Hi Christian.
I always look at the response in real time so that I can see the response change as I move my hand around the edges.
This is the best way to understand how this works.
let's not forget that weights placed in from the edges also changed frequencies on the panel.
you have used an all on method and an all off method ,which is best ?
But it is very clear how this has changed the panel character.
the response in the 150hz to 300hz has increased by about 10db ,this will give voices a much warmer fuller sound.
The response below 150hz has gone the other way and reduced ,up to 10db in places.
If tests are done in real time ,you can fine tune these dips and peaks ,far quicker and easier.
You can watch them going up and down as you move your hands or foam pads around the panels.
Heavier panels will be less sensitive to the touch than EPS, you did not say which material you were using.
So if you placed your finger tips on to an EPS panel it will make a great difference, but placing you finger tips on a 3mm ply will make very little difference.
This is why I treat different panel materials ,differently.
You can't assume that one panel will react the same as another even if it is the same material but a different size or shape.
Different panels have different characteristics and quirks.
Steve.
I always look at the response in real time so that I can see the response change as I move my hand around the edges.
This is the best way to understand how this works.
let's not forget that weights placed in from the edges also changed frequencies on the panel.
you have used an all on method and an all off method ,which is best ?
But it is very clear how this has changed the panel character.
the response in the 150hz to 300hz has increased by about 10db ,this will give voices a much warmer fuller sound.
The response below 150hz has gone the other way and reduced ,up to 10db in places.
If tests are done in real time ,you can fine tune these dips and peaks ,far quicker and easier.
You can watch them going up and down as you move your hands or foam pads around the panels.
Heavier panels will be less sensitive to the touch than EPS, you did not say which material you were using.
So if you placed your finger tips on to an EPS panel it will make a great difference, but placing you finger tips on a 3mm ply will make very little difference.
This is why I treat different panel materials ,differently.
You can't assume that one panel will react the same as another even if it is the same material but a different size or shape.
Different panels have different characteristics and quirks.
Steve.
No problem. We all make assumptions, possible explanations and going to proof is difficult for DIYer. The important is to try to go forward, sorting good and bad ideas. and for that, ideas and proposals are welcome.
So I can report (as some of you knew already and tried to tell me) that the Zobel network doesn't seem to affect the FR of my high inductance exciters at all. If it seems to good to be true, apparently it is 
Started preparing a 25 kg/m3 EPS panel today to do some tests on, but I do feel like I do have to try making sandwich composites as well. Like I mentioned before I'm a bit sceptical of getting a good result without vacuum. This is an application that is extremely sensitive to any delamination over time which would make the plate distort and rattle, and you also want to minimize the amount of epoxy used. The reason to use vacuum in the first place I would have thought is to avoid delamination while using minimal amount of resin?
Started preparing a 25 kg/m3 EPS panel today to do some tests on, but I do feel like I do have to try making sandwich composites as well. Like I mentioned before I'm a bit sceptical of getting a good result without vacuum. This is an application that is extremely sensitive to any delamination over time which would make the plate distort and rattle, and you also want to minimize the amount of epoxy used. The reason to use vacuum in the first place I would have thought is to avoid delamination while using minimal amount of resin?
Leob.
Although I would recommend using XPS for using with epoxy skins, I do intend applying epoxy to one of my 1cm EPS test panels .
My panels have already got a thin coating of pva , so if they work together (stick) the epoxy should not soak into the panel too much.
I would not recommend sanding XPS before coating with epoxy, the epoxy seems to stick to the shiny XPS very well ,without soaking in to the surface, which it would do if sanded.
I will just apply the epoxy with a roller as in the tech ingredients video, but without the fibre glass mesh.
This is in the hope that natural gravity will flatten out the surface to some extent ?
If I was applying this to an uncoated low quality EPS , such as the one I am using, I would be a little worried that the expanded polystyrene balls would start to pull off from the surface ,when rolling the epoxy ?
Steve.
Although I would recommend using XPS for using with epoxy skins, I do intend applying epoxy to one of my 1cm EPS test panels .
My panels have already got a thin coating of pva , so if they work together (stick) the epoxy should not soak into the panel too much.
I would not recommend sanding XPS before coating with epoxy, the epoxy seems to stick to the shiny XPS very well ,without soaking in to the surface, which it would do if sanded.
I will just apply the epoxy with a roller as in the tech ingredients video, but without the fibre glass mesh.
This is in the hope that natural gravity will flatten out the surface to some extent ?
