It's really simple, the resonators are simply a half wavelength long.
IE: if you have a tweeter with an impedance peak at 1000Hz, then create a series of resonators that are spread out about an octave BELOW the impedance peak and an octave above. The metamaterial is largely there to nullify the impedance peak, similar to what B&W did with it's "Nautilus" speakers.
There's a dude on Facebook who dug up a patent from a speaker from the 1980s that lays out how the metamaterial thing works in layman's terms.
The only real difference between the 40 year old patent and what Kef is doing is that:
1) Kef has a plethora of transmission lines in their metamaterial
2) The folding of the Kef transmission lines is way more complex.
But otherwise? Same idea.
One interesting thing that I realized from tinkering with these, is that they don't seem to have any effect on planar tweeters. That's what led me to speculate that the reason they work so well is because conventional drivers have a big ol' impedance peak.
If I'm right, then compression drivers would be especially good candidates for metamaterials.
Kinda surprised that Celestion hasn't jumped on the bandwagon, considering that Kef is next door and they have the same owners.
But anyways, yeah, give these a try. You can design one in 3D in about 2-4 hours and I can't think of anything that will make a tweeter sound better for two bucks worth of plastic.
IE: if you have a tweeter with an impedance peak at 1000Hz, then create a series of resonators that are spread out about an octave BELOW the impedance peak and an octave above. The metamaterial is largely there to nullify the impedance peak, similar to what B&W did with it's "Nautilus" speakers.
There's a dude on Facebook who dug up a patent from a speaker from the 1980s that lays out how the metamaterial thing works in layman's terms.
The only real difference between the 40 year old patent and what Kef is doing is that:
1) Kef has a plethora of transmission lines in their metamaterial
2) The folding of the Kef transmission lines is way more complex.
But otherwise? Same idea.
One interesting thing that I realized from tinkering with these, is that they don't seem to have any effect on planar tweeters. That's what led me to speculate that the reason they work so well is because conventional drivers have a big ol' impedance peak.
If I'm right, then compression drivers would be especially good candidates for metamaterials.
Kinda surprised that Celestion hasn't jumped on the bandwagon, considering that Kef is next door and they have the same owners.
But anyways, yeah, give these a try. You can design one in 3D in about 2-4 hours and I can't think of anything that will make a tweeter sound better for two bucks worth of plastic.
FYI/FWIW, don't know about all compression drivers, but historically the ones designed around W.E./Lansing, etc., already have ~IB size rear chambers, i.e. the cover is only for protection and keep it from comb filtering with the horn's output. If wanting to increase power handling down around the driver's rated XO point one must add loading caps to reduce its rear chamber Vb.
Very interesting. Could you post the link to that 'dude on Facebook who dug up a patent from a speaker from the 1980s that lays out how the metamaterial thing works in layman's terms'?
When I calculate the 1/2 wavelength at 680Hz (resonant frequency of SB tweeter) I get 10 inches, which doesn't comport with the pics you posted.
When I calculate the 1/2 wavelength at 680Hz (resonant frequency of SB tweeter) I get 10 inches, which doesn't comport with the pics you posted.
The details are scattered across the other thread, but most of the data you need is here:
https://www.diyaudio.com/community/threads/3d-printed-metamaterials.360739/post-6841680
Squeezing a ten inch long path into the metamaterial isn't too hard when you fold it a few times. The metamaterial I made was something like 3" x 3" x 3"
If I made it again I would try and schedule it down quite a bit. This is still a sealed box - albeit a strange one - and a liter of air is way more air than a 1" tweeter requires.
https://www.diyaudio.com/community/threads/3d-printed-metamaterials.360739/post-6841680
Squeezing a ten inch long path into the metamaterial isn't too hard when you fold it a few times. The metamaterial I made was something like 3" x 3" x 3"
If I made it again I would try and schedule it down quite a bit. This is still a sealed box - albeit a strange one - and a liter of air is way more air than a 1" tweeter requires.
None of this is super scientific, but here's some opinions:
1) The main advantage of the acoustic metamaterial seems to be that it flattens out the impedance curve. How and why that reduces distortion is beyond me, but my measurements indicate that distortion is lowered, by quite a bit.
