What diffraction simulation software do you recommend?
I'm still gone chamfer the sides of the braces. It will be 1" thick by the way. Even though it might be tiny, I still believe it will make a difference. Scan-Speak Illuminator drivers have aerodynamic motor design for the same reason. If not, i the idea will still definately make me feel better.
I use Knauf Acoustifit on the insides. The surface of the baffle is lined in 3 mm felt. I haven't really experimented with what's most effective for in the midrange chamber, also because it's hard to find info on what kind of insulation is most effective for what frequency range. And I honestly don't want to experiment myself much because it all takes lots of money, and time. I will try to when I'm finishing this next pair. But I want to have a clear idea of what will probably be most effective. Because I don't want to waste lots of money on materials I'll end up not using.
I can change the "q" of the crossover slopes but it doesn't seem to make the slopes steeper, it just makes the knee sharper and eventually point up. I think I posted pictuers previously in this thread. When I get to measure the speaker I'll try different q's and see if it affects the actual acoustic response.
I haven't made measurments again yet, because I find it to be too hard in my room. I'm planning on taking the speakers to work (high end store) again and measure them there. We have a large listening space, I'll definately be able to make good measurments there. And my colleagues also want to hear the speakers with the new tweeters.
I'm still gone chamfer the sides of the braces. It will be 1" thick by the way. Even though it might be tiny, I still believe it will make a difference. Scan-Speak Illuminator drivers have aerodynamic motor design for the same reason. If not, i the idea will still definately make me feel better.
I use Knauf Acoustifit on the insides. The surface of the baffle is lined in 3 mm felt. I haven't really experimented with what's most effective for in the midrange chamber, also because it's hard to find info on what kind of insulation is most effective for what frequency range. And I honestly don't want to experiment myself much because it all takes lots of money, and time. I will try to when I'm finishing this next pair. But I want to have a clear idea of what will probably be most effective. Because I don't want to waste lots of money on materials I'll end up not using.
I can change the "q" of the crossover slopes but it doesn't seem to make the slopes steeper, it just makes the knee sharper and eventually point up. I think I posted pictuers previously in this thread. When I get to measure the speaker I'll try different q's and see if it affects the actual acoustic response.
I haven't made measurments again yet, because I find it to be too hard in my room. I'm planning on taking the speakers to work (high end store) again and measure them there. We have a large listening space, I'll definately be able to make good measurments there. And my colleagues also want to hear the speakers with the new tweeters.
The good thing about your problem frequency range is it can be gated, so you don't need a special room.
My suggestion of dropping the woofer response from the equation might be valid. Once the ear adjusts to a lack of bass, I wonder whether we can differentiate harsh midrange / treble vs. pleasant recordings when the woofer was playing... and identify the same pattern without the woofer. the without woofer is to eliminate some sort of woofer related distortion problem but also take out some room effects allowing listening outside to be compared (which has no room effects). Basically harsh everywhere = speaker issue.
HD sweeps might say one thing... but I wonder if targeted frequency IMD measurements would be useful. To identify the frequencies of interest, I would use a manually adjusted sine wave "sweep" to find where it "jumps out". Using an SQL meter should confirm what your ears are telling you. If all frequencies have the same SPL measurement- then moving onto IMD measurements might be helpful. With the suspect frequencie(s) identified, select products and factors based on the 2nd and 3rd harmonic.
All the above might seem a lot of effort, but after all, when you do it, do it right. Go all out
My suggestion of dropping the woofer response from the equation might be valid. Once the ear adjusts to a lack of bass, I wonder whether we can differentiate harsh midrange / treble vs. pleasant recordings when the woofer was playing... and identify the same pattern without the woofer. the without woofer is to eliminate some sort of woofer related distortion problem but also take out some room effects allowing listening outside to be compared (which has no room effects). Basically harsh everywhere = speaker issue.
HD sweeps might say one thing... but I wonder if targeted frequency IMD measurements would be useful. To identify the frequencies of interest, I would use a manually adjusted sine wave "sweep" to find where it "jumps out". Using an SQL meter should confirm what your ears are telling you. If all frequencies have the same SPL measurement- then moving onto IMD measurements might be helpful. With the suspect frequencie(s) identified, select products and factors based on the 2nd and 3rd harmonic.
All the above might seem a lot of effort, but after all, when you do it, do it right. Go all out
It does not matter much if on-axis simulation at low...mid frequencies is needed. VituixCAD Diffraction supports off-axis up to 180 deg and has off-axis export sequence so it should be obvious selection for a bit more advanced design method.
BUT enclosure should be designed so that simple diffraction simulation is not needed for the tweeter. Wavefront simulation could be more usable though knowledge and common sense should be adequate to avoid directivity problems which cannot be handled with crossover design. Rectangular box with sharp edges and "optimized" tweeter position just don't work.
If you run 12 db (recommended in active, 24 db will be even harder) you should pick 0.5 and activate invert on the mids. I think 12db is 0.5...
Before you build anything new, to save you a huge amount of time, you should put some time into doing you own experiments.
