Have you seen this?
www.somasonus.net
Box Construction Methods | somasonus
www.somasonus.net
-
WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!
Speaker enclosure vibrations - a few measurements with accelerometer
- Thread starter Thomas_A
- Start date
OP
- Thread Starter
- #22
Yes, I've looked it through once regarding ports. I have not looked through the CLD tests though. I did my CLD cabinets 2005, and at that time I just went with recommendations how to do it from another forum and that was to build the speaker using MDF boards of the same thickness glued together with 1 mm of DG-A2 glue. It was a bit tricky and the glue was really a mess to work with, but at least the results were satisfactory.Have you seen this?
Box Construction Methods | somasonus
Very interesting.
As we can see in your measurements, the constrained layer cabinet has higher distortion from the cabinet at a lower frequency , at 250 Hz.
The cabinet without constrained layer has its highest cabinet resonance and distortion at 500 Hz .
Regarding lowering the resonance frequency of the cabinet :
This is in my opinion not entirely good - in my opinion its better to have more stiff materials like aluminium or birch plywood, putting the cabinett resonances up in frequency where they are less audible and more easy to control. If its done right.
Genelec has made an outstanding cabinet with the loudpeaker G3, the same box as 8030c. The strongest resonance is at 637 Hz from the cabinet. This is a better result than the class leading Kef ls50 Meta.
Genelec G Three active loudspeaker
My review samples of Genelec's G Three powered loudspeaker came with a little hand-sized green and tan cardboard card featuring a poem in bold black letters dated 1898: At the cottage window a little bird sang. And the light of the window did flicker. And look. The roof up it sprang and the...
www.stereophile.com
”I investigated the enclosure's vibrational behavior with a plastic-tape accelerometer. It was extremely inert. The only resonant mode I found was on the sidewall, at 637Hz (fig.1), but this is vanishingly low in level, even at SPLs >90dB.”
Last edited:
Nice thread going on. I have researched quite a bit on cld so want to chime in. From physics we know that at resonance frequency vibration can be damped by damping, higher than resonance by mass, and lower than resonance by stiffness. We probably should reconise where the vibration happens compare to resonance frequency to target it. When we deal with a part to lower it, other part looks like raised in comparison, so we will need to know how good we are at too.
The principle how cld work is using shear between layers. If you bend a book, the pages will shear past each other, and cld work against it, turn the shear movement into heat, so the shear will not bounce back as much. The material used should have "damping", which means turn movement into heat, and the material is called viscoelastic material. There is a graph showing the hysteresis of velocity or something (I forgot), the area inside of hysteresis is the amount of damping.
Not every material that is soft will have viscoelasticity. Silicon does not, so it will act like spring, keep on bouncing. Some viscoelastic material are green glue related product and acrylic tape from 3m.
Fr the research I have see, they suggest the two outer layer of cld should have very high stiffness. The viscoelastic material don't have to be thick, 1mm is probably already very thick.
My point of view on this is that it depends on everything from stiffness of outer layer to hysteresis shape or stiffness of viscoelastic material. I really don't know if all their hysteresis shape will look the same and how stiffness affect the damping. If the stiffness of the outer material is high, the resonance frequency is high, I guess maybe we need higher stiffness for our viscoelastic material or hysteresis shape suit for the stiff outer layer.
The principle how cld work is using shear between layers. If you bend a book, the pages will shear past each other, and cld work against it, turn the shear movement into heat, so the shear will not bounce back as much. The material used should have "damping", which means turn movement into heat, and the material is called viscoelastic material. There is a graph showing the hysteresis of velocity or something (I forgot), the area inside of hysteresis is the amount of damping.
Not every material that is soft will have viscoelasticity. Silicon does not, so it will act like spring, keep on bouncing. Some viscoelastic material are green glue related product and acrylic tape from 3m.
Fr the research I have see, they suggest the two outer layer of cld should have very high stiffness. The viscoelastic material don't have to be thick, 1mm is probably already very thick.
My point of view on this is that it depends on everything from stiffness of outer layer to hysteresis shape or stiffness of viscoelastic material. I really don't know if all their hysteresis shape will look the same and how stiffness affect the damping. If the stiffness of the outer material is high, the resonance frequency is high, I guess maybe we need higher stiffness for our viscoelastic material or hysteresis shape suit for the stiff outer layer.
