first of all I apologize if my english is not good.
if the frame with the woofer does not vibrate, well you have one less problem.
the problem is U22REX which makes the baffle vibrate as you rightly say.
On HifiSentralen owners of LX521 speak of vibrations up to 2mm.
What is needed is to increase the weight of the brachet baffle and everything else attached to the baffle.
a speaker vibrates more or less depending on how it is built and the forces transmitted by the connected speakers.
and here I gave you the example of owner lx521 with aluminum baffle, he says it was the best upgrade to the sound, I believe it but I think he would still improve by making aluminum all the pieces that hold the baffle
an Orion owner made all the aluminum cabinet, I bet his Orion sounded different from Orion built in wood.
his was a radical choice to get safe and get the most out of it, perhaps someone would say excessive, but I have learned that there is always an improvement when the structure is more rigid and heavier.
Now let's get to the rules of vibration mechanics that help design properly.
The construction of a baffle or an entire box
it must be rigid and must have a weight that contrasts the force generated by the moving mass of each speaker,
the optimal weight is obtained by respecting the ratio 1/1000.
U22REX has 19.8gr of moving mass, therefore the weight of the baffle needed would be 19.8kg.
instead the weight of the baffle complete with all the pieces that compose it and including the screwed speakers will be perhaps 6 ~ 7kg?
that's why Linkwiz added the bridge, not only to isolate the baffle from the V frame with the woofers that trembles like an earthquake due to the little weight insufficient to contrast the strength of the 2 membranes.
and in fact lx521magic proposes the realization with harder and heavier wood like panzerholz or aluminum.
the fact that you can or cannot hear the difference when you make a change is a consequence of the conditions in which you observe the sound and resolution of your hi-fi, higher resolution or more accurate reproduction of small nuances that remain hidden in hi-fi with low resolution.
your vibration analysis is not accurate, I also felt small vibrations with my finger, they seemed harmless but when I eliminated them the sound improved.
From my experience I tell you that every little vibration creates distortions or colors to the sound.
how many mistakes and how much sound are you missing
you can't tell until you've eliminated the vibrations.
I have not made any criticism of your work, you are free to experiment, I wanted to give you some tips to solve problems.
I'm also working on something derived from LX521.
for the baffle I also experimented, U22REX vibrates.
when the baffle is small I don't see any other solution than to make a metal baffle, I was thinking of a metal with a higher specific gravity than aluminum.
I have no idea how much it will cost.
I had also thought of marble, I love narmo, but it is difficult to find someone who does a good job, maybe I'm wrong but how do pieces of marble come together?
with glue?
and will it be mechanically stable like metal stopped by bolts?
I don't know I have to investigate ...
valtergio --
OK, your latest comments make more sense, thanks. So we are in agreement -- the 521 baffle is mechanically weak and prone to vibration, not just from the bass bin but also from its own 8" driver.
In my case, the W-frame bass frame is really a non-factor, at least until the vibrations from the 8" driver can be reduced.
I didn't comment on the sound because the 2nd baffle was not just CLD, but also slightly smaller, and bolted to the bass bin -- too many variables introduced to draw conclusions about sound differences between the old baffle and new baffle.
But I can say overall, the sound with the 2nd baffle is a bit cleaner all around. Less sibilance & harshness in the upper ranges, though it wasn't bad with the 1st baffle. It's not possible to pinpoint why, it could be ALL the differences between the baffles.
On my Orions, the 8" mid-driver is mounted to a solid hard maple post, which itself is bolted to the double-thick BB top panel of the bass bin. The driver does not make firm contact with the front baffle. The driver frame barely touches a foam strip around the baffle mounting hole. The whole point was to keep the mid-driver from vibrating the front baffle, and to keep any vibrations from the woofers from affecting the mid-driver. The fact that the mounting post is attached to the woofer box is an obvious flaw... but the LX521's bridge is too clumsy a solution. Also the broad side panels that form the legs of that bridge have to act like sails, picking up airborne vibrations from the very close bass bin anyway. A minimalist metal structure with legs servings the same purpose would work better, but I'm less interested because my metal working skills are nil & it still makes this a 2-part speaker.
Marble has its own resonances & would be ridiculous to shape -- harder than aluminum, I'm sure. And we could also get into an endless debate about mass & stored energy vs rigidity & damping, but let's not.
