Christian.
The difference is huge !
Gobel is trying to stop all reflections from the edges and prevent DML modes, this is a bending wave panel !
The methods he is using has been around for hundreds of years.
on beaches you can see water breakers or should I say wave breakers which are made of wood mostly and go out into the see ,to break up the waves, for the bathers.
The slits on the gobel panel do the same but with extra damping applied.
Steve.
Good question to know if there is something in practice behind that... The paper might be much in advance to me! I will read at least the introduction.Yea, maybe, that's why I qualified it with 'presumably'
I did a simple calc last year with simple textbook equations and got around 1000 Hz I think for 3cm, which is in the ballpark, and I left it there.
To get an accurate value, you would need to use the correct higher-order theory, not the simple Kirchoff-Love theory.
Foam for example is much thicker than metal plates and deforms much more, so the poisson ratio, shear strength and deformation effects would be important. It would be nice to have it line up, but will it lead to anything you can use, practically?
I just found this paper may be of interest if you want to pursue it: https://matelys.com/publications/AMCTCG20.pdf
Christian
Hello Steve,
Got it... I thought at the first look it was masking tape but not, it is sponges. Thank you I see now.
Christian
Steve,
Academic papers say : infinite plate no sound... so finite plate with infinite edge absorption, does it make sound? A DML with a frequency dependent edge absorption makes more sens to me for now for smoothing the FR.
I understand and I see your example at sea shore but which example of real bending wave panel (in the meaning of full absorption) do we have (apart the Göbel for which I don't know paper showing this full absorption)?
Far from my little knowledge.
Christian
Here are some responses from the 4-ellipse panel I showed yesterday. They are all in-room with 1/12 smoothing. I'm sure I did one outside, but have not found it. All curves are comparable in the sense that the volume was set constant and the speaker was places in the same position in the room. The speakers were set diagonally in the room, facing toward an large (~2m x 2m) doorway into the hall, in an effort to reduce room modes.
In comparison with previous study of a plain panel of similar size (1200x600x30), I noted that the response seemed to have more peaks, and so a greater overlap between peaks, giving fewer large variations. I also believe there are more peaks below 100 Hz from the lobes, and this helps the bass response, but I need to repeat outside.
Paul
In comparison with previous study of a plain panel of similar size (1200x600x30), I noted that the response seemed to have more peaks, and so a greater overlap between peaks, giving fewer large variations. I also believe there are more peaks below 100 Hz from the lobes, and this helps the bass response, but I need to repeat outside.
- Green curve is panel in free air. Note the relatively linear decrease below ~400Hz at about 20 dB/decade.
- Blue curve is panel wedged into baffle with light foam strips. I remember this sounded really good IMO, although I didn't listen for a long time. Bass holds up much better.
- Magenta is full support probably closer to simple support, panel siliconed to 7mm ply. Vibration and strong resonance through the baffle and frame.
Paul
Again on coincidence frequency, I'd be very interested to observe it somehow on the panel. I believe that at coincidence you should see an abrupt increase in the SPL measured at 90 degrees to the panel normal (ie along the edge of the panel).
I looked for it and never found any evidence of this. Maybe will try again when I get time to drag everything out and make a start.
I looked for it and never found any evidence of this. Maybe will try again when I get time to drag everything out and make a start.
Christian.
I don't think anyone can claim full absorption ,even gobel ?
the ohm Walsh driver is another bending wave drive unit.
The idea is to damp , if not all, but as much as possible ,any reflections from the edges.
Even cone drivers use the roll surround to help prevent reflections back into the cone.
as do bmr drivers.
These bending wave speakers go to extreme measures to stop dml modes by stopping , or should I say trying to prevent reflections from the edges.
This is the vast difference between bending wave units and DML units.
DML units encourage reflections !
Other forms of damping or clamping are inbetween these two extremes.
Steve.
I don't think anyone can claim full absorption ,even gobel ?
the ohm Walsh driver is another bending wave drive unit.
The idea is to damp , if not all, but as much as possible ,any reflections from the edges.
Even cone drivers use the roll surround to help prevent reflections back into the cone.
as do bmr drivers.
These bending wave speakers go to extreme measures to stop dml modes by stopping , or should I say trying to prevent reflections from the edges.
This is the vast difference between bending wave units and DML units.
DML units encourage reflections !
Other forms of damping or clamping are inbetween these two extremes.
Steve.
Thank you Paul
You are right, the blue curve looks nice.
The difference between the green (free air) and the blue curve (foam suspension) is similar to what I see in my current tests between a free air panel (or even one suspended by 2 strings) and one with a foam suspension (even not all around the panel). My test panel is just a rectangular 520x30cm) piece of XPS 9mm thick.
Christian
I add this idea in my test list... we'll see. It has perhaps not a big importance but as it is a remarkable value, it might be a value to be able to measure.
Christian.
If you mean how do bending wave units sound ?
They sound very capable, but usually flawed by putting them in a box.
I have only heard the ohm Walsh and manger units.
The gobel I have only seen in YouTube videos and the speakers are massive and very expensive.
But they are still only using the one small bending wave panel per speaker.
I'm not sure how any of these units would handle very loud heavy music ?
I have heard that the Walsh unit has a problem with high db .?
