A Study of DMLs as a Full Range Speaker

[homeswinghome]

I think high frequency is micro vibration
It cannot produce large area micro vibration
Excluding AMT (Air Motion Transformer, Heil Tweeter)
So it should be right to install many Exciters
More coils allow more areas to generate high-frequency sound

If you mean that several exciters used in one DML are a solution to get more HF...
it might be true for the mid high frequencies if the panel absorption is too high leading to a reduce surface in this range
not so sure in high frequency if alone, one of those exciters is not able to produce HF (due to panel mechanical impedance, voice coil mass, coil inductance).
Before going this way, if needed, I would explore the way of a reduce voice coil diameter exciter which was already suggested here.

But when listening closely
I found that the coil was slightly noisy when it was reciprocating at high frequency
So it should be more perfect to use the tweeter
Having one coil responsible for too much of the frequency range would seem to be slightly noisy right?
So I'm also hesitant to try multiple coils!?

I don't see to which noise you are referring to? For DML, I have in mind 2 sources of "noise" :

• What is called in papers the membrane self noise. It is the frequencies not in the input signal produced by the material non linearity. It can be shown with a "contamination test". Search on the forum with this words. I posted about that. It is a matter of material choice to reduce it (this is a technical statement, is it necessary?)
• The exciter noise which in my understanding the noise produced by the air trapped at the rear of the membrane in the voice coil area. Not to difficult to detect in the FR with the spectrogram view or a measure on the rear side. Some solution here is needed has the peak in the FR is important. I don't have the solution...

When multiplying the exciters, you reduce the level for each (for a given overall level)... so does it really increase the "noise"?

 

[Audiofrenzy]

I'm not sure what you mean by micro vibration.
The voice coils have a maximum excursion of around 3mm. This would be for max SPL at, say, 100Hz depending on your panel dimensions.
Only a fraction of this excursion would be required to generate the same SPL at 1khz and even a smaller fraction at 10khz. ALL frequencies are micro-vibrations depending how loud you want to go. And the whole panel vibrates in a quasi-random fashion if there's no damping involved. The better the positioning of the exciter, the more parts of the panel will vibrate.
You will most certainly not get better HF simply by using more exciters. You will get more power handling, and more SPL, and if you get the positioning right then maybe a flatter response curve, and maybe a bit more HF. But not as a rule.

If it's not mounted to a panel, then the voice coil should be absolutely dead quiet when driving any frequency, except for a little bit of HF air action coming from the spider suspension, and also a little bit from the flat glue ring that sticks to the panel. At high power you might also notice some distortion but which is usually hidden by the SPL coming from the panel.

If you want to use a cone tweeter then you are going straight back to the problems inherent in cone speakers, i.e. beaming, lobing, dead spots, hot spots, bad imaging, shallow sound stage and a very small sweet spot. Of course, you will also need a cross-over and you need to try keep the cross-over frequency away from the 1khz-4khz band.
A tweeter on a DML panel is not a very elegant solution. (Yes I know Tectonic have used a planar tweeter (or is it a dipole ribbon?) in some of their panels, and I would concede that's an okay-enough compromise. But it still needs a cross-over!)

I would like to know what is meant by "noisy" noise. I see this word is used a lot on the forum. Is this what is meant instead of distortion? Or is it mechanical buzzing and rattling? Rubbing? Over-excursion?

"Absolutely dead quiet"= except for a little bit of HF air action? Either you are trying to confuse/mislead people or you have a very poor choice of words that contradict each other. The same way you said that 8ohm exciters lack HF. That is very misleading as they dont lack HF they just have a tad bit less. Its like saying well I only buy 4ohm tweeters because 8ohm tweeters lack HF.

Elegant? Again choice of words. Then contradicts himself by saying yes I know tectonic uses a planar tweeter.

Then he says "keep the cross over freuency away from the 1-4khz band". Most used crossover points are in the 1-4khz range.

Then my last question is have you tried implementing a tweeter in your design or is it just all talk and no walk as usual?

 

[Audiofrenzy]

Seems additional weights are in most of the commercial DML. One of their role is to modify the modes and smooth the FR.

Audiofrenzy,
You have deeply studied the Bertagny and used their solutions.
What is the reason of the additional tweeter? Is its band limted because of the optimisation of the bass/mid despite the high density EPS?
We often see the back of the speaker to show what is inside. Is it a true open back speaker or is there some damping/attenuation material?

There is more to using weights then to just smooth out the FR as it all depends on ones design.

