Sonar, tou.
Using 20db increments would flatten any response out ,and make it look good!
Using 5db increments would be more informative or at least no larger than 10db.
Also ,the response can change dramatically by moving the microphone from left to right and up and down ,even by only a few inches.
I usually use peak hold and wave the microphone around the seating area to get an idea of the in room response.
If you are lucky you can get similar responses.
Steve.
Using 20db increments would flatten any response out ,and make it look good!
Using 5db increments would be more informative or at least no larger than 10db.
Also ,the response can change dramatically by moving the microphone from left to right and up and down ,even by only a few inches.
I usually use peak hold and wave the microphone around the seating area to get an idea of the in room response.
If you are lucky you can get similar responses.
Steve.
You are right with teh measurement from sonar.
The measurement of the modipo is from a well known german magazine for loudspeaker building. I think they know how to measure. Additionally it is approved from some other people. They assuemd this to be a type of Nxt at LF:
Modipo -Hörbericht von Bertramxxl - Seite 3
It makes an addition of 5-6db at 20Hz. It would be interestig to now how to achieve this initionally...
The measurement of the modipo is from a well known german magazine for loudspeaker building. I think they know how to measure. Additionally it is approved from some other people. They assuemd this to be a type of Nxt at LF:
Modipo -Hörbericht von Bertramxxl - Seite 3
It makes an addition of 5-6db at 20Hz. It would be interestig to now how to achieve this initionally...
Last edited:
Spedge,
You are right about the 20db scale. It looks too good to be true. I am new to REW and calibrated mics.
The purpose of these tests are to see how different materials behave. IMHO I feel that we cannot take DMLs as pure full range speakers. They really need LF and HF support, there's though some 'magic', a sweet spot hard to ignore somewhere in the mids.
Sonnar,
My "subjective" experience runs counter to your thoughts on this.
The larger panels the more LF they produce. I think room size limits bass more than the capacity of a well formed DML.
Rigidity of panels vs weight gets you all the HF and MF.
As I've mentioned before, I suspect that the higher the rigidity vs weight the better the panels do everything. If correct, aluminum or titanium honeycomb core with aluminum or titanium skins likely best.
Also, large panels with a high rigidity to weight ratio don't need multiple exciters.
Additional reinforcement at the edges may also help, but one can get these honeycomb panels with that already incorporated.
Unfortunately such panels don't come cheap.
My current thought model in order of suspected contribution:
I had hoped to do some measurements over the past weekend, but spent most of it designing and beginning to assemble a more permanent support frame for my panels.
I'll get to the measurements and will post a recording or two as soon as life permits.
My "subjective" experience runs counter to your thoughts on this.
The larger panels the more LF they produce. I think room size limits bass more than the capacity of a well formed DML.
Rigidity of panels vs weight gets you all the HF and MF.
As I've mentioned before, I suspect that the higher the rigidity vs weight the better the panels do everything. If correct, aluminum or titanium honeycomb core with aluminum or titanium skins likely best.
Also, large panels with a high rigidity to weight ratio don't need multiple exciters.
Additional reinforcement at the edges may also help, but one can get these honeycomb panels with that already incorporated.
Unfortunately such panels don't come cheap.
My current thought model in order of suspected contribution:
Rigidity vs weight
Panel size
Panel edge rigidity
Suspension/isolation
Amplification
Exciter placement
Panel size
Panel edge rigidity
Suspension/isolation
Amplification
Exciter placement
I had hoped to do some measurements over the past weekend, but spent most of it designing and beginning to assemble a more permanent support frame for my panels.
I'll get to the measurements and will post a recording or two as soon as life permits.
I found interesting that the aes paper comes to the conclusion, that the min and max frequency only depands on the mass of the coil Mc, the magnet Mm and the stiffness Rp:
Fmax= Rp/(2pi*Mc)
Fmin= Rp/(2pi*Mm)
The size has more to do with the efficiency.
What do you think. I have found another study, with the conclusion that with an array of exciter (more magnets) you can get LF from very small panels...
Edit:
Probably we could use two exciteres. One for LF and one for HF. This should make it possible to use more different panels... not only alu or titan
Fmax= Rp/(2pi*Mc)
Fmin= Rp/(2pi*Mm)
The size has more to do with the efficiency.
What do you think. I have found another study, with the conclusion that with an array of exciter (more magnets) you can get LF from very small panels...
Edit:
Probably we could use two exciteres. One for LF and one for HF. This should make it possible to use more different panels... not only alu or titan
Last edited:
Very interesting and useful. I can incorporate this into a Mathematica model.
Interesting.
I've only made a single set of DMLs (aluminum honeycomb panels roughly 71" x 28") they play loud (and do so seemingly effortlessly) and they have more bass than any speakers I've ever had in my space before this. I "feel" like I've never before heard all the bass the room could support.
