the ideas. Sorry I have more questions than answers to you... what you have chosen is away from I have tested or even read. Where to start?
About the bracing, against sagging risk why not. From the DML point of view it will increase the stiffness and so the first resonance which is not a goal in DML with plywood in my understanding. The intriguing point is how do you think to proceed to adjust the frequency? I come to DML from loudspeakers and electronics not from music instrument making.
More intriguing is your choice to create an interface with polyurethane. I have currently poplar plywood panel playing with the exciter glued on the back side with no lack of treble. Neither excess. It is even a quality for me of the plywood to have good natural balance in the treble; Other materials seem having a lack or an excess. Which background to this technique with polyurethane? Have you already tested it?
About acrylic, I am sure not to go to it because of its too poor efficiency and its HF extension is not clear. Some mentioned a lack of treble, other an excess...
About the multiexciter design, there is no consensus in this thread about it. The main stream thinking is to say it will create a multisource network creating interference and comb filtering. So basically the risk of limit in high frequency. There are voices to say not because the combination of vibration in the panel is more complex. I stay away of this problem : I have enough power, even with plywood which is a low efficiency material, for my use. Any way i more in the opinion of a concentrated source in HF due to the short wave length.
About the way to combine electrically the exciters, I won't worry to much. It is easy to change... if you have access. Some tests before implanting them ?
A good point of the DML is the sound field decreases more slowly than for a pistonic loudspeaker. So for a similar power you have probably a gain of 3dB.
I don't know which power you use in loudspeakers but if i remember the thread about "how much power?" in this forum, 25W is enough for 75% of the uses. In addition most of the DML are not used below something like 200Hz.
To re-attach the exciter is a problem with no good solution for me... at least with double side tapes from my local hardware store. One pair I have was glued with epoxy for its second use... One method from Spedge I haven't tested is PVA glue.
As having HF was not a problem in what I tested, I currently use a simple double side tape to make tests in the low to mid even upper mids with Spedge technique for hanging the exciter by the wires. When I will go for a complete new design, iI will decide between a new pair or probably epoxy.
As there are no strong basis for now, DML is a domain that needs many tests.
Christian
Yes, I'm leaning towards 4x DAEX25FHE-4 per panel now. That would be 100w per panel with 4 ohm in parallel-series, which I have the feeling might be slightly more than the panel can handle comfortably. But I do want to try to push it to its limits, and is about the same price as 2x DAEX30HESF-4 for 80w.
I guess I should order 2x 30HESF and 4x 25FHE and make two panels to compare.