Thanks Christian,
I am unfamiliar with both Rohacell and PolyEther. But from their material properties it looks like their density is too high for good efficiency.
I know you have posted those equations before, and I do need to apply them some time to see if I can make them work in real life. In your experience, do they line up with your tests?
Last edited:
I concur. Those pre-made panel skins do look a little too thick. And, same as you, I will have to try make my own CF skins.
At this stage it looks to me like EPS is still the least dense core material, but I'm not sure how it will react with resin/epoxy adhesives. Watch this space.
wharfedale and fane had dml´s quite some time ago. it was called loudpanel. they suggested to build a ceiling of panels just above the dance floorWhile there is a very long thread in full range regarding DML speakers, I think they are also very interesting for PA applications, and deserve their own thread since it is quite a different application then when making a pair of speakers for home use.
So here might be a good place to discuss things like what materials offer the best sensitivity, how to get maximum power density per plate, what power amplifiers are good to use with DML speakers and how do you construct a good frame and protect the exciters and plates from rain, etc.
Personally I started looking at DML speakers simply out of curiosity since they just seemed very novel in the the way they actually work as well as how they perform. Initially I hoped to get a pair of half decent home speakers at best, but quickly realised both how much I like the sound, but also how interesting the properties are for PA applications.
So why is DML interesting for PA sound?
1. SPL does not reduce as much over distance. This means a more even spread of the sound over the audience. The decay rate is more like a line array even if you have a single plate set up as a point source.
Basically if you stand on a dancefloor with plates directed to the middle from each corner, it feels like the stacks are half the distance away compare to where they actually are. It is hard to describe, and it is a weird feeling when you walk around between them.
2. The diffuse waves generated by DML really helps with acoustic issues. Not only by reducing the amount of reflections we perceive in a closed space, but also the interference between multiple speakers. It is also weird and seems to defy logic sometimes, but in combination with the previous point it means the whole dancefloor sounds like a sweet spot with both loud and clear sound.
3. Dispersion is really wide, covering a lot of area with a few speakers.
4. Good sensitivity, even when looking at the usual on axis response, and amazing total efficiency when taking the dispersion into account.
5. Low distortion even at high SPL.
6. Despite the diffuse nature of the waves emitted, spatial imaging is excellent. Even when standing very close to one speaker stack you can hear stereo effects from speakers on the other side of the dancefloor in a way that I have not experienced before.
I make a small festival each summer, and as it happens this year the guy who I usually rent from couldn't do it, so I decided to quickly try to put together a DML system since I had done enough experiments to be confident it would be viable.
I will add post in this thread describing the system I built in more detail and add some videos of it in action, but for now a quick description.
4 plates per stack, each plate made of :
1 graphite enhanced EPS plate, 500x330x25mm:
https://www.dekokopf.com/neopor-styroporplatten-3er-set-50x33x2-5cm.html
4 DAEX30HESF-4 exciters:
https://www.soundimports.eu/en/dayton-audio-daex30hesf-4.html
Frame built with wood and 3d printed corners.
I have 4 stacks, so a total of 16 plates. Each stack is driven by one of these:
https://www.china-sanway.com/D2S-2CH-900W-DSP-Built-in-Amplifier-Module-Class-D-pd353583.html
Subs was rented 4xL'Acoustics SB218.
I tried to make construction as quick and simple as possible, and think I succeeded with that, but still of course it become stressful to finish in time. I didn't have much time to tune it with the subs, and never got time to test it properly at war levels before the party, so I was nervous to say the least. The guy I rented the subs from brought tops as backup if it would seem like a flop.
We turned it on and without even tweaking the XO it just was way beyond expectation, sounding really loud and clear, and no problem to keep up with the subs, rather the other way around. Sound guy admitted he was very sceptical initially, and probably thought I was crazy not using his Kara tops instead, but admitted he was very impressed when he heard it
During the party I just got constant comments about the sound. Everyone positive, and lots of people that would never do that usually. The actual sound geeks present was over the top ecstaticCommon comments where "everywhere seems like a sweet spot", "I always have to wear earplugs, but not with this sound", "best sound I have ever heard" and "it is like listening to quality headphones but from a PA system". After a couple of days of hearing the same comments it almost went from total relief that I was not crazy and it was not a flop to it becoming a bit too much and annoying
Since it was such a resounding success I'm really surprised there hasn't been more interest in DML technology for PA applications.
One tricky aspect with DML is the patent situation. Since I do this as a private person and not for profit (my festival is also non commercial) it is not an issue, but getting started with DML commercially might not be so easy. The only commercial manufacturer of DML PA speakers are Tectonic. They seem to produce a very high quality product, but at a price level only big established rental firms can afford. However, those tend to be quite conservative and go with the products that are typically requested, and have no interest in having to sell in customers on something completely new. And Tectonic also seem to have very limited marketing and no presence outside US, and importing their product to the EU would end up very costly.
