A Study of DMLs as a Full Range Speaker

Jump to Latest
H

homeswinghome

Member
Joined 2012
github.com/Homeswinghome/AudioDIY_DML
Audiofrenzy said:
My advice, change the size and shape of the DML panel to a 8inch square.
Click to expand...
Thank you for the feedback
For now, the rounded shape shown no advantage compare to a square angle one.
About your proposal, why :
  • 8 inches? why not going to the 15" or a bit more which is the width of the woofer below?
  • square? a square is said to cumulate the mode in each direction at the same frequency which is not good for the smoothness of the FR.
 
  • Like
Reactions: 1 user
A

Andre Bellwood

Member
Joined 2022
homeswinghome said:
Hello
This is the DAEX25 shared some posts ago
View attachment 1197278
This is the DAEX25SHF-4
View attachment 1197279
Can we conclude the DAEX25SHF-4 has a better efficiency (higher BL even if 20db difference in the charts is strange), a better Xmax (not specified) but a lower HF capability? The difference in the mids are strange, this area being more driven by the material and the geometry than the exciter. In the LF, the roll off comes most probably from the panel material and dimensions.
Click to expand...
Agreed, Christian.
That 20db better sensitivity is very strange indeed. But I am tempted to get one to confirm for myself.
 
S

Sandasnickaren

Member
Joined 2015
Audiofrenzy said:
Why would you make the DML panel the same size or larger then the woofer below? Assuming the woofer is being used for the LF while the DML panel is used for the Mid/mid-bass frequencies and the tweeter is being used for the high frequency.

Like I've mentioned before every size driver has an ideal frequency response. 8inch is ideal for mid/midbass frequencies.
Physics dictate that the larger the size of the diaphragm the more ideal it gets for the LF while the smaller the diaphragm becomes its better suited for mid and high frequencies This is why I keep on saying that DML's arent that much different then conventional cone drivers when it comes to certain laws of physics.

A square shape is the most ideal shape for the most accurate/balanced sound. Since all sides are equal it will have the most uniform pattern as all the vibrational waves will reach the edges at the same time. In a long rectangular shape the vibration waves will reach the shortest edge before they reach the longest edge making the pattern less uniform. This is the reason why Dayton used a 12"X12" square for there general exciter frequency plots as its the most accurate shape.

This is the reason why mostly all conventional cone drivers are round because its evenly proportioned in turn will be the most accurate and balanced sound. With an oval shaped driver it will no longer be as accurate and or as balanced as a circle shaped driver due to the uneven shape. This is the reason why you wont find a oval shaped driver in higher end loudspeakers. Oval shaped drivers are used in cars, tv's due to limited space sacrificing sound quality.

Any shape that does not have all equal edges will tend to exaggerate the sound in certain frequencies and the greater the differences in the length of the edges the more exaggerated the sound becomes.

The tweeter you are using is around 1inch right? When asked why are tweeters small the most common answer is beaming which is true but its not the only reason and or not really the main reason. The reason tweeters are small is because it needs to focus the high frequencies and to do this the diaphragm needs to be small. When trying to recreate high frequencies on a large diaphragm the highs are dispersed through out the whole diaphragm making it sound more diffused. This is why some people say that DML's lack high frequencies as the treble is not prominent sounding. This also happens in larger full range conventional cone drivers. To solve this problem some speaker manufacturers use "WHIZZER" cones to focus the highs as the highs start from the middle of the voice coil and it spreads outwards to the edges of the cone. The larger the surface area the further the highs have to travel to the outer edges making it sound very diffused. Another way speaker driver manufacturers solve this problem is by using a very stiff dust cap dome shaped material. If you look at xsuper9988 (post 10,755) cone tweeter most dont realize that the small hard dome shaped dust cap in the middle is what focuses the high frequencies so that it sounds more prominent.

Both Spedge and Eucyblues use this dust cap dome shape. Another person I forget his name also used a whizzer cone.