OP
- Thread Starter
- #25
To be fair, I have not seen any harmonic distortion results from accelerometer measurements from Stereophile, only the primary frequency as you show above. So I don't know what these say. It is more complicated since an enclosure which is glued together should be "linear". There is something else giving nonlinearities, and the suspicion is the driver to enclosure interface. One solution is to glue the driver to the enclosure, but I do not want to experiment with that. At least not at the moment...Very interesting.
As we can see in your measurements, the constrained layer cabinet has higher distortion from the cabinet at a lower frequency , at 250 Hz.
The cabinet without constrained layer has its highest cabinet resonance and distortion at 500 Hz .
Regarding lowering the resonance frequency of the cabinet :
This is in my opinion not entirely good - in my opinion its better to have more stiff materials like aluminium or birch plywood, putting the cabinett resonances up in frequency where they are less audible and more easy to control. If its done right.
Genelec has made an outstanding cabinet with the loudpeaker G3, the same box as 8030c. The strongest resonance is at 637 Hz from the cabinet. This is a better result than the class leading Kef ls50 Meta.
Genelec G Three active loudspeaker
My review samples of Genelec's G Three powered loudspeaker came with a little hand-sized green and tan cardboard card featuring a poem in bold black letters dated 1898: At the cottage window a little bird sang. And the light of the window did flicker. And look. The roof up it sprang and the...
”I investigated the enclosure's vibrational behavior with a plastic-tape accelerometer. It was extremely inert. The only resonant mode I found was on the sidewall, at 637Hz (fig.1), but this is vanishingly low in level, even at SPLs >90dB.”
View attachment 222835
Last edited:
I think ( without knowing everything ) that vibration attenuation treatment must and can be done in different ways depending on the frequency the loudspeaker will play . One seems not to be helped using constraining layer constructions in subwoofers , where ordinary stiff 2 cm birch plywood without bracing seems to be good enough up to 150 Hz ( 48 dB/ oct crossover and cube formed, 40*40 cm ).
So, then you have 150-20000 Hz where ( maybe ) a clue to low cabinet vibration comes from B/W Matrix structure , raising all resonanses above 600 Hz making them higher Q resonances and therefore maybe less audible. Or using very stiff material.
BBC had , in the 1977 report, another interesting approach above 250 Hz - it was using only 10 mm plywood and a soft bitumen material with equal thickness.
Read more here, page 10 :
Last edited:
OP
- Thread Starter
- #27
Agree, subwoofer enclosures are not a real problem below 100 Hz, unless there are special cases with non-linearities. With respect to frequency, I think distortion it more audible in the upper voice range than in the lower. So it all depends.
When you look at measured distortion, there are several examples of speakers where you see the usual high distortion in the lowest bass followed with a new peak in the midrange. Genelec is no exception to this, although it may be restricted to 2nd order distortion, e.g.
Genelec 8361A Review (Powered Monitor)
This is a review and detailed measurements of the Genelc 8361A studio monitor (powered active speaker). It is on kind loan from a member and costs US $4,995. The 8361A looks identical to the rest of the series except massively scaled up: It doesn't look it but it is extremely heavy. It has a...
www.audiosciencereview.com
The measured issues with boxes and drivers may also in part be related also to what is found in this thread:
The REAL Problem of March Audio's Sointuva WG (Review, Measurements and Reinforcements with Klippel device)
I am not good at English. Therefore, this post was written with the translation help of many people in Korea. I am very grateful to them. I am an audiophile publishing speaker reviews to audio communities. Currently, I am using Klippel Distortion Analyzer 2, Earthworks M30 mic, and a DIYed...
www.audiosciencereview.com
Last edited:
- Joined
- Mar 20, 2020
- Messages
- 168
- Likes
- 1,089
Genelec has made an outstanding cabinet with the loudpeaker G3, the same box as 8030c. The strongest resonance is at 637 Hz from the cabinet. This is a better result than the class leading Kef ls50 Meta.