First lesson, that also took me some time to get to understand, is the difference between textbook filters and accoustic response. In practice very simple: In order to get ie. a perfect LR12 filter, one need to understand this:
LR12 in DSP and LR12 accoustic slope is only the same in a perfect world, that does not exist ie. if the driver is perfectly flat.
In order to get an actual ie. LR12 filter you do this:
Get the texbook filter and import them into for instance REW, using this for reference. (Textbook filter are easy to make in Rephase).
Then overlay the actual response of the ie. tweeter. Is the tweeter response the same? It should. Here it gets complicated, because you have to iron out room reflections (if you have 1 meter too any reflection surface you can measure tweeter in that room). You have to take into consideration diffration, what the tweeter can handle and so.
From this, you can mingle with DSP settings.
I have immense respect for all the stuff you can do in VituixCAD, but I think you should start with learning the basic first.
First lesson, that also took me some time to get to understand, is the difference between textbook filters and accoustic response. In practice very simple: In order to get ie. a perfect LR12 filter, one need to understand this:
LR12 in DSP and LR12 accoustic slope is only the same in a perfect world, that does not exist ie. if the driver is perfectly flat.
In order to get an actual ie. LR12 filter you do this:
Get the texbook filter and import them into for instance REW, using this for reference. (Textbook filter are easy to make in Rephase).
Then overlay the actual response of the ie. tweeter. Is the tweeter response the same? It should. Here it gets complicated, because you have to iron out room reflections (if you have 1 meter too any reflection surface you can measure tweeter in that room). You have to take into consideration diffration, what the tweeter can handle and so.
From this, you can mingle with DSP settings.
I have immense respect for all the stuff you can do in VituixCAD, but I think you should start with learning the basic first.
I have done some experiments with felt on the front baffle. To my big surprise, it was NOT worth the effort. It takes som some time covering the whole baffle, with tweeter perfectly flush to the felt.
One could argue, that if some high-end speakers do this, it might be good right? In my measurement and listening experiments, it did not improve anything. Keep in my this was with a satori tweeter, with a trapez baffle that was measering great from the beginning.
One could argue, that if some high-end speakers do this, it might be good right? In my measurement and listening experiments, it did not improve anything. Keep in my this was with a satori tweeter, with a trapez baffle that was measering great from the beginning.
I had a Ti Audax dome tweeter. To make it sound the best; I used a semi-circle of felt above the dome. A full circle made it too dull and no felt made it too bright. I suppose it varies greatly from tweeter to tweeter and baffle size, shape, etc. I read that pure (100%) lamb's wool was the best for this. I also read that some of the ready made tweeter diffraction rings were synthetic and they could actually make the sound worse rather than help.
Felt Rings verses Felt Blocks for Baffle Diffraction Reduction
Interesting experiments in the above link.
Felt Rings verses Felt Blocks for Baffle Diffraction Reduction
Interesting experiments in the above link.
Internal damping of midrange. Thick felt on all internal wall is a good starting point. The knauf stuff you have, I believe is great. In regards to filling, I have found no better than fluffy sheepwool. Troels Graveson uses this in all his newer constructions. I think Jantzen sell this, albeit the amount you get, is enough for a whole life of experiments. Amount of filling: This needs experiments.
In regards to front baffle shape, I have found that trapez-like shape, think Avalon speaker style, is the way to go. Keep in mind that most of those speakers have their tweeter positioned quite close to the top. Yes, making huge roundovers is better, but not easy to achieve without CNC machine. This software is easy to start with, getting a basic visualization of what is going on: Tolvan Data
From there, the by far best way to experiment is practice, is arming yourself with dens cardboard, stanley knife and painters tape. From there you can make many iterations in a short amount of time.
If not convinced about this simple ways of using materials, I can say that in the construction of Belab 90, they used blu tack, in order to perfect the shape of the front baffle.
Why not just make a new baffle of you current speaker?
In regards to front baffle shape, I have found that trapez-like shape, think Avalon speaker style, is the way to go. Keep in mind that most of those speakers have their tweeter positioned quite close to the top. Yes, making huge roundovers is better, but not easy to achieve without CNC machine. This software is easy to start with, getting a basic visualization of what is going on: Tolvan Data
From there, the by far best way to experiment is practice, is arming yourself with dens cardboard, stanley knife and painters tape. From there you can make many iterations in a short amount of time.
If not convinced about this simple ways of using materials, I can say that in the construction of Belab 90, they used blu tack, in order to perfect the shape of the front baffle.
Why not just make a new baffle of you current speaker?
If you have V. Dickason's LDC; look in the chapter about shapes of enclosures and damping. These were originally done by Harry Olson MANY years ago! For my upper and lower mids; I have the shape "J". Almost zero baffle face. I got the idea from the original B&W 801 mids enclosure. Shape "L" and "J" seem to give the flattest FR with the least disturbances (except the sphere). Shape "L" is like the Avalon trapezoid shape. My mid enclosures have no parallel internal surfaces. I use foam pyramids directly behind the driver magnet; then poly' stuffing behind that. I need to add felt damping to the internal surfaces but we are not dealing in the bass octaves here so the foam and poly' do quite well even without adding the felt.