The ear shape you are using is intended to be used more as a woofer. You already have a conventional cone woofer so the DML shape should be used as a mid/midbass driver which should be around 8 inches or smaller square or you could even use a short rectangle say 4 inches more then the width so a 8"X12".

Again physics dictate that the right size diaphragm has a ideal frequency response including the shape.
Click to expand...
Maybe a square shape is not the ideal format.
At least not according to this study. It’s from the Chalmers thesis.
A critical review of bending wave loudspeaker technology and implementation
Master’s Thesis in the Master’s programme in Sound and Vibration KUONAN LI

Thomas

IMG_3169.png
 
  • Thank You
Reactions: 1 user
V

Veleric

Member
Joined 2016
Concerning this quote, from the Kuonon Li thesis:

1690997147602.png


Does anyone know where the 0.9482 ratio that he refers to actually comes from? The NXT patents I know of mention 1.134 and 1.37 ( their reciprocals, 0.882 and 0.73), but I don't recall ever seeing 0.9482 anywhere other than the Li thesis.

Also, I recall other references to NXT's "gold" and "silver" ratios (which I presumed to refer to the 1.134 and 1.37, but perhaps not). Is anybody aware of any NXT/Azima, etc published literature that talks about the gold and/or silver ratios? I'm wondering where that terminology originated. Was it NXT papers or patents? Or someone else's papers? Or made up on one of the forums?

Eric
 
A

AjohnL

Member
Joined 2023
Veleric said:
Does anyone know where the 0.9482 ratio that he refers to actually comes from?
Click to expand...
I have seen from NXT several times in different places. One thesis looked around at what was available commercially. There was variation. From memory 0.8 and up.

The ones I am listening to uses 2 exciter positioned diagonally near the centre. Specs

PPS 1 Panel LP1 Sub-Bass
Nominal Impedance 8 Ohms 8 Ohms
Power Handling 70 Watts 70 Watts
(Programme)

Sensitivity (1W@1M) 87dB 87dB

Freq. Response (-6dB) 80Hz - 20KHz 40Hz - 170Hz
(DML connection)
System Crossover Freq. 160Hz
System LF Response (Fb) 35Hz
Drive Units 2x25mm Exciter 170mm Bass

H x W x D mm 555 x 500 x 27 235 x 460 x 210
H x W x D inches 21.8 x 19.7 x 1.1 9.25 x 18.1 x 8.26

The claimed DML FR is directly connected , not via the crossover which is in the woofer. The panels are edge supported.

;) Posting has removed the spaces in the formating. They are still there when I edit it.
 
E

Eucyblues99

Member
Joined 2021
homeswinghome said:
Hello
This is the DAEX25 shared some posts ago
View attachment 1197278
This is the DAEX25SHF-4
View attachment 1197279
Can we conclude the DAEX25SHF-4 has a better efficiency (higher BL even if 20db difference in the charts is strange), a better Xmax (not specified) but a lower HF capability? The difference in the mids are strange, this area being more driven by the material and the geometry than the exciter. In the LF, the roll off comes most probably from the panel material and dimensions.
Click to expand...
You do need to try one of these, they're nothing like the run of the mill exciters, they're a V8 compared to the 4 cylinder ones 😊
Eucy
 
  • Like
Reactions: 1 users
A

AjohnL

Member
Joined 2023
Veleric said:
The NXT patents I know of mention 1.134 and 1.37
Click to expand...
I think ratio must relate to tuning but not that sure NXT info is that relevant. They clearly did a lot of research - you can get an idea from this
https://simeoncanada.com/wp-content/uploads/2018/12/NXT-Technology-Review.pdf
Notice cad, 200 material database and access to a number of papers. Then what this physically came up with. ;) At a cost of course, that's business. They will also have done work on exciters. Not just them as well on panels. Patents tied down what companies could do. That's the benefit of having them.