”I investigated the enclosure's vibrational behavior with a plastic-tape accelerometer. It was extremely inert. The only resonant mode I found was on the sidewall, at 637Hz (fig.1), but this is vanishingly low in level, even at SPLs >90dB.”Genelec G Three active loudspeaker
My review samples of Genelec's G Three powered loudspeaker came with a little hand-sized green and tan cardboard card featuring a poem in bold black letters dated 1898: At the cottage window a little bird sang. And the light of the window did flicker. And look. The roof up it sprang and the...
You can find an article I wrote on measuring cabinet vibrational behavior like this at https://www.stereophile.com/features/806/index.html
John Atkinson
Technical Editor, Stereophile
Oh ! Thanks !
The use of these devices here and in your article leave me wondering why the manufacturers advice on mounting isn't being followed.
From reading the data sheet of the PVDF sensor, it calls for the device, when used as a vibration sensor, to be mounted by the contact legs. I see this as the sensor standing perpendicular to the surface being measured ... allowing the sensor to vibrate freely above in open air ... not taped flat to the surface being measured for vibration.
Here, Thomas_A, you've just double face taped the device you're using to the enclosure walls and left the connector lead free to interact with the sensor. The data sheet for your device requires secure attachment of both the sensor and the the connecting wire lead for a good measurement.
It's nice to see measured effects of enclosure vibration using both these types of sensors but how accurate is the data being presented?
Last edited:
dasdoing
Major Contributor
it's actualy obvious if you think about it. hard coupling want absorb anything, the soft feet will "mute" the surface like a pad on a snaredrum
OP
- Thread Starter
- #32
You can look at the initial measurements made the day before and judge for yourself regarding repeatability. Below the harmonic distortion from the two enclosures. (The action-reaction force measurements with soft and hard feet are not comparable between days, as I wrote there were difference in the setup (more stable metal support for the speaker and plugged port).)
Springs Under My Speakers: What's Happening?
Kinda sorta maybe related…mounting my fullrange like this (by magnet) so that it was independently mounted from its baffle (through a hole) significantly increased high volume clarity on my PAP open baffles. So I do suspect changing the vibration environment can make significant improvements...
www.audiosciencereview.com
As I wrote in the beginning, double-adhesive tape is not recommended for accurate readings, especially for high frequencies.I did testing for the lead taped and not, did not make any difference here. However, the readings follows well what should be expected. If double-sided tape should be used, it must be thin. Which it is here.
Last edited:
OP
- Thread Starter
- #33
There was a lengthy discussion about this matter here:
Speaker spikes, pads etc. couple or de-couple your speakers
Speaker spikes. pads etc couple or de-couple your speakers Last week i bought a Vandersteen Model 1B speakers an they sound great. Love their musicality an staging. When i bought them beneath each corner felt pads where placed (plastic surrounding with felt in it an a nail through). The ones...
www.audiosciencereview.com
I don't know when the "speaker spike" story started, but perhaps end of 1970s or beginning of 1980s. Neither physics or measurements have indicated any benefits using spikes when it relates to make the speaker move less during playing. Instead the speaker resonates more, often somewhere in the bass region.
OP
- Thread Starter
- #34
Double-sided tape for accelerometers, a note from MSI here:
Have you tried a nearfield measure on your port? Can be quite revealing in my experience.
Or perhaps some nearfield mic measurments of cabinet walls?
Or perhaps some nearfield mic measurments of cabinet walls?
Some constructors ( i.e I.Ö ) says it might be an advantage to use different thickness of MDF in the loudspeaker cabinet. If the baffle are made of 22 mm MDF , then it should be better to make the sides of 16 mm and the back of the speaker 19 mm . This avoids the ”tuning fork effect”.
I have no practical measurement experience of this, but I have tried 3 mm bitumen in some DIY speakers before, also in the jbl 530 , HYBRID and monitor audio rx6, and the best sound results ( better than without bitumen ) was gained with bitumen glued on about 50 % of the loudspeakers internal wall, placed asymetrical , so that the left wall was damped with bitumen and the opposite wall was undamped with no bitumen. This was done at first on only one speaker, so I could compare the sound with the other loudspeaker cabinet without bitumen.
Putting bitumen on all internal walls made the sound much worse in all three loudspeakers. This was a hard lesson, and took some hours to get rid of some of the bitumen pads.
My theory of this experience is that its a good thing for sound, If one can spread the cabinet resonances to different frequencies . This can probably be done using different thickness in the walls of the box, or using bitumen glued on some of the walls.