In the end, with nothing else on the market, no one hears DML systems and no one demands them from the rental firms.
So it seems to be up to DIY'ers has to build some systems and get the word around!
Anyone else doing DML for PA applications, or is interested in doing that?
Hello. Yes it fits quite well. Not the time right now but I will try to find later what I did on that. If not by the advanced research with keyword "efficiency" and posted by "homeswinghome"
@Andre Bellwood
Hello André,
I made some archeology thanks to the advanced search function. See :
- #5449 which introduce the D/µ³ (link to papers)
- #4664 with this very nice graph from my test (more posts in this one : #4201, 4203, 4256)
The graph is SPL@2m for different panels I tested according to log10(E/rho³).
Single point on the left : acrylic
3 mid points : different plywood
4 on the right : EPS/XPS
By the way, this graph is not listed in the history file... I am going to correct that.
Christian
Thanks André
Your question gives me the opportunity to test some tools I am learning to produce documentation. Nothing really new : pandoc with a source in markdown to pdf with a bibliography... Thinks that didn't exist when I was a student (it was the time of the ZX81...).
Please find attached the same results in a pdf file. Probably a bit academic in the format but I don't know how to change that (standard Pandoc format)!
Let me know your feedback, if some corrections needed. I will push it after in the Github page.
Christian
I use the easycomposites laminating epoxy to glue the exciters in place, and seems to be no reaction with the EPS.
As I understand you want a really light core as but with excellent compressive strength in the direction exciter is pushing to get good HF response.
If you use high grade EPS, compressive strength is good, but it is probably unnecessarily dense as core between carbon skins. Perhaps if cutting very high grade EPS in a honeycomb pattern it could work. And if you instead use low grade EPS to get similar weight to a honeycomb material you will get low compressive strength, and likely an inefficient and dull panel.
Haven't actually seen those wharfedale before. Seems to be a NXT licensed product, and not any good for PA use. Only goes down to 160Hz and a surprisingly poor 87dB @1m sensitivity with 70w "program" power handling.
I think this is why the polycarb Twinwall disc works so well, being a better "conductor" for HF than what EPS is. I'm waiting on tenterhooks for materials to arrive so I can start testing various honeycombs with Aramid/Kevlar/CF skins.
Last edited:
Hello André
Generally speaking units are SI (N, kg, s, m...). I will check and add (some other details to add on my side for a cleaner doc)
Has anyone tried a 2-way / 3-way system using different materials to cover HF/MF/LF? I know this is essentially what's being done with the disc, but what about having HF covered by it's own exciter/panel? Like twinwall "tweeter" with MF covered by acrylic or EPS, supplemented with a sub for LF.
I'm also after the goal of best SPL:weight ratio for mobile PA setups, getting fullrange out of a single panel isn't as important to me which is the goal of the other thread.
I do like the fact the disc still has HF radiating from within the main panel though, maybe panels could also be hung in a MTM configuration?
Last edited:
Hello Vajiao,
Leob's panels are as far as I remember the first PA application (memory!), for most of the applications where trying to have a wide range panel is a target (no crossover). So I think few examples of multi ways. Have a look around "Bertagni" : a single EPS membrane with 2 drivers. Basically if you have several drivers with different materials, you'll get different levels (efficiencies will be different). And unfortunately, it seems that material good for HF have a lower efficiency.
At this step, it seems too early for me to conclude that André's panel is comparable to a 2 way system... I think more investigations are needed to conclude. One would be to test one piece of the EPS material alone and one of the same surface from the twinwall material. Same efficiency?
Indeed. I'm using a single driver but which seems to be working in different bands because of different mechanical impedances presented to it. The fascinating thing about it is that there seems to be no loss of efficiency between the different systems at play on the panel.
I have tried a quasi-"three-way" panel running from a single driver as per attached pic, but that needs a helluva lot more work. I didn't spend enough time on it.
Yes that is a CAD sketch of the panel that I eventually made out of EPS. But the FR was so bad I did not bother going further.
I tried something similar with Correx. Same problem: The FR is atrocious. Besides, Correx is way too heavy to be efficient.
The principle behind the offset ovals is that, in a baffle with a pistonic tweeter, the edge diffraction is almost completely neutralized if you put the tweeter in the correct position.
But it absolutely does not seem to work on DML panels.
Sorry do not misunderstand, I would simply like to understand one aspect, being a forum that deals with applications for professional use, would it not be more obvious to start from the one that made tectonic with carbon instead of starting from scratch with other types of materials? the latter have always been for domestic use and keep the cost low, but if the priority is quality, why experiment with materials that surely tectonic will have tested extensively and then use carbon? I repeat it is just my curiosity, experimenting is always positive