The details of a set up I posted are an example. The ratio is amazingly close to 0.9 but hang on the panel is a composite of materials with different characteristics. It uses 2 exciters on a central diagonal. That factor and their spacing is probably significant. The whole thing gets more complicated than typical work mentioned in this thread. These panels also have a couple of pieces of light wadding in them. They also use NXT exciters.

Worse still the exciters internally are connected by a rather small electronic component. Andre suggested this is to limit power. Probably correct. Resistance increases as power levels go up. I've used winisd to design a woofer with an extended bass shelf ~1.5db down. I played with adding series resistance. At 1ohm the shelf levelled off. I've no idea how accurate this software is but in some respects a change makes sense. I wonder what happens with an exciter? I made the change in case the spl turned out to be too high to match the panels. Something else to wonder about.

LOL No measurements yet. Frankly I prefer to listen for decent time first.That in itself has been interesting. I was using Yamaha floor standing previously.

I also have some commercial licensed PC DML speakers. Ratio is ~1.5 but the edges have a slight radius rather than being straight maybe just for looks but would increase production costs especially on the frame but looks can figure in that. They are about 185mm tall, shortest dimension, the rad increases size in the corners. Implementation is poor but sound, well not too bad but the woofer amp and controls has problems. Once again like others that I have found info on the panel is edge supported. Glued in place in this case. Just pressed on in the others.

Given my background I wont disregard previous work that resulted in an actual product so in my case logically panels will be edge supported. They mostly seem to be on here anyway. Probably wasting my time if I look at anything else. To simplify that best bet appears to be pick a width and vary the length. Easy to change the frame aspect ratio then. Size ;) well large PC size minimises materials. A4 has an aspect ratio of 1.414. Somewhere to start. I have only seen panels that use central exciter arrangement. Bit of a ??????
 
  • Like
Reactions: 1 user
V

Veleric

Member
Joined 2016
AjohnL said:
I think ratio must relate to tuning but not that sure NXT info is that relevant.
Click to expand...
I'm not sure exactly what you mean by that. Perhaps you mean that you can get good results from many other aspect ratios, and I would agree with you about that for sure.
But what I'm looking for is evidence of whether or not NXT ever actually recommended the aspect ratio 0.9482 (regardless of whether or not it is actually a good recommendation). I've read a lot of the NXT patents and publications, and thought that I was aware of their recommended aspect ratios, and understood their logic behind those recommendations. But the 0.9482 (or 0.95) is not one I remember from any NXT publication, and I'm frankly somewhat doubtful that they ever actually recommended it (despite the assertion in the Li thesis). I was just hoping that if I'm wrong about that, someone could show me proof of that!
Eric
 
A

AjohnL

Member
Joined 2023
Audiofrenzy said:
Another theory of mine is that since the front and the back waves are 180 degrees they interfere with each other and cancel each other out. With a frame and or damping it reduces this cancelling effect and the deeper the depth of the frame the lesser the effect.
Click to expand...
That relates to the ones I am listening to. Meant to be hung on the wall. The shallow "rear enclosure" is filled with a foam under light pressure, little more than just contacting the panel. The foam is just cut away where the exciters are - oval as they use 2. The foam is bound to damp the panel.

180 degrees? I came across a paper suggesting test the polar response to see if the panel is optimised.
 
X

xsuper9988

Member
Joined 2023
X

xsuper9988

Member
Joined 2023
X

xsuper9988

Member
Joined 2023
Audiofrenzy said:
Be careful of low efficiency exciters. I bought some exciters with really low efficiency just to try it out and it was unusable due to very low volume output. I cant return them because I already wired/soldered them in series. :cry:
Click to expand...
I was expecting to pick up a few more to see if there would be a better result.
Because its bandwidth looks really good
What a shame...disappointed...

Thank you very much for your sharing and notification~
 
You must log in or register to reply here.
Top Bottom