Of course actual measurements may show all of this as just in my head ;-)
Size may mostly have to do with efficiency, but (naively) the larger the vibrating panel, or cone, or ribbon, or planner, the more bass.
All of this gives all of us very interesting lines of research.
Here is the study i mentioned before:
A novel miniature matrix array transducer system for loudspeakers -ORCA
In the study he says that with one exciter at longer waves (LF) on small panel you get less bass response. But he come to the conclusion that with more (Array) and stronger exciters you can solve this, because the panel then begins to move as complete with the frequency.
Mey English is not so good. Probably the best if you take a look on the study....
A novel miniature matrix array transducer system for loudspeakers -ORCA
In the study he says that with one exciter at longer waves (LF) on small panel you get less bass response. But he come to the conclusion that with more (Array) and stronger exciters you can solve this, because the panel then begins to move as complete with the frequency.
Mey English is not so good. Probably the best if you take a look on the study....
This still looks lumpy to me,and the hf roll off above 6k I would regard as unacceptable for a full range panel ,which seems to be an ongoing problem with exciter and panel mismatch,which I believe is unnecessary.
Of course you could use eq but do we really believe you can bring back what has been mechanically lost?
This graph has a 100db window with the trace squeezed in at the very top,it could be more informative if centred more on the waveform .
I know it takes time to learn how to use these new measurement techniques ,so it is not a criticism ,but just an observation .
All the best.
Steve.
I'm quite new to this, but I was wondering whether someone already tried to build a round, circular panel speaker, and moreover, whether this produces good quality sound. The tech ingredients yt channel used one round panel in their experiment, but they did not dug deeper than that. Because of aesthetically purposes I would propose a round one. To extend on this, I heard lower tones are not always that well for flat panel speakers, would it work to add a woofer speaker in the center of my (future) round panel? something like my attached sketch?
kind regards, Bas
kind regards, Bas
Attachments
Baswess99
Hi.
You do not mention panel size and materials?
Placing a cone or exciter in the centre of a circular panel is not a good idea,for well known reasons.
Yes I have made large and small round panels,but with offset exciters.
Not sure why round panels would be more aesthetically pleasing ,but I'm sure you have your reasons.
What is your preferred xo point?
Steve.
Hi.
You do not mention panel size and materials?
Placing a cone or exciter in the centre of a circular panel is not a good idea,for well known reasons.
Yes I have made large and small round panels,but with offset exciters.
Not sure why round panels would be more aesthetically pleasing ,but I'm sure you have your reasons.
What is your preferred xo point?
Steve.
Thought I'd try and make the point,with my deq set to 20db setting ,even though it has 10db dotted lines,at one mtr.
This is just one of my card panels ,that I haven't managed to get around to taking apart yet.
I just wired it up and set the sub levels ,and that was it,3mins max,no eq.
The response looks very nice ,to me anyway,even with a load of junk piled in front of my transmission lines!
The card panel is doing everything above 300hz and stays within about +-6db with strong output up to and beyond 20k ,it only starts to get lumpy below 300hz ,when the TLs kick in ,according to this setting anyways?
If I move it up to 30db with 15db lines it starts to look like a straight line!
I never use the 20 db setting as it hides too many problems.
EQ would be very easy with this panel,by just lowering the midrange response a little,but normally I would try and keep the response as close to +-3db as possible,without EQ.
Steve.
This is just one of my card panels ,that I haven't managed to get around to taking apart yet.
I just wired it up and set the sub levels ,and that was it,3mins max,no eq.
The response looks very nice ,to me anyway,even with a load of junk piled in front of my transmission lines!
The card panel is doing everything above 300hz and stays within about +-6db with strong output up to and beyond 20k ,it only starts to get lumpy below 300hz ,when the TLs kick in ,according to this setting anyways?
If I move it up to 30db with 15db lines it starts to look like a straight line!
I never use the 20 db setting as it hides too many problems.
EQ would be very easy with this panel,by just lowering the midrange response a little,but normally I would try and keep the response as close to +-3db as possible,without EQ.
Steve.
Attachments
Last edited:
I'd like to use a xps foam panel from approximately 65-70cm dia, 20mm thickness, as I heard in several experiments to be optimal, right?
my initial idea was to place a exciter according to the 2/5 / 3/5 principle, and the woofer as can be seen in the picture as a separate lower producing speaker. But as you show already, the round panel can produce good sound at full range, so I guess a woofer is not needed at all? Correct me if im wrong, I'm quite a beginner in diy audio.
my initial idea was to place a exciter according to the 2/5 / 3/5 principle, and the woofer as can be seen in the picture as a separate lower producing speaker. But as you show already, the round panel can produce good sound at full range, so I guess a woofer is not needed at all? Correct me if im wrong, I'm quite a beginner in diy audio.