If I was applying this to an uncoated low quality EPS , such as the one I am using, I would be a little worried that the expanded polystyrene balls would start to pull off from the surface ,when rolling the epoxy ?
Steve.
Christian,
I did the algebra and got the same equation as you for B/µ. So unless we both made the same mistake, it's correct. I didn't open and check the spreadsheet yet, but I will when I have the time to really look at it.
But meanwhile, a few comments. First, I should clarify what I meant when I said the the fg/nomex composite in the video "should" have a B value of 10 Nm. What I meant was simply that 10 Nm is what I estimated for B with my spreadsheet, for the construction as it was described. That is, for that particular thickness of nomex and that particular thickness of fiberglass, the stiffness should be about 10 Nm. I did not mean to suggest that I thought that 10 Nm was (or was not) a good target value for B. In truth, I don't often look directly at B when I design a panel (but rather it's effects). So I don't really think in terms of B by itself as a parameter to target. I do know that less than 1 is really flimsy and more than 100 is really stiff, but other than that I don't really have a target in mind.
Back to your spreadsheet. One nitpick: Sometimes you are using B and sometimes D for panel stiffness. Both are commonly used of course, but I know you meant to use only one or the other here.
Finally. I suspect your spreadsheet calculation are correct. But I'm trying to understand the practical use. That is, it's not really practical to take a panel with a particular density, and then change (only) it's stiffness. Density and stiffness are so dependent on each other. I only know how to imagine a panel construction, and then estimate both density and stiffness. What don't I understand?
I don't mean to sound negative, I only want to better understand your thinking.
Eric
Last edited:
The good side is story is you tried the Zobel network. For many characterics of DML I have had to test by myself. Keep the components. If you don't get the HF back with it, it is good for you amp... probably with a smaller capacitor. If I remember there is a 1µF/10Ohm network at the output of my amp.So I can report (as some of you knew already and tried to tell me) that the Zobel network doesn't seem to affect the FR of my high inductance exciters at all. If it seems to good to be true, apparently it is
Started preparing a 25 kg/m3 EPS panel today to do some tests on, but I do feel like I do have to try making sandwich composites as well. Like I mentioned before I'm a bit sceptical of getting a good result without vacuum. This is an application that is extremely sensitive to any delamination over time which would make the plate distort and rattle, and you also want to minimize the amount of epoxy used. The reason to use vacuum in the first place I would have thought is to avoid delamination while using minimal amount of resin?
Eric,
Your feedback is not negative at all but constructive as expected; and thank you for checking. I better understand also your approach.
As I propose a naming, I will clean the document of the "D".
My intention is to add an approach of the bending stiffness by making test beam (30cm x 3cm) and checking its deviation under a weight. I already did some test beams with materials I have. Not very precise materials are designs might be sorted. This help (in my opinion) to have an idea of what is too or not enough stiff. Stiffness is not something of a common understanding. We are used to evaluate weight, dimensions... but stiffness...
Christian
Very clear NaRenaud. Thanks a lot.
Is there some advice about well known "low viscosity resin" and "release spray"?... I don't know what will be next with DML for me but considering all those posts, composite seems being a serious way.
Christian
Just to come back on the 2 exciter per panel...
Below first graph shows a set of FR measurements I made in various conditions. Red is one exciter as "reference", all the other are 2 exciters, the second exciter was added in different positions, this was on 3 days so not exactly same position of the panel or the mic in the room.
In the first graph, only a 1/6 octave smoothing is applied. There are some differences but not really big.
The second graph is exactly the same set of measurement but a frequency dependent window is applied. There are more important differences.
I have no final conclusion for now... it seems something happening with 2 exciters. Difficult to say with in addition the question how REW works with the frequency dependent window. Can we hear the difference? I don't know. Not tested.
For now I will remain cautious with the 2 exciters design.
If a clear advantage is found in the next future in other frequency band for the 2 exciter design, it is possible to add a capacitor across the connections of one exciter to have only one emitted in HF. I did it with what I have as capacitor : 6.8µF. It the purple/blue line.
Below first graph shows a set of FR measurements I made in various conditions. Red is one exciter as "reference", all the other are 2 exciters, the second exciter was added in different positions, this was on 3 days so not exactly same position of the panel or the mic in the room.
In the first graph, only a 1/6 octave smoothing is applied. There are some differences but not really big.
The second graph is exactly the same set of measurement but a frequency dependent window is applied. There are more important differences.
