A Study of DMLs as a Full Range Speaker

They will be close to the wall, hopefully enough to hide the panel. If not, I will maybe try some black spray on the canvas back. If I understand it correctly, the canvas is just suspension and damping when the panel is relatively large, so a bit of spray paint should not make much difference.
 
Pelanj.
You can have the panel up close to the wall as long as you do not have it flat on the wall, as this will roll off the frequencies below about 500hz and sound pretty bad.
You can have one side on a hinge on the wall for instance, and pull the other side out by at least 6 inches , or even straight out as burnt suggests, this will give you full range panels again , hopefully down to 40hz if not pushed too hard.
If on a hinge you could push them back when not in use.
As for spraying the back, you could just staple on some fabric or speaker cloth.
Just some thoughts.
Steve.
 
  • Like
Reactions: pelanj
+ @pway
Hello Paul, hello Eric
For the materials, what about the most standard :
  • EPS E 6MPa, density 18kg/m³
  • XPS E 21MPa, density 31kg/m³
  • poplar plywood E 9GPa, density 530kg/m³
For the thickness 15 to 20mm for EPS or PS, 2, 3 or 5mm for plywood
For the dimensions why not Leob's choice 500x330mm (25mm thick), for plywood Burntcoil's tall blonde 1160x300mm, Eric 406x584mm or 305x1200mm.
Eric, are those figures ok for you?
Paul, those information are gathered in the history file. Follow the link "github.com" under my signature on the left of the post.
For the FEM tool, Elmer might be interesting as it is multi-plateform (my computer is under Linux Manjaro...), LISA doesn't work in the windows emulation. Some tutorial needed for me!
Christian
Hi Christian
Not sure how much difference it makes, but I'd use 400 kg/m3 for Poplar ply
Eucy
 
Eucy,
The natural frequencies scale as 1/rho^0.5. So all the natural frequencies would increase by a factor of (530/400)^0.5 or about 1.15, if the density is changed from 530 to 400. I measured about 440 for my 5 ply poplar.
Eric
+ @Eucyblues99 , @pway
Thank you for the feedback. I think that to make a common data base of materials would be useful. If you have some data, lists, tables... We'll find a way to organize it. All the ideas welcome.
Christian
 

Multitone distortion measurement​

Some month ago, I posted measurements of multitone test of different materials for DML panels. I just found this thread Multitone distortion measurement on drivers?. Very interesting. One of the posts points to New multi-tone speaker measurement method at LowBeats. This page makes a proposal for a metric : the multitone signal is played, the SPL measured in dB at 1m. In parallel the spectrum is displayed (ie REW). The volume is increased up to produce a distortion that meet what the page defines as "distortion limits". The metric is the reached SPL. Here is the description of the limits :
After thorough listening tests, the following guidelines have proven to be practical. At frequencies below about 300 hertz, the distortion limit for the "recommended upper sound pressure level" is 3 percent (distortion reduction = 30 dB), above this the generally applied rule 1000 / frequency in hertz applies. The distortion limit at 1000 Hertz is therefore 1% (distortion reduction = 40 dB).
The "recommended upper sound pressure level" of a speaker is the level at which it reaches the distortion limit defined above (microphone distance = 1 meter) in any frequency band in the sound spectrum.
 
  • Like
Reactions: pway and BurntCoil
i made a recording of the whizzer cone running with my TLs , XO at 450hz
i turn the sub on and off so you can hear what the whizzers are doing.
the whizzer has only one blob of bluetack holding it on to the exciter , i preferred this mounting .
i managed to get it to go down to 450hz, in the end with a much better frequency response .
steve.
 

Attachments

Multitone distortion measurement​

Some month ago, I posted measurements of multitone test of different materials for DML panels. I just found this thread Multitone distortion measurement on drivers?. Very interesting. One of the posts points to New multi-tone speaker measurement method at LowBeats. This page makes a proposal for a metric : the multitone signal is played, the SPL measured in dB at 1m. In parallel the spectrum is displayed (ie REW). The volume is increased up to produce a distortion that meet what the page defines as "distortion limits". The metric is the reached SPL. Here is the description of the limits :
Will REW create a multitone signal like this, or would we need to build one with Audacity?
Edit: never mind, looks like there are options for evenly spaced log frequency signals, amongst others.
 