Baswess99.
I have no idea what size is optimal for 20mm xps,I have not seen frequency plots that get anywhere near 10k without large amounts of EQ,no matter what size panel?
It would be interesting to see these experiments.
I have only used xps with a maximum of 5mm thickness ,which did have better results in the HF.
The only small panel I can think of that produces a decent low end down to 40hz was the artist canvas picture panels.
Even this I would use with sub.
I also don't believe I have shown a round panel (or any other shaped panels for that matter ) producing what i would call fullrange ,in realistic room conditions.
Mounting a low frequency driver to your dml panel would also hinder the performance of your dml panel.
I'm not sure what to suggest without knowing how you intend to mount your panels.
Hanging from ceiling. On a wall , standing on a surface?
Or as part of a stand alone speaker?
Steve.
I have no idea what size is optimal for 20mm xps,I have not seen frequency plots that get anywhere near 10k without large amounts of EQ,no matter what size panel?
It would be interesting to see these experiments.
I have only used xps with a maximum of 5mm thickness ,which did have better results in the HF.
The only small panel I can think of that produces a decent low end down to 40hz was the artist canvas picture panels.
Even this I would use with sub.
I also don't believe I have shown a round panel (or any other shaped panels for that matter ) producing what i would call fullrange ,in realistic room conditions.
Mounting a low frequency driver to your dml panel would also hinder the performance of your dml panel.
I'm not sure what to suggest without knowing how you intend to mount your panels.
Hanging from ceiling. On a wall , standing on a surface?
Or as part of a stand alone speaker?
Steve.
I am now getting an annoying 'buzzing' sound from both Dayton Audio DAEX25FHE-4 Framed exciters. It only goes when I push the exciter firmly against the plywood panel.
I can either rig up a modified support to add pressure to the exciter, or look for alternatives. The Monacor EX-40/8 would be a good match for my 8 ohm bass driver, but I see few mentions or reviews.
I'm running the exciters / DML panel crossed over from 320hz (24db/oct LR) and with EQ on the miniDSP get a very flat response to 17khz before it drops off. A Monacor 15 inch bass driver in a sort of U frame OB thing does the rest. The panel size is 400mm x 565mm.
So, I'd be grateful for any suggested alternatives to my existing Dayton Audio units and hopefully something a bit more reliable.
I can either rig up a modified support to add pressure to the exciter, or look for alternatives. The Monacor EX-40/8 would be a good match for my 8 ohm bass driver, but I see few mentions or reviews.
I'm running the exciters / DML panel crossed over from 320hz (24db/oct LR) and with EQ on the miniDSP get a very flat response to 17khz before it drops off. A Monacor 15 inch bass driver in a sort of U frame OB thing does the rest. The panel size is 400mm x 565mm.
So, I'd be grateful for any suggested alternatives to my existing Dayton Audio units and hopefully something a bit more reliable.
Lordtarquin
There are usually three reasons buzzing can occur.
1.
Suspension sag.
Turn the panel upside down and see if this helps.
2.
The glue contact with the exciter and panel is coming loose.
Removing the exciter and re-gluing in place.
3.
The foot is coming loose from the coil former.
Gently remove foot and re glue back into place.
I have had all of these problems at one time or another, all these problems are cheep and easily fixed.
Steve.
There are usually three reasons buzzing can occur.
1.
Suspension sag.
Turn the panel upside down and see if this helps.
2.
The glue contact with the exciter and panel is coming loose.
Removing the exciter and re-gluing in place.
3.
The foot is coming loose from the coil former.
Gently remove foot and re glue back into place.
I have had all of these problems at one time or another, all these problems are cheep and easily fixed.
Steve.
Thanks Steve.
After trying many things, it turned out to be a new mounting arrangement I was trying. Basically, the panel was resonating against the foam cushioning I was using.
By having the panel 'floating' in the frame, attached at 20 points with thin strips of 3mm neoprene foam on both sides of the panel. The buzz has gone, and with EQ, the panel is now essentially flat over its desired range, although it does fall off rather sharply above 15khz, being about 7db down at 20khz. There may be a bit of wiggle room to boost the very top end.
Cheers
Simon
After trying many things, it turned out to be a new mounting arrangement I was trying. Basically, the panel was resonating against the foam cushioning I was using.
By having the panel 'floating' in the frame, attached at 20 points with thin strips of 3mm neoprene foam on both sides of the panel. The buzz has gone, and with EQ, the panel is now essentially flat over its desired range, although it does fall off rather sharply above 15khz, being about 7db down at 20khz. There may be a bit of wiggle room to boost the very top end.
Cheers
Simon