The orginal disk also a plastic, so I think it's ok for gluing. Some sanding also help.
My idea is:
1. Just replace the demage disk by the blue line part. Pink area glue into the voice coil.
2. or, Screw mount system. Pink area of the cap glue into voice coil. Blue line part glue to panel.
Attachments
No I don't have xo.
But I did use capacitor to roll off 100hz, and also 200hz. A little help on the knocking but still there. And I don't want to low the volume when listen EDM music. So, maybe just will use them on double exciter system.
Eucy,
I don't ever recall reading that such harmonics are related to the bending wave mechanism, nor seeing such harmonics reported as a "feature" of DML panels. If you happen to recall where you saw that mentioned I'd be interested to read it.
Eric
I haven't tried bracing.
But filling the gaps is where all the fun/engineering/science is!
"Coerce" was indeed a poor choice of words. Because what I am really doing (or at least trying to do) is not to force a panel to do something it does not naturally want to do, but rather, to design the panel system (panel+support+exciter+etc) so that the system naturally wants to do the exact thing that I want it to do.
Eric
Characteristic is probably a better word
I'll look when I get back home.
However, we have a great assortment of panels represented here..we could do some group single tone testing and see where that lands us...differing panels, supports etc .
It would be a far wider sample than that of a research paper. 👍
Eucy
Eric.
I have still yet to see a patent that talks about improving sound quality, instead of filling gaps.
If using a panel full range , you can increase the warmth in the lower midrange by using weights or clamping or foam damping.
All of these options coerc the frequencies (or move them around) filling in gaps.
As long as everyone understands that they will not only be filling a single gap , but altering the whole panel sound.
I have used these techniques on many types of panel, but I find the panels start to sound like ordinary cone speakers, or BMR,which is fine if you want you panel to sound like a standard cone type speaker.
We all have different tastes and objectives , and we are all trying to build our own ideal panels, I presume.
As long as everyone understands that even long strips of foam on a long thin panel will restrain ( coerc) the panel considerably ,affecting the sound quality of the panel as a whole.
Steve.
I have still yet to see a patent that talks about improving sound quality, instead of filling gaps.
If using a panel full range , you can increase the warmth in the lower midrange by using weights or clamping or foam damping.
All of these options coerc the frequencies (or move them around) filling in gaps.
As long as everyone understands that they will not only be filling a single gap , but altering the whole panel sound.
I have used these techniques on many types of panel, but I find the panels start to sound like ordinary cone speakers, or BMR,which is fine if you want you panel to sound like a standard cone type speaker.
We all have different tastes and objectives , and we are all trying to build our own ideal panels, I presume.
As long as everyone understands that even long strips of foam on a long thin panel will restrain ( coerc) the panel considerably ,affecting the sound quality of the panel as a whole.
Steve.
Can do some more tests on my panels, but I have done a bit of single and multi tone testing, as well as looked at harmonic distortion graphs in REW, and have observed the opposite...very low harmonic distortion. So I really don't think it is a feature of DML to introduce harmonics, and you have an issue somewhere in the system if that is what you are seeing.
They can of course amplify harmonics if you have strong modes coinciding, but to it seems perfectly clear that the plate will generate less harmonics than a cone speaker, especially when pushed hard.
Not really possible to assess sound quality as such in a technical way other than analysing response in frequency and time domains. The papers talk about frequency response mostly because that is the major factor when determining the sound quality of a panel. A panel with a 10dB+ hump somewhere in the low mids will just not sound as good as one that doesn't. Maybe on some track in some room if you are lucky, but generally it will be perceived as inferior quality.
When it comes to the time domain things are a bit more complicated, but clamping the plate affects both frequency domain by moving the modes in the plate so that peaks can be spread for an even FR, but also makes the plate ring less, giving a tighter more dampened sound.
IMO it will always be best to aim for an even FR with well distributed peaks. However the ringing "chorusey" sound of unclamped plates is a more interesting effect that I can understand wanting to preserve. It does sound great on some tracks, but gives to much colour IMO for the kind of things I usually play.
Leob.
You seem to take everything to the extremes, a 10db peak in the lower midrange, ouch !
I don't remember anyone saying they suffer from large amounts of distortion?
I only mentioned that I noticed a lower peak in my test, not that it was loud ?
If a panel is suffering from loud ringing ,then it has to be sorted.
If I could not sort the ringing with my usual methods, I would bin the panel.
Steve.
You seem to take everything to the extremes, a 10db peak in the lower midrange, ouch !
I don't remember anyone saying they suffer from large amounts of distortion?
I only mentioned that I noticed a lower peak in my test, not that it was loud ?
If a panel is suffering from loud ringing ,then it has to be sorted.
If I could not sort the ringing with my usual methods, I would bin the panel.
Steve.
I don't think it is extreme to see a 10dB peak somewhere on a DML panel that is not well tuned. 5dB peaks would be an excellent panel, like a Tectonic, but 10dB is what I see in my panels now with no EQ, and I hope to get it down a few dB by more careful tuning.
In my experience all panels I tested rings more or less in that they introduce artifacts visible when looking at an impulse response. They ring much less if they are clamped though. And that makes them sound more like a normal speaker since the added phase distortion caused by the ringing is so low that it is masked.
I haven't talked about loud harmonics, if you have noticeable individual harmonics when playing a sine, you have some issue, and it is not something that is naturally a part of DML. In that case it would be noticeable on my panels as well. .