Eucy,
The natural frequencies scale as 1/rho^0.5. So all the natural frequencies would increase by a factor of (530/400)^0.5 or about 1.15, if the density is changed from 530 to 400. I measured about 440 for my 5 ply poplar.
Eric
Consulting an equation for eigenfrequencies of rectangular plate, it looks like:

f proportional to rho^-0.5 as you say
f directly proportional to h (thickness)
f proportional to E^0.5
 
Christian,
I found this article about using Elmer that is close the the way I’ll be using it, with gmsh and paraview.
I may use the gui just to get an initial sif file, then go to manual edit of sif and command line.
I will probably end up using pyelmer as it streamlines the process and allows multiple runs with parameterised variables for optimisation or experiment.

https://medium.com/innovating-resourcefully/getting-started-with-elmer-fem-finite-element-method-and-gmsh-using-python-55d3c158aa8

Eric,
Did you try different shapes with Lisa, and if so did you normalise the area? I’m thinking of normalising to say 0.5 m^2, which is close to a practical size.
 
Will REW create a multitone signal like this, or would we need to build one with Audacity?
Edit: never mind, looks like there are options for evenly spaced log frequency signals, amongst others.
Hello Paul
In the post 4656 the measurements I did. The link to the wave file, go to "test waveform". The file was posted here 4958.
This is the tests from Leob : 4953 with a multitone from REW. Go in the generator in REW and then multitone. I haven't tried it.
Christian
 
Christian,
I found this article about using Elmer that is close the the way I’ll be using it, with gmsh and paraview.
I may use the gui just to get an initial sif file, then go to manual edit of sif and command line.
I will probably end up using pyelmer as it streamlines the process and allows multiple runs with parameterised variables for optimisation or experiment.

https://medium.com/innovating-resourcefully/getting-started-with-elmer-fem-finite-element-method-and-gmsh-using-python-55d3c158aa8

Eric,
Did you try different shapes with Lisa, and if so did you normalise the area? I’m thinking of normalising to say 0.5 m^2, which is close to a practical size.
Thank you Paul
The list of tools sounds ok too me, using some command lines or subprocess call from Python too (I already have done that for more simple task like exporting a pdf like DMLmaterial doc from a python script). The challenge is in the new formats of file (.geo, .sif) but I think it is a good proposal. If you have simple example of the intermediate files and commands (even the in windows syntax), I would appreciate to start with.
Christian
 
Hello Paul
In the post 4656 the measurements I did. The link to the wave file, go to "test waveform". The file was posted here 4958.
This is the tests from Leob : 4953 with a multitone from REW. Go in the generator in REW and then multitone. I haven't tried it.
Christian
Wow thanks for such a comprehensive response! A lot of water under the bridge since I was last active on the forum. Looks like a useful technique.
 
If you have simple example of the intermediate files and commands (even the in windows syntax), I would appreciate to start with.
Christian
No problem, I can upload a the geo, sif file I used whenever you are ready. I've not done much though yet myself. Have you managed to install Elmer on Manjaro? In the first instance, just use the GUI and work through the tutorial.

One gotcha: I wish I had seen that medium article first, as it used the msh2 version option when running gmsh. Elmer grid does not know how to process the latest version & I wasted a lot of time discovering that myself 🙁

The challenge is in the new formats of file (.geo, .sif)
I have found gmsh a bit clunky to use from the gui, but doable for the simple 2D models I need. The geo files are just a simple language, plain text.
Sif files have a simple structure too, the complexity comes from researching the exact algorithm options you need.
Pyelmer is attempting to integrate the whole workflow, looks OK. I hope to check it out this week.
https://pypi.org/project/pyelmer/
 
Eric,
Did you try different shapes with Lisa, and if so did you normalise the area? I’m thinking of normalising to say 0.5 m^2, which is close to a practical size.
Paul,
I must admit that I have modelled a lot more rectangles than anything else!
I agree that about 0.5 m2 is a good practical size. But most often I start with the material parameters of the plate I have in hand, and see what size it would have to be to get the response I'm after.
Eric