Again it depends on the design. That is one factor out of many many factors. I give out some hints but none of you pick it up. All materials act as a filter to certain degrees some more then others. A hard stiff/denser material like thin Balsa or Bass wood will have better HF then EPS, not just because its stiffer but because the midrange frequencies is reduced (filtered) on stiffer materials so the highs are not drown out as much by the mid frequencies. This is why the frequency response should be fairly flat as you dont want some frequencies to drown out the other.

The only damping material is the foam surround material used to hold the EPS diaphragm in place.

 

[Audiofrenzy]

Is fiberglass and resin part of your DML cookbook?

No, but they make a good bullet proof vest for around 30 bucks.

 

[Audiofrenzy]

Not generally, no.
Since distortion is generally very low in a good loudspeaker, (less than 5 parts in 1000 for a good loudspeaker), it's unlikely that you would pick up the distortion in an FR graph.
Of course there are those who claim that distortion is good in hifi 🤦‍♂️ but that's probably because they have never heard a really good system at play, and they probably think that their personal tastes carry more weight that that of experienced recording engineers. Or maybe they're just blown away by their little Bluetooth USB speakers.

That's actually one of the problems with really good systems: Amateurs are so used to cheap sound that they will choose the worst, most distorted option almost every time in blind tests.

Again why is no one questioning this post? SIGH!!!!!

 

[xsuper9988]

Distortion is not an information of the FR curve by itself as the FR is the linear aspect of the transfer function but tools like REW thanks to the logsweep measurement has the capability to mesure it. You can add easily the different distortion orders or the total distortion upon the FR;

I just started more than half a year
I don't know how to judge the distortion
But I guess it should be what I said before
Overcorrection on EPS panels using EQ
So how can we generate a lot of "second harmonic distortion"?
Because I'm curious what does "Second Harmonic Distortion" sound like?

Seems additional weights are in most of the commercial DML. One of their role is to modify the modes and smooth the FR.

Audiofrenzy,
You have deeply studied the Bertagny and used their solutions.
What is the reason of the additional tweeter? Is its band limted because of the optimisation of the bass/mid despite the high density EPS?
We often see the back of the speaker to show what is inside. Is it a true open back speaker or is there some damping/attenuation material?

You may not have noticed old posts
The impression is that the post mentioned that he has a set of "Bertagni"
So he should be able to get a lot of references from it
I don't seem to have seen anyone offering "Bertagni" in my country
But it is estimated that the price of such high-end antiques should be very high

 

[xsuper9988]

I'm not sure what you mean by micro vibration.
The voice coils have a maximum excursion of around 3mm. This would be for max SPL at, say, 100Hz depending on your panel dimensions.
Only a fraction of this excursion would be required to generate the same SPL at 1khz and even a smaller fraction at 10khz. ALL frequencies are micro-vibrations depending how loud you want to go. And the whole panel vibrates in a quasi-random fashion if there's no damping involved. The better the positioning of the exciter, the more parts of the panel will vibrate.
You will most certainly not get better HF simply by using more exciters. You will get more power handling, and more SPL, and if you get the positioning right then maybe a flatter response curve, and maybe a bit more HF. But not as a rule.

Then if "very small energy" it can emit sound over a large area
The dome tweeter should be able to produce a very large sound with low power
But judging from the results, even if it is a high-power dome tweeter
Its diaphragm is still small

If it's not mounted to a panel, then the voice coil should be absolutely dead quiet when driving any frequency, except for a little bit of HF air action coming from the spider suspension, and also a little bit from the flat glue ring that sticks to the panel. At high power you might also notice some distortion but which is usually hidden by the SPL coming from the panel.

Everyone tested this before
You can take a spare Exciter test
He still has high frequency, low frequency sounds

I would like to know what is meant by "noisy" noise. I see this word is used a lot on the forum. Is this what is meant instead of distortion? Or is it mechanical buzzing and rattling? Rubbing? Over-excursion?

I use FR speaker as mid-high
And the scope of his responsibility is too large (1K~20K)
And the EQ is overcorrected
When I was very close to the FR speaker, I felt a slight abnormal sound

 

[xsuper9988]

If you want to use a cone tweeter then you are going straight back to the problems inherent in cone speakers, i.e. beaming, lobing, dead spots, hot spots, bad imaging, shallow sound stage and a very small sweet spot. Of course, you will also need a cross-over and you need to try keep the cross-over frequency away from the 1khz-4khz band.
A tweeter on a DML panel is not a very elegant solution. (Yes I know Tectonic have used a planar tweeter (or is it a dipole ribbon?) in some of their panels, and I would concede that's an okay-enough compromise. But it still needs a cross-over!)