I mentioned a desire to create a CLD board with aerolam at the core and BB as outer layers. I will try to complete that project this weekend and compare it to the 2nd baffle. It will be light for its size & very rigid. (You might recall that the Celestion SL600/700 used modified aerolam for its enclosure. Celestion SL700 loudspeaker Page 2 | Stereophile.com)
Finally, going back to my version of the AINO-Gradient, I'm simplifying & strengthening the upper driver mounting bracket. The 12" mid driver is solid, no problem there, I'll post photos later, but I'm having trouble getting the NEO8 mounted well. I suspect part of the reason for the excessive sibilance I hear from the system so far is the less-than-rigid mounting of he NEO8. Juha, you used a piece of BB with recessed cutouts for the NEO8/3, then bolted the bottom part to the top of your solid carved wood post. Was it glued as well? What thickness BB? Surely it has some flex?
OK, your latest comments make more sense, thanks. So we are in agreement -- the 521 baffle is mechanically weak and prone to vibration, not just from the bass bin but also from its own 8" driver.
In my case, the W-frame bass frame is really a non-factor, at least until the vibrations from the 8" driver can be reduced.
I didn't comment on the sound because the 2nd baffle was not just CLD, but also slightly smaller, and bolted to the bass bin -- too many variables introduced to draw conclusions about sound differences between the old baffle and new baffle.
But I can say overall, the sound with the 2nd baffle is a bit cleaner all around. Less sibilance & harshness in the upper ranges, though it wasn't bad with the 1st baffle. It's not possible to pinpoint why, it could be ALL the differences between the baffles.
On my Orions, the 8" mid-driver is mounted to a solid hard maple post, which itself is bolted to the double-thick BB top panel of the bass bin. The driver does not make firm contact with the front baffle. The driver frame barely touches a foam strip around the baffle mounting hole. The whole point was to keep the mid-driver from vibrating the front baffle, and to keep any vibrations from the woofers from affecting the mid-driver. The fact that the mounting post is attached to the woofer box is an obvious flaw... but the LX521's bridge is too clumsy a solution. Also the broad side panels that form the legs of that bridge have to act like sails, picking up airborne vibrations from the very close bass bin anyway. A minimalist metal structure with legs servings the same purpose would work better, but I'm less interested because my metal working skills are nil & it still makes this a 2-part speaker.
Marble has its own resonances & would be ridiculous to shape -- harder than aluminum, I'm sure. And we could also get into an endless debate about mass & stored energy vs rigidity & damping, but let's not.
I mentioned a desire to create a CLD board with aerolam at the core and BB as outer layers. I will try to complete that project this weekend and compare it to the 2nd baffle. It will be light for its size & very rigid. (You might recall that the Celestion SL600/700 used modified aerolam for its enclosure. Celestion SL700 loudspeaker Page 2 | Stereophile.com)
Finally, going back to my version of the AINO-Gradient, I'm simplifying & strengthening the upper driver mounting bracket. The 12" mid driver is solid, no problem there, I'll post photos later, but I'm having trouble getting the NEO8 mounted well. I suspect part of the reason for the excessive sibilance I hear from the system so far is the less-than-rigid mounting of he NEO8. Juha, you used a piece of BB with recessed cutouts for the NEO8/3, then bolted the bottom part to the top of your solid carved wood post. Was it glued as well? What thickness BB? Surely it has some flex?
Guys, what do you think is the right/best way to measure the effect of these unwanted vibrations?
Basically any wideband transducer suffers from enormous modulations in upper range, with wideband/music signal. How should we measure and quantify this doppler modulation? Is it distortion, spl variations or what? How do we measure modulation? With dualtone or pink noise/RTA?
Theoretization and fingertips is just speculation!
Doppler Distortion in loudspeakers
Speaker Doppler Distortion | Electronics Notes
Basically any wideband transducer suffers from enormous modulations in upper range, with wideband/music signal. How should we measure and quantify this doppler modulation? Is it distortion, spl variations or what? How do we measure modulation? With dualtone or pink noise/RTA?
Theoretization and fingertips is just speculation!
Doppler Distortion in loudspeakers
Speaker Doppler Distortion | Electronics Notes
OK, but it still has to hang from something -- the frame for the NEO8 in the A-G 1.3 build is the challenge. It needs to be well supported but not impede the back wave of either the NEO8 or the 12" mid.