Leob.
I had a quick flip through old photos of panel plots , and found most of them well within +- 5db in my usable frequency range .
Some reaching down to 200hz or lower depending on the panel materials.
Usually dml panels suffer dropouts, dips, below about 300hz.
Rising again below about 100hz.
Impulse response just shows the dml at work, damping the DML will obviously reduce this.
Steve.
I had a quick flip through old photos of panel plots , and found most of them well within +- 5db in my usable frequency range .
Some reaching down to 200hz or lower depending on the panel materials.
Usually dml panels suffer dropouts, dips, below about 300hz.
Rising again below about 100hz.
Impulse response just shows the dml at work, damping the DML will obviously reduce this.
Steve.
If you are fine with a 300Hz cutoff, then making a DML with flat response is sure a lot easier and should not require much tuning. I don't really see why you would do that though. Without a sub it will be really lacking, and few subs handle 300Hz well. So you would end up needing a 3-way system to get good low mids.
And +/- 5dB is 10dB
As I said that is about double of Tectonic. I have some difficulty reading the graphs you post, and cannot see any scale and resolution seems low. But from the detailed REW graphs posted here and for example the ones in Dayton's data sheets, 10dB difference between peaks and valleys is certainly not extreme.And if it was true that the tail in the impulse response was simply DML at work, reducing it would surely result in less output? I don't see reduced effectivety from clamping, only tighter impulse response.
Hello Leob,If you are fine with a 300Hz cutoff, then making a DML with flat response is sure a lot easier and should not require much tuning. I don't really see why you would do that though. Without a sub it will be really lacking, and few subs handle 300Hz well. So you would end up needing a 3-way system to get good low mids.
And +/- 5dB is 10dB
And if it was true that the tail in the impulse response was simply DML at work, reducing it would surely result in less output? I don't see reduced effectivety from clamping, only tighter impulse response.
- 10dB is it before EQ? I don't remember seeing a Tectonic FR without EQ. I had a quick search in this thread but I only found my most of last February saying the same...
- I am a bit lost in the vocabulary of the previous posts... In my understanding : DML is synonym of panel (the vibrating flat membran). The tail in the impulse response is the signature of strong modes (some resonance frequencies with a low damping). A tighter impulse response means less resonances, less peaks, a smoother FR.
Don't remember where I saw the FR for Tectonic, but you can also look at their recommended EQ which as I recall specifies a cut of a little over 5dB at the most.
The length of the impulse response doesn't necessarily correlate with the frequency content. The ringing is affecting the time domain mostly, so you can have a dampened plate with smooth FR as well as a ringing plate with smooth FR.
I guess it is hard to say what is the actual initial bending wave and what is what I call ringing in an undampened plate, since both is the plate resonating in a way.
As you know some plates with have quite a long tone when you tap them. By damping the plate that tone becomes more like a short knock. The long tone is the plate ringing and is not at all needed to recreate the signal and will not make the panel louder. We care about the actual tap sound being loud, but it should be short for clean reproduction.
As I understand it that graph is the actual filter response, not the FR of the panel. But yes, that shows around 5dB difference, and should mean that the actual raw response of the panel is off by the same amount.
Leob.
This picture which I think I posted before?,is of one of johnnogs proplex 2ft panels, with the exciter just plonked in the centre.
With no eq or damping but with an excellent response plot.
Strangely I can only hear clean sounds with no ringing 🤗
The tectonic link explains why.
Under ,how dml works.
The last computer animation at the bottom clearly shows the impulse in the centre of the panel with the dml actions there after.
Steve.
https://tectonicproaudio.com/about-the-technology/
This picture which I think I posted before?,is of one of johnnogs proplex 2ft panels, with the exciter just plonked in the centre.
With no eq or damping but with an excellent response plot.
Strangely I can only hear clean sounds with no ringing 🤗
The tectonic link explains why.
Under ,how dml works.
The last computer animation at the bottom clearly shows the impulse in the centre of the panel with the dml actions there after.
Steve.
https://tectonicproaudio.com/about-the-technology/
Attachments
Maybe you can get some useful info from that picture as you know the device and can actually read what the axis represents. I have no idea what the scale is, and cannot read it, so it is pretty meaningless to me unfortunately.
I think you are misunderstanding what I mean by ringing. The bending wave propagating in the plate is not what I mean, but the residual vibrations in the plate after that. Think when you throw a something into a bowl of water. That will result in several waves. Sound waves in the water, water waves in the bowl, and the bowl will vibrate.
If the bowl is the DML plate and you are exciting it, not only will you have bending waves in the plate, but the plate will move slightly just like the bowl. You will first hear the bending wave, where the distortion in the time domain is masked just like Tectonic claims. But if the plate is just hanging in a string the plate will keep vibrating after the sound waves inside it's mass has stopped vibrating, and the vibration will cause phase distortion as well as a extended tail of the IR.
You claim you prefer the sound of unclamped plates, so not sure I understand your point. You think clamping only affects FR and doesn't affect the IR?
You think clamping will always result in worse FR?
To me it is pretty clear just touching a free hanging plate you affect more than the FR. Maybe calling it ringing is confusing, but it is clear that the plates sound is affected by the residual waves when unclamped. Whatever it is that is happening, one can hear the result...smearing of the phase (chorus) and a reverb like tail.