I don't think there are specific needs for a tweeter because used with a DML..
The tweeter introduces :

• a crossover with the question linked to that (phase, distance between the source center...)
• a big difference in the radiation pattern. No HF to the rear side, a reduction of the radiation angle when the frequency increases.

From the only FR I have seen from the Yamaha speaker, it stops emission too early;

I have done several groups and tests so far
It is DML+OB medium and high frequency (FR speaker)
I don't have a lot of speakers to test
I have recently purchased some that have not yet arrived
But I can't hear any incongruity in "DSP crossover"???
Maybe I adjusted it with EQ so I can't hear it

The questions you mentioned are very similar
But I adjusted it with EQ
I really don't feel anything about the "crossover" part

a big difference in the radiation pattern. No HF to the rear side, a reduction of the radiation angle when the frequency increases.

I will test this part again
Whether the DML is evenly distributed in the treble part

Thank you very much everyone for sharing and replying~

 

[xsuper9988]

If you mean that several exciters used in one DML are a solution to get more HF...
it might be true for the mid high frequencies if the panel absorption is too high leading to a reduce surface in this range
not so sure in high frequency if alone, one of those exciters is not able to produce HF (due to panel mechanical impedance, voice coil mass, coil inductance).
Before going this way, if needed, I would explore the way of a reduce voice coil diameter exciter which was already suggested here.

this is not my solution
This solution is currently used by @Andre Bellwood
He used four exciters
I might try it later
Maybe a good solution

I don't see to which noise you are referring to? For DML, I have in mind 2 sources of "noise" :

• What is called in papers the membrane self noise. It is the frequencies not in the input signal produced by the material non linearity. It can be shown with a "contamination test". Search on the forum with this words. I posted about that. It is a matter of material choice to reduce it (this is a technical statement, is it necessary?)
• The exciter noise which in my understanding the noise produced by the air trapped at the rear of the membrane in the voice coil area. Not to difficult to detect in the FR with the spectrogram view or a measure on the rear side. Some solution here is needed has the peak in the FR is important. I don't have the solution...

When multiplying the exciters, you reduce the level for each (for a given overall level)... so does it really increase the "noise"?

I posted above
I'll add another explanation
I think that's the sound of coils
Should I call it distortion overcorrected with EQ?

I think could use more Exciters
But the uniformity and smoothness of the frequency should be enough
Otherwise overcorrection will have "coil sound" (I don't know what to call it, let's call it that)

Thank you for your reply and sharing~

 

[xsuper9988]

Before going this way, if needed, I would explore the way of a reduce voice coil diameter exciter which was already suggested here.

It is hard to find one with small diameter and enough power
There are also high unit prices as @Andre Bellwood said before.

When multiplying the exciters, you reduce the level for each (for a given overall level)... so does it really increase the "noise"?

I am currently a three-way frequency
Subwoofer + DML +FR
I am not using multiple Exciters

a big difference in the radiation pattern. No HF to the rear side, a reduction of the radiation angle when the frequency increases.

Add the missing part
I use an OB style FR
It is open with "positive waves" + "negative waves"

But my FR is not good enough
I think over correcting high frequencies
So there is the "coil sound" produced by the distortion

I have ordered some new FR and dome tweeters
I'll do a comprehensive test later

I think the main advantage of DML is to replace the expensive large size cone speaker, and its diffusion is better

 

[xsuper9988]

Which Yamaha model is this loudspeaker? Seems the "DML" and the tweeter are firing in same direction (not the case of other models being 3 ways as other Yamaha models with such "natural sound" or ear speakers).

NS full series introduction PDF Download​

NS-10(1 Way)

NS-15、NS-18(2 Way)

NS-20、NS-30(3 Way)

​

 

[toddincabo]

"Absolutely dead quiet"= except for a little bit of HF air action? Either you are trying to confuse/mislead people or you have a very poor choice of words that contradict each other. The same way you said that 8ohm exciters lack HF. That is very misleading as they dont lack HF they just have a tad bit less. Its like saying well I only buy 4ohm tweeters because 8ohm tweeters lack HF.

Elegant? Again choice of words. Then contradicts himself by saying yes I know tectonic uses a planar tweeter.

Then he says "keep the cross over freuency away from the 1-4khz band". Most used crossover points are in the 1-4khz range.

Then my last question is have you tried implementing a tweeter in your design or is it just all talk and no walk as usual?