That first link is a good article, about a well thought out and implemented experiment, but it refers specifically to a kind of IM distortion test in a single loudspeaker driver. His conclusion is that you can hear Phase Modulation Distortion (with 36.8Hz & 4.65kHz signals simultaneously fed to a single 10" driver) but...
However, I'm trying minimize vibrations in the platform upon which a mid/high driver is mounted. I don't think I really need to measure this*; I'm sure the deleterious effect is there, additive to other distortions, and definitely not beneficial. I just want to get it lower than it is right now in my 521 baffles. (*at least not yet.)
Among commercial open baffle speakers that seem to be successful, I can't think of a single one that actually uses a baffle board as little supported as the LX521. Can anyone?
For example, PureAudio's complete perimeter support for their wide (and large) baffle boards seems much more the norm. Kyron Audio's Kronos appears to use a drive mounting arrangement reminiscent of the last versions of the Orion -- bass drivers are mounted via magnet structure to an external metal post, and not attached to the front baffle at all. Mid/high drivers are mounted more conventionally to a baffle separate from the bass baffle.
A friend had multi-unit dipoles, drivers attached from frame or magnet to a heavy steel pole. Dynamic drivers made the pole to walk araound in the room... A free-hanging 12" or 8" driver with Mms of tens of grams must loose much oflow freq efficiency because of inertia loss, with more distortion obviously.
Magneto- or electroplanar units don't have moving mass at all practically, so no problems when hanged with elastic wire/rubberbands.
So we have two problems - doppler/inertia of the unit itself and vibrations conducted and radiated via the frame.
Box speakers have wall resonances, and then large surface area x amplitude make the spl of the radiated "distortion" audible. Drivers with minimal frame have minimal surface area to transduce these vibrations.
When a driver unit has narrow passband, it is not so prone to these interferences. A typical hifi 2-way speaker's woofer has to reproduce 5-6 octaves (30-3000Hz and that's where highest energy and most information is), but a 4-way speaker leaves only 2 octaves per unit! But to be honest, a multiway has other problems with crossovers and interdriver spacing!
I believe and my experience tells, that smoothness of spl response, uniform directivity and low distortion without peaks will tell all this well enough! Frame/box resonances will show up as response- and distortion peaks!
Magneto- or electroplanar units don't have moving mass at all practically, so no problems when hanged with elastic wire/rubberbands.
So we have two problems - doppler/inertia of the unit itself and vibrations conducted and radiated via the frame.
Box speakers have wall resonances, and then large surface area x amplitude make the spl of the radiated "distortion" audible. Drivers with minimal frame have minimal surface area to transduce these vibrations.
When a driver unit has narrow passband, it is not so prone to these interferences. A typical hifi 2-way speaker's woofer has to reproduce 5-6 octaves (30-3000Hz and that's where highest energy and most information is), but a 4-way speaker leaves only 2 octaves per unit! But to be honest, a multiway has other problems with crossovers and interdriver spacing!
I believe and my experience tells, that smoothness of spl response, uniform directivity and low distortion without peaks will tell all this well enough! Frame/box resonances will show up as response- and distortion peaks!
Last edited:
I made good progress on the aerolam CLD baffle board, which ended up becoming very well damped and amazingly rigid... but it also was a total error right from the start. 12mm BB + 10mm aerolam + 12mm BB = 34mm. Unfortunately the depth of the 4" Scanspeak upper mid driver is only 40mm, which means the 78mm diameter hole for that driver is like a tunnel & over-impedes the back radiation. See pics. Why I didn't realize this before I even started can only be attributed to overeager blindness, a fool rushing in. My big excuse: I didn't have suitable thinner BB panels on hand & couldn't be bothered to buy more at the lumber store.
I'll probably give the new baffle a try anyway, mount the 2 mid drivers w/o bothering with the tweeter, and measure/listen to the backward radiation. If it measures/sounds significantly different from the last baffle, this one will have to be shelved & I'll try again with 9mm BB. I've still got 7/8th of the original 4x10' aerolam sheet to mess around with.