I have no final conclusion for now... it seems something happening with 2 exciters. Difficult to say with in addition the question how REW works with the frequency dependent window. Can we hear the difference? I don't know. Not tested.
For now I will remain cautious with the 2 exciters design.
If a clear advantage is found in the next future in other frequency band for the 2 exciter design, it is possible to add a capacitor across the connections of one exciter to have only one emitted in HF. I did it with what I have as capacitor : 6.8µF. It the purple/blue line.
Christian,
Look for this, or something similarly described as a laminating epoxy resin from a French supplier.
https://www.easycomposites.co.uk/el2-epoxy-laminating-resin
This starter kit is pretty cool. You could make two (pretty small) panels with this amount of material. It includes the carbon fiber, epoxy resin, release film, and PVA mold release. You would need a glass plate or mirror, and something to use as a core, like PS foam or balsa or nomex or whatever you want to try in addition to the kit. It's a cheap way to give it a try.
https://www.easycomposites.co.uk/carbon-fibre-laminating-starter-kit
Being in the US I have never purchased from them, but their videos are very good, and I expect their products are also good.
Eric
Regarding the discussion of framing, I can affirm that mounting in a frame definitely boosts the bass. This is because, like box speakers, it prevents the cancellation of high and low pressure areas on the front and back of the speaker. Especially so for panels, because most of the sound from panels at lower freqs already comes from corner and edge modes - uncanceled areas which remain after the + and - nodes in the middle have cancelled.
I also tend to agree its detrimental to clamp though. One reason is that it's very difficult to brace the frame sufficiently that it doesn't itself resonate, rattle or hum. Probably a heavy cast iron frame would sound better when clamped
I think a planar baffle with minimal or intermittent suspension is the way to go, and allows you make create something you can stand up and position like a 'normal' speaker, with additional bass boost from the floor.
I also tend to agree its detrimental to clamp though. One reason is that it's very difficult to brace the frame sufficiently that it doesn't itself resonate, rattle or hum. Probably a heavy cast iron frame would sound better when clamped
I think a planar baffle with minimal or intermittent suspension is the way to go, and allows you make create something you can stand up and position like a 'normal' speaker, with additional bass boost from the floor.
Steve,
One thing I've been wondering about is this: what do they mean (exactly) when they say a panel is in "pistonic" mode?
They seem to frame it in terms of the exciter displacement. But exciter displacement, large or small, doesn't necessarily relate (in any obvious fashion, anyway) to "pistonic" motion.
Typically, "pistonic" implies rigid body motion, without bending, of the driving surface. But the tectonic panel is pretty well fixed at four points, so there's no way it can move even the slightest bit, without bending. So, in the literal sense at least, there is no way a tectonic panel fixed at four points can move like a piston (i.e without bending). The only way it can move is by bending!
So they must mean something different. Are they suggesting that at some frequencies, the panel is somehow "quasi-pistonic"? That is, not truly pistonic, but, "kind of" pistonic, but in a bending (i.e non pistonic) kind of way? Do they mean that only the fundamental mode is the main contributor to the SPL? Or do they mean that standing waves are the main contributors? Or do they mean something else? Any ideas?
Eric
Thank you Eric
I remember you have already pointing this source... To update the history file is in low priority currently; I should probably put it a bit more to the top.
Christian
Steve
Is it the "Conversations" icon (letter) just beside the "Alert" (bell) and Login one on the top right you are looking for? Then select the name you want to discuss with...
Good remark Eric
In the "basic tests" I have been doing, I made measurements of the same XPS small panel in 4 suspension conditions :
- Free panel (just connected to the voice coil)
- String suspension by the upper short side
- Foam suspension letting the corner area free
- Same adding 4 "fixed" points
Second graph shows the same 2 string suspension (green), the foam suspension letting the corner (dark blue) and the same with 4 additional fixed points (light green) : the foam suspension reinforced the 200Hz region, reducing a little the 50Hz hump while the 4 points "kill" the 50Hz hump.
So all what is mentioned in the patent is perhaps not related to the final design but sometimes more general...
I am wondering if a "pistonic mode" is visible... Even in a proximity measurement of the free panel (no suspension) there is almost nothing below the 50Hz hump.
All of that are tendencies... the foam results are probably very sensitive to the relation foam/panel. Here the foam is probably quite strong (only 5mm height) compare to the panel.