Ewwww, don't buck the Master! The Master has spoken. But I swear, The Master does remind me of the "glass doesn't bend" guy.

 

[xsuper9988]

Then if "very small energy" it can emit sound over a large area
The dome tweeter should be able to produce a very large sound with low power
But judging from the results, even if it is a high-power dome tweeter
Its diaphragm is still small

Hi @Andre Bellwood
I made a mistake in describing too briefly in the morning

Then if "very small energy" it can emit sound over a large area
The dome tweeter should be able to produce a very large sound with low power
But judging from the results, even if it is a high-power dome tweeter
Its diaphragm is still small



Then if "very small energy" it can emit sound over a large area
The dome tweeter should be able to sound a very large area at low power (very small energy)
But judging from the results, even if it is a high-power dome tweeter
Its diaphragm is still small

 

[Andre Bellwood]

NS full series introduction PDF Download​

NS-10(1 Way)
View attachment 1181753 View attachment 1181754

NS-15、NS-18(2 Way)
View attachment 1181755
View attachment 1181756
View attachment 1181757

NS-20、NS-30(3 Way)
View attachment 1181750 View attachment 1181751

​

I don't think these are DML. Are they open baffle?
I had a pair of Leak speakers like this, with a polystyrene cone, many years ago.

I see that they mention NS10's in that brochure...
Yamaha NS10's are probably the most widely used professional studio monitors. They are industry standards. But those on your attachment must be the original concept as I don't recognize them. They must have changed and evolved to the standard that eventually became so popular.

NS-10's

 

[Andre Bellwood]

Hi @Andre Bellwood
I made a mistake in describing too briefly in the morning

Then if "very small energy" it can emit sound over a large area
The dome tweeter should be able to produce a very large sound with low power
But judging from the results, even if it is a high-power dome tweeter
Its diaphragm is still small



Then if "very small energy" it can emit sound over a large area
The dome tweeter should be able to sound a very large area at low power (very small energy)
But judging from the results, even if it is a high-power dome tweeter
Its diaphragm is still small

Higher frequencies require less power and excursion to produce high SPL.
For example a 1,000W woofer can be perfectly balanced with a 200W mid (2nd order x-over at 500Hz) and a 50W tweeter (2nd order x-over at 3khz). And a 2,000W/channel amplifier will drive such a system adequately.

In addition to the above, the reason that tweeters are small in size is because of the moving mass issue--a small mass can produce higher frequencies than a large mass can when driven by the same force--and also because of the beaming problem. A large cone will start to beam the sound when the cone size is comparable to the wavelength.

 

[xsuper9988]

Higher frequencies require less power and excursion to produce high SPL.
For example a 1,000W woofer can be perfectly balanced with a 200W mid (2nd order x-over at 500Hz) and a 50W tweeter (2nd order x-over at 3khz). And a 2,000W/channel amplifier will drive such a system adequately.

In addition to the above, the reason that tweeters are small in size is because of the moving mass issue--a small mass can produce higher frequencies than a large mass can when driven by the same force--and also because of the beaming problem. A large cone will start to beam the sound when the cone size is comparable to the wavelength.

Because the information I read most is OB
But I started from DML
I haven't played passive crossover
But hearing you say that I got a new knowledge
It turns out that the treble can have a small wattage
I'll have a full test after two weeks
Thank you so much for sharing~

 

[Audiofrenzy]

Ewwww, don't buck the Master! The Master has spoken. But I swear, The Master does remind me of the "glass doesn't bend" guy.

???

Here is some UB40 to calm Todd down a bit.

 

[twocents]

No carpet today? I can hear that.

 

[Audiofrenzy]

Its not because of the carpet as the carpet is still there just around 5inches or so out of the videos view. There are 2 things different from this video. First I modified my design to have more highs frequency, the second is I moved each speaker on the sides so that each of my phones mic faces the speakers with the cell phone in the middle so it sounds brighter due to each speaker being close to the mic. Last but not least UB40 track already is very bright to begin with.

Tomorrow I will lessen the brightness with a simple technique. I was going to do it today but my ADD was acting up so I got distracted.

 

[homeswinghome]

I am currently a three-way frequency
Subwoofer + DML +FR
I am not using multiple Exciters

Add the missing part
I use an OB style FR
It is open with "positive waves" + "negative waves"

OK. Seems I missed one of your previous post... DML + FR is not usual in the designs shared here. Have you already described how the frequency ranges are splitted between them? what is their relative efficiency?