There is I believe an old Walsh forum on this site , with people building their own units ?
You might get a better idea from them?
Steve.
If you mean how do bending wave units sound ?
They sound very capable, but usually flawed by putting them in a box.
I have only heard the ohm Walsh and manger units.
The gobel I have only seen in YouTube videos and the speakers are massive and very expensive.
But they are still only using the one small bending wave panel per speaker.
I'm not sure how any of these units would handle very loud heavy music ?
I have heard that the Walsh unit has a problem with high db .?
There is I believe an old Walsh forum on this site , with people building their own units ?
You might get a better idea from them?
Steve.
I am at the moment still listening to these very low quality ply panels made from old fruit and veg crates.
I thought they were thinner but they are 3mm.
I am enjoying the sound ,but they do have a few problems to sort out ,mainly caused by the exciter I believe ?
they have oodles of treble without a large peak and a very lively sound.
lots of detail, leaning towards the EPS sound and efficiency , which is the surprising thing ?
I have not done any comparisons yet, but they can definitely go a lot louder without the exciter getting hot.
Which is nice 😃
Steve.
I thought they were thinner but they are 3mm.
I am enjoying the sound ,but they do have a few problems to sort out ,mainly caused by the exciter I believe ?
they have oodles of treble without a large peak and a very lively sound.
lots of detail, leaning towards the EPS sound and efficiency , which is the surprising thing ?
I have not done any comparisons yet, but they can definitely go a lot louder without the exciter getting hot.
Which is nice 😃
Steve.
@Leob
Hello Leob,
I made some measurements to see what happens at high frequencies with different materials
In the first FR below :
- blue 2.8mm plywood 17x27cm. The HF seems a bit smoother but not go very far beyond 10k.
Seems this light EPS sample has a lower capability to give HF.
There are 2 things to see in the spectrograms (see below)
Christian
Free EPS 7mm
XPS 9mm on 2 strings
Plywood
Hello Leob,
I made some measurements to see what happens at high frequencies with different materials
In the first FR below :
- blue-green (the lower one) is the XPS 9mm 20x30cm. See the hump at 9k. 20k is much lower
- orange is a 7mm piece of EPS (medium density I guess, it is from a packaging) 13x17cm. The hump is at 8k. 20k is low also
- green (upper) is a large 60x80cm 20mm light EPS (15kg/m³). No hump. The level was increased by 10dB for reading. If you remove those 10dB, this panel is below the others from 4k and above.
- blue 2.8mm plywood 17x27cm. The HF seems a bit smoother but not go very far beyond 10k.
Seems this light EPS sample has a lower capability to give HF.
There are 2 things to see in the spectrograms (see below)
- The free panels (only linked to the voice coil) might show important level far in time from the origin
- The PS panels show this resonance in the 8 to 10k area that plywood doesn't have.
Christian
Free EPS 7mm
XPS 9mm on 2 strings
Plywood
Thank you Steve. Yes it was my question.
Years ago, I built a pair of Walsh units based on a steep cone made from a plastic foil usually used to wrap gifts. On the top I had a full range speaker without its membrane. When we imagine the steep cone is between the coil and the suspension, high level might be a problem. I remember having a not too bad result. With the lack of information at that time, I stopped this experience.
Christian
It is I think exactly because of that I am with a plywood design for now and also because I have difficulties to get the same bandwith from EPS (see my post 5334 just before). For the efficiency, I have chosen poplar which available in the DIY store where I go usually.
Christian
Very interesting Christian, thanks!
Going to do a similar test to compare with the ply and polycarbonate using the new exciters. Been enjoying listening to the plates so much that I don't really want to remove the exciters to do experiments
And like I don't have enough projects going, I got hold of a nice sub that I hope should be able to keep up with my test plates to get a feeling about how it works out as a system, but it needs repairing. Was happy to open it up and see a cap looking like the hulk with the plastic bursting open, so hoping that replacing that is all there is to it.
I've looked for it myself by holding positioning the mic at just a bit less than 90 deg and looking for a peak in response that "should" be there. I've never found it. But, if you have good enough equipment, apparently you can. This figure is from a Paper by B Zenker (link below) and shows the change in directivity above the coincidence frequency, indicated by the dotted line.
https://www.researchgate.net/public...-_Evaluation_of_the_Voice_Coil_Break-Up_Modes
The Tectonic site includes this plot which shows a similar change in character above about 5 kHz, coincidentally (haha).
I'm thinking that these plots like these would actually be a useful basis for discussing the "regions" of the frequency response of a DML.
Eric
Attachments
Further along the lines of defining various regions of the frequency response, it would be useful to look at the surface acceleration plot along with the polar frequency response:
https://www.researchgate.net/public...zing_the_Off-Axis_Radiation_below_Coincidence
It's especially cool that you can see in the surface acceleration plots that the (1,1) mode is operating at about 70 Hz, the (1,3) mode at 100 Hz and the (1,5) mode at 180 Hz! I wish I could make those plots.
Eric
https://www.researchgate.net/public...zing_the_Off-Axis_Radiation_below_Coincidence
It's especially cool that you can see in the surface acceleration plots that the (1,1) mode is operating at about 70 Hz, the (1,3) mode at 100 Hz and the (1,5) mode at 180 Hz! I wish I could make those plots.
Eric