Christian
When they talk of pistonic mode, I believe they're talking of the dominant source of sound in the far field. They're not, necessarily, saying that other modes are completely absent. But the the mode that is most pistonic in character (1,1) is indeed the lowest. At lower and lower frequencies, the higher order modes with adjacent + and - antinodes tend to cancel more and more, but a single antinode can only cancel with its opposite at the rear of the panel, like conventional speakers.
Hello Pway,Regarding the discussion of framing, I can affirm that mounting in a frame definitely boosts the bass. This is because, like box speakers, it prevents the cancellation of high and low pressure areas on the front and back of the speaker. Especially so for panels, because most of the sound from panels at lower freqs already comes from corner and edge modes - uncanceled areas which remain after the + and - nodes in the middle have cancelled.
I also tend to agree its detrimental to clamp though. One reason is that it's very difficult to brace the frame sufficiently that it doesn't itself resonate, rattle or hum. Probably a heavy cast iron frame would sound better when clamped
I think a planar baffle with minimal or intermittent suspension is the way to go, and allows you make create something you can stand up and position like a 'normal' speaker, with additional bass boost from the floor.
Reading your post, I am not able to figure out where you are in the frame/suspension design. Before asking for more details, I had a look to your other posts and I understand you have read a lot and also have probably experimenting at lot?
But unfortunately, I haven't found any pictures or measurements of your design.
Might it be possible your share where you are and a bit more about suspension and frame shape according to you?
To more detailed points :
- Which material finally have you chosen?
- In post 3619, you propose "cusps on the edges opposite the exciter". Is it in the same way than Tectonic rubber supports?
- You seem also have experienced non rectangular shape. I also made some tests but I haven't found improvement on that (not to say the FR are even more irregular with such shapes). What is your opinion on this way?
Eric.
thank pway, but it is even more complicated than that 😁
you can get yourself in a lot of trouble mentioning pistonic mode, with cone drivers , let alone dml ?
people will start to show animations of cone break up and warping ,saying cones are not true pistons 😱
I have been saying for years that free floating panels are a combination of pistonic ,then ,bending wave, then dml, in that order.
Which is more dominant depends on the panel and whatever you do to it !
This is why it is very difficult to make sweeping statements about what is a dml panel and how does it work.
A 2ft panel ,approximately , should , if I remember correctly, with pistonic cone drivers , start to roll off somewhere in the 200hz area.
But the art panel and many of my smaller free floating panels all have an increase in output somewhere below 100hz ?
Taking burntcoils tall blondes ,for instance.
You have a long narrow panel with the exciter at the top , only the top of the panel can in anyway be called pistonic, the rest of the panel (as I see it) must be bending wave and then dml.
This panel has a surprisingly strong LF output.
Tectonic explain that they expand the central pistonic area by using 4 exciters , what they fail to mention of course is basically they are expanding the oil can affect also.
If the centre of the coil area (similar to cones) does not act like a piston what chance has the centre between 4 exciters?
It all becomes a can of worms, and you wish you'd never even mentioned it.
The only way to get around it is to say all drive units including dml are mainly ,or should I say sort of pistonic within the coil area ?
What happens after that is up to the designer ?
Steve.
thank pway, but it is even more complicated than that 😁
you can get yourself in a lot of trouble mentioning pistonic mode, with cone drivers , let alone dml ?
people will start to show animations of cone break up and warping ,saying cones are not true pistons 😱
I have been saying for years that free floating panels are a combination of pistonic ,then ,bending wave, then dml, in that order.
Which is more dominant depends on the panel and whatever you do to it !
This is why it is very difficult to make sweeping statements about what is a dml panel and how does it work.
A 2ft panel ,approximately , should , if I remember correctly, with pistonic cone drivers , start to roll off somewhere in the 200hz area.
But the art panel and many of my smaller free floating panels all have an increase in output somewhere below 100hz ?
Taking burntcoils tall blondes ,for instance.
You have a long narrow panel with the exciter at the top , only the top of the panel can in anyway be called pistonic, the rest of the panel (as I see it) must be bending wave and then dml.
This panel has a surprisingly strong LF output.
Tectonic explain that they expand the central pistonic area by using 4 exciters , what they fail to mention of course is basically they are expanding the oil can affect also.
If the centre of the coil area (similar to cones) does not act like a piston what chance has the centre between 4 exciters?
It all becomes a can of worms, and you wish you'd never even mentioned it.
The only way to get around it is to say all drive units including dml are mainly ,or should I say sort of pistonic within the coil area ?
What happens after that is up to the designer ?
Steve.