I did learn some hard lessons about working with this material. One is that shards of the super-thin aluminum honeycomb slice through skin like razors. Another is that you cannot cut it with any angled router bit; cuts other than perpendicular don't end up well. One has to take great precaution not to breathe the fine fibers of the fiberglass skins that get thrown up when cutting. I was almost hasmat-suited, and a festool dust extractor was constantly on the router (used for all the cutting, mostly with a high speed 1/4" bit meant for use on CNC routers).
It was quite cold when I applied the latex/silicone caulking, which didn't want to spread well at all, but I was surprised after it was dry (with multiple clamps at high pressure) not see any patches (in the edges of the cut driver holes) where the goop dried thicker. It all seems quite consistently thin -- not even visible as a layer, tho admittedly the goop does change color from white to clear as it dries.
My big excuse: I didn't have suitable thinner BB panels on hand & couldn't be bothered to buy more at the lumber store. I'll probably give the new baffle a try anyway, mount the 2 mid drivers w/o bothering with the tweeter, and measure/listen to the backward radiation. If it measures/sounds significantly different from the last baffle, this one will have to be shelved & I'll try again with 9mm BB. I've still got 7/8th of the original 4x10' aerolam sheet to mess around with.
I did learn some hard lessons about working with this material. One is that shards of the super-thin aluminum honeycomb slice through skin like razors.
Another is that you cannot cut it with any angled router bit; cuts other than perpendicular don't end up well. One has to take great precaution not to breathe the fine fibers of the fiberglass skins that get thrown up when cutting. I was almost hasmat-suited, and a festool dust extractor was constantly on the router (used for all the cutting, mostly with a high speed 1/4" bit meant for use on CNC routers).It was quite cold when I applied the latex/silicone caulking, which didn't want to spread well at all, but I was surprised after it was dry (with multiple clamps at high pressure) not see any patches (in the edges of the cut driver holes) where the goop dried thicker. It all seems quite consistently thin -- not even visible as a layer, tho admittedly the goop does change color from white to clear as it dries.
Attachments
Decided to shelf the last aerolam/BB CLD trial. Too much work to just "try" it when I know it's not really going to be ideal.
Round #3 is 6mm BB in front, 10mm aerolam in middle, and 9mm BB in back. Total depth will be just 25mm, which is only 5-7mm deeper than Linkwitz originally specified. This will still provide much greater stiffness (and internal damping) than 25mm BB alone.
The 9mm in back allows a bigger chamfer for rear radiation, especially important for the small 4" driver. I'll cut those chamfered holes in the 9mm BB board before gluing the 3 layers together, because the angled router bit can't go deep enough with the aerolam in place.
Got cut by the aerolam while just moving the piece! Might have to get some good gloves I can wear all the time on this project.
Round #3 is 6mm BB in front, 10mm aerolam in middle, and 9mm BB in back. Total depth will be just 25mm, which is only 5-7mm deeper than Linkwitz originally specified. This will still provide much greater stiffness (and internal damping) than 25mm BB alone.
The 9mm in back allows a bigger chamfer for rear radiation, especially important for the small 4" driver. I'll cut those chamfered holes in the 9mm BB board before gluing the 3 layers together, because the angled router bit can't go deep enough with the aerolam in place.
Got cut by the aerolam while just moving the piece! Might have to get some good gloves I can wear all the time on this project.
Attachments
You could take the baffles to a cnc shop and have them mill-out the rear driver chamfers. Not sure what that would do to the connection for driver & baffle though..
-and yes, a resonant tunnel is bad.
Scroll-down to the Scan Speak Discovery 10F/4424G00 test and particularly the editorial just below it (about that driver):
Zaph|Audio
-and yes, a resonant tunnel is bad.
Scroll-down to the Scan Speak Discovery 10F/4424G00 test and particularly the editorial just below it (about that driver):
Zaph|Audio
Attachments
Last edited:
I have the tools to make at least a 45 degree chamfer in BB, but the aerolam cannot be chamfered at all, cnc or otherwise. The aluminum honeycomb just falls apart if you try, the outer skin needs to be intact on both sides.
With the 6mm BB in front, I can create a 3.5mm recess for the driver frame to be flush with the front baffle, which means a depth of 11.5mm before the 78mm diameter hole expands into the rear chamfer. That's probably not bad. If you cut a deeper/wider chamber, the board strength would be compromised a lot more.