I read that some say the newest Satori Be dome with the large wave guide sounds as good if not better than a RAAL ribbon. Of course, for $400+ US it BETTER sound good!
Satori TW29BNWG-4 Beryllium Dome Tweeter with Waveguide
Satori TW29BNWG-4 Beryllium Dome Tweeter with Waveguide
I red the article about the felt. Hmm, I might actually get rid of the felt then.
I will do my best with optimising the damping inside the midrange chamber. I will probably do it with real sheep wool and will also try thick felt on the walls if I can find something of good thickness (3 mm probably won't do the trick). I found a website of some sheep farm where they sell real sheep wool for real cheap here in the Netherlands.
Great to hear that the waveguide version of the Satori performs so well. Won't be getting it though. Because it doesn't physically fit in my design and because I just got my new BliesMa's. I have a similar experience with them. The BliesMa's are so freaking fast and have such low distortion that they sound like ribbons in terms of detail retrieval. I recently heared the new Veddan Origin system which uses a very special 270⁰ magnetostat. Becausr of it's huge surface area yet practically non existent mass it retrieved literally every detail from the recording and can play over 100 dB without any audible distortion at all. My experience with the BliesMa is similar. Detail like I've never heared from a dome or something even close and I have also played around 100 dB with extremely little distortion. It was rather unpleasent because of the faulty frequency response than from distortion. Very, very impressive.
I will do my best with optimising the damping inside the midrange chamber. I will probably do it with real sheep wool and will also try thick felt on the walls if I can find something of good thickness (3 mm probably won't do the trick). I found a website of some sheep farm where they sell real sheep wool for real cheap here in the Netherlands.
Great to hear that the waveguide version of the Satori performs so well. Won't be getting it though. Because it doesn't physically fit in my design and because I just got my new BliesMa's. I have a similar experience with them. The BliesMa's are so freaking fast and have such low distortion that they sound like ribbons in terms of detail retrieval. I recently heared the new Veddan Origin system which uses a very special 270⁰ magnetostat. Becausr of it's huge surface area yet practically non existent mass it retrieved literally every detail from the recording and can play over 100 dB without any audible distortion at all. My experience with the BliesMa is similar. Detail like I've never heared from a dome or something even close and I have also played around 100 dB with extremely little distortion. It was rather unpleasent because of the faulty frequency response than from distortion. Very, very impressive.
My kind of tweeter for sure; out of my price range though for now anyway. Maybe you might try some cork on the outside? I often use cork. 100% lamb's wool is said to be the best for inside; many people and papers say so anyway. My main system has separate enclosures for each driver pair; the upper and lower mids have almost zero baffle face (exposed area) and no internal surfaces are parallel to minimize diffraction in front and minimize stand waves inside...Even if my woofer boxes have parallel sides; I try to break it up with internal surfaces that are not parallel, I have even used styrofoam cones and other odd shapes inside to reduce standing waves; angled support bracing, etc. The foam pyramids help as well to help keep back wave energy from returning through the cone...
Most people say Be is the one of the very best dome tweeter materials...I have had a few ribbons which I always liked but vertical directivity is a common issue. I would like to try a good quality AMT soon. My Fostex ST cost over $200 US each. I set my limit on pricing per driver about there. Some day I may go for more serious high end but not anytime in the near term future...
I ordered 1 kg of lambs wool from the farm. Seems like a pretty big bag. I ordered the type that supposedly has smaller fibres which I think would be the better type for sound absorbtion. The real deal, top quality for only €18 including shipping.
Genuine wool is more effective for on the baffle so I looked for real wool felt for on the baffle. It's too expensive though. One piece of 5 mm 100 x 183 cm black wool would cost €235,-.
Genuine wool is more effective for on the baffle so I looked for real wool felt for on the baffle. It's too expensive though. One piece of 5 mm 100 x 183 cm black wool would cost €235,-.
Last edited:
And about damping of the enclosure itself. I wan't to dampen the walls of the midrange chamber. I used butyl matts on my current pair. And this might not be a very scientific way to test but when I gave the panels a simple knock I couldn't hear a difference between with and without the butyl. So I really doubt if it's any effective.
What I was thinking about doing is having a heavy mass stuck to the walls with a soft rubbery material inbetween. I was thinking about 3 mm plates of lead with a rubbery foam adhesive, I'd still have to find what exactly. I think this would be a far more effective way. Appart from it's heavy mass the lead itself is very well damped as well. PMC does the same thing on their Fenestria speaker with the side walls.
What I was thinking about doing is having a heavy mass stuck to the walls with a soft rubbery material inbetween. I was thinking about 3 mm plates of lead with a rubbery foam adhesive, I'd still have to find what exactly. I think this would be a far more effective way. Appart from it's heavy mass the lead itself is very well damped as well. PMC does the same thing on their Fenestria speaker with the side walls.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.