2) Conversely, the metamaterial experiments convinced me that possibly some of the good things that people hear from planars and ribbons may be due to a flat impedance curve. I made a metamaterial enclosure for one of those BG NEO3 clones, and it did absolutely nothing. I believe that was because the impedance curve is flat to begin with.
3) I have not personally built a sealed transmission line, but I have a hunch that it should yield the same results as metamaterial, maybe even better. The main issue with sealed transmission lines is that they tend to be ridiculously large. Completely impractical for a woofer, but perfectly fine for a tweeter.
4) A densely stuffed sealed enclosure won't achieve the same results as a sealed transmission line. It will reduce the impedance curve to an extent, but it won't nuke it entirely like a sealed TL can
For me, possibly the biggest takeaway from this entire project is that there are probably a lot of good reasons to run without a back chamber at all. Obviously, dipoles and cardioids have a different radiation pattern. But I also wonder if the lack of reflection from the inside of the box has some benefits that are difficult to measure.
I think it's very difficult to detect whether that's an issue, because if the reflection WAS an issue, the reflection will get combined with the initial signal in way less than a single millisecond. For instance, with the SB26 tweeter, that reflection will happen in less than one tenth of a millisecond.
To complicate things even further, I definitely wonder if reflections inside of the back chamber are minimized by making the chamber so small that the reflection occurs in a tiny fraction of a millisecond. Ring radiators are a good example of this; the back chamber is absolutely TINY and if that backwave reflection is a problem, it might be less of a problem if it happens so quickly that it's nearly indistinguishable from the front wave. The back chamber in an Eminince N151M ring radiator is so small, the depth of the chamber is less than a centimeter deep. Sound travels 34 centimeters in a single millisecond, so a reflection in a back chamber that's that small will take less than a tenth of a millisecond. That might be so fast that it's inaudible and perhaps that has something to do with the perceived "clarity" of ring radiators.
1) The main advantage of the acoustic metamaterial seems to be that it flattens out the impedance curve. How and why that reduces distortion is beyond me, but my measurements indicate that distortion is lowered, by quite a bit.
2) Conversely, the metamaterial experiments convinced me that possibly some of the good things that people hear from planars and ribbons may be due to a flat impedance curve. I made a metamaterial enclosure for one of those BG NEO3 clones, and it did absolutely nothing. I believe that was because the impedance curve is flat to begin with.
3) I have not personally built a sealed transmission line, but I have a hunch that it should yield the same results as metamaterial, maybe even better. The main issue with sealed transmission lines is that they tend to be ridiculously large. Completely impractical for a woofer, but perfectly fine for a tweeter.
4) A densely stuffed sealed enclosure won't achieve the same results as a sealed transmission line. It will reduce the impedance curve to an extent, but it won't nuke it entirely like a sealed TL can
For me, possibly the biggest takeaway from this entire project is that there are probably a lot of good reasons to run without a back chamber at all. Obviously, dipoles and cardioids have a different radiation pattern. But I also wonder if the lack of reflection from the inside of the box has some benefits that are difficult to measure.
I think it's very difficult to detect whether that's an issue, because if the reflection WAS an issue, the reflection will get combined with the initial signal in way less than a single millisecond. For instance, with the SB26 tweeter, that reflection will happen in less than one tenth of a millisecond.
To complicate things even further, I definitely wonder if reflections inside of the back chamber are minimized by making the chamber so small that the reflection occurs in a tiny fraction of a millisecond. Ring radiators are a good example of this; the back chamber is absolutely TINY and if that backwave reflection is a problem, it might be less of a problem if it happens so quickly that it's nearly indistinguishable from the front wave. The back chamber in an Eminince N151M ring radiator is so small, the depth of the chamber is less than a centimeter deep. Sound travels 34 centimeters in a single millisecond, so a reflection in a back chamber that's that small will take less than a tenth of a millisecond. That might be so fast that it's inaudible and perhaps that has something to do with the perceived "clarity" of ring radiators.