In fact, Zaph's photo looks too flimsy, there's not enough material holding the screws for the driver, imo. Maybe you can get away with this in a normal baffle supported on all edges, but not the unsupported 521 baffle. The whole point of my exercise is to make the baffle less bendy & less resonant, so...
With the 6mm BB in front, I can create a 3.5mm recess for the driver frame to be flush with the front baffle, which means a depth of 11.5mm before the 78mm diameter hole expands into the rear chamfer. That's probably not bad. If you cut a deeper/wider chamber, the board strength would be compromised a lot more.
In fact, Zaph's photo looks too flimsy, there's not enough material holding the screws for the driver, imo. Maybe you can get away with this in a normal baffle supported on all edges, but not the unsupported 521 baffle. The whole point of my exercise is to make the baffle less bendy & less resonant, so...
You could take the baffles to a cnc shop and have them mill-out the rear driver chamfers. Not sure what that would do to the connection for driver & baffle though..
-and yes, a resonant tunnel is bad.
Scroll-down to the Scan Speak Discovery 10F/4424G00 test and particularly the editorial just below it (about that driver):
Zaph|Audio
..consider infill for that area around the driver "injected" into the honeycomb - something you can chamfer.
Better yet, just carve it out larger and put in a coupling ring that's effectively decoupled from the baffle. ..tweeters and small midranges are best decoupled from the baffle.
Better yet, just carve it out larger and put in a coupling ring that's effectively decoupled from the baffle. ..tweeters and small midranges are best decoupled from the baffle.
Last edited:
-and yet there are no measurement artifacts.
(..something about "baby and the bath-water" applies here.)
Don't know about that -- has it been measured? Surely someone, after all these years, has looked closely at the mechanical weakness of that baffle? And then there's the OT LX521 owner who went with an 18mm thick solid aluminum baffle -- linked by valtergio -- and reports:-and yet there are no measurement artifacts.
(..something about "baby and the bath-water" applies here.)
Obviously subjective, no measurements...
There wouldn't be in this condition with such a small driver (even open baffle). It's also sectioned (panel within a panel) with a felt coupling. (..it's a test baffle.)
You can scroll down and look at the extensive measurements. He also notes the modest difference with the Voice Coil measurement of the driver.
Oh there will be differences with different baffles and mounting.
The most significant isn't from panel resonance control/damping (from my experience), but rather just how rigid (and massive) the mounting surface is: just like the subjective comment you've posted. Brass usually providing the best result and braced on it's own rigid "post" decoupled from the rest of the baffle. It's sort of like there is no energy lost relative to the vc moving back and forth. Decoupling just makes sure that the other drivers aren't adding a little bit of extra motion to that particular driver - especially when you've got woofers with lots of mass moving vs. low mass mid's or tiny mass tweeters.
You can scroll down and look at the extensive measurements. He also notes the modest difference with the Voice Coil measurement of the driver.
Oh there will be differences with different baffles and mounting.
The most significant isn't from panel resonance control/damping (from my experience), but rather just how rigid (and massive) the mounting surface is: just like the subjective comment you've posted. Brass usually providing the best result and braced on it's own rigid "post" decoupled from the rest of the baffle. It's sort of like there is no energy lost relative to the vc moving back and forth. Decoupling just makes sure that the other drivers aren't adding a little bit of extra motion to that particular driver - especially when you've got woofers with lots of mass moving vs. low mass mid's or tiny mass tweeters.
Last edited:
Scott, your comments point back to the Linkwitz Orion approach to the 8" driver mounting. In my 521 baffle the 8" is the primary source of large vibration, so mounting it via magnet to a back post and detaching it from the baffle would have the biggest impact. Worth thinking about, tho aesthetically challenging and probably detrimental to the dipole effect.
Backside support/connection of the 8" mid is difficult to make yes. But I don't think it would harm the radiation pattern significantly. Rearside "on-axis" response doesn't have to be smooth, it is more about total energy radiated bacwards that matters.
Another problem is how to fix the magnet to the pole/rest - I use velcro strip which isn't as rigid as I'd like it to be. But I don't want to use epoxy glue either... Extra challenge would come if the magnet had a vent in the middle like many SB Acoustics woofers.
Another problem is how to fix the magnet to the pole/rest - I use velcro strip which isn't as rigid as I'd like it to be. But I don't want to use epoxy glue either... Extra challenge would come if the magnet had a vent in the middle like many SB Acoustics woofers.