Using constrained layer damping would then only be beneficial If used on some of the walls, not all of them.
Another approach would be an optimal aligment pressure on driver montage to the baffle and make this of steel or aluminium, like its done in Genelec G3. But this is hard to do DIY.
Last edited:
Linkwitz have done a lot of measurements :
”There are several ways to reduce modal panel vibrations. Because the vibration energy from the driver decreases rapidly with increasing frequency it is advantageous to push the panel vibration modes up in frequency where the excitation energy is small. This is best accomplished by increasing the panel stiffness, but often goes together with increasing the mechanical Q of resonance. Dampening the panel by using a constrained layer that dissipates energy will reduce Q. Panel stiffness is also obtained by extensive bracing. As my rule of thumb, no un-braced box panel area should be larger than 4 inch squared for 3/4 inch thick wood panels. That is a lot of bracing, but it pushes modes into the low kHz range. ”
————
Linkwitz seems to come to the conclusion that its better to push the panel vibration modes up in frequency - i.e. by increasing the panel stiffness.
————
More from Linkwitz:
A) Drivers with a stamped metal baskets are prone to exhibit a high Q resonance when tightly clamped to the baffle. The magnet moves relative to the voice coil at the resonance frequency. Energy is stored and also readily transmitted from the moving mass of the cone into the cabinet.
B) Soft mounting the driver basket to the baffle using rubber grommets reduces the resonance frequency. A 2nd order lowpass filter is formed that reduces the transmission of vibration energy from the moving cone to the baffle and cabinet. The resonance must occur below the operating range of the driver.
C) If the driver is mounted from the magnet and the basket rim touches the baffle only softly, then the magnet-basket resonance cannot occur and the transmission of vibration energy into the baffle is minimized.
—————-
Looking at those picture - Just a thought… The Genelec G3 mounting of the driver is done by clamping the magnet from behind to the baffle wall. Maybe this is a better way of mounting drivers than using screws at the front panel ?
”There are several ways to reduce modal panel vibrations. Because the vibration energy from the driver decreases rapidly with increasing frequency it is advantageous to push the panel vibration modes up in frequency where the excitation energy is small. This is best accomplished by increasing the panel stiffness, but often goes together with increasing the mechanical Q of resonance. Dampening the panel by using a constrained layer that dissipates energy will reduce Q. Panel stiffness is also obtained by extensive bracing. As my rule of thumb, no un-braced box panel area should be larger than 4 inch squared for 3/4 inch thick wood panels. That is a lot of bracing, but it pushes modes into the low kHz range. ”
————
Linkwitz seems to come to the conclusion that its better to push the panel vibration modes up in frequency - i.e. by increasing the panel stiffness.
————
More from Linkwitz:
A) Drivers with a stamped metal baskets are prone to exhibit a high Q resonance when tightly clamped to the baffle. The magnet moves relative to the voice coil at the resonance frequency. Energy is stored and also readily transmitted from the moving mass of the cone into the cabinet.
B) Soft mounting the driver basket to the baffle using rubber grommets reduces the resonance frequency. A 2nd order lowpass filter is formed that reduces the transmission of vibration energy from the moving cone to the baffle and cabinet. The resonance must occur below the operating range of the driver.
C) If the driver is mounted from the magnet and the basket rim touches the baffle only softly, then the magnet-basket resonance cannot occur and the transmission of vibration energy into the baffle is minimized.
—————-
Looking at those picture - Just a thought… The Genelec G3 mounting of the driver is done by clamping the magnet from behind to the baffle wall. Maybe this is a better way of mounting drivers than using screws at the front panel ?
Last edited:
- Joined
- Jan 15, 2020
- Messages
- 6,895
- Likes
- 16,894
For the front and rear baffle this could theoretically make sense as you don't want them to have the same eigenfrequencies, although alone due to the frontal driver placements they would be different. For the side ones it would also make sense if the dimensions are identical to the front and rear one (which is rare and should be better avoided also due to internal air volume modes), in the end its better though to just engineer them all sufficiently stiff (for example through brackets) and sufficiently damped (through viscoelasticity).
- Joined
- Jan 15, 2020
- Messages
- 6,895
- Likes
- 16,894
That's a selective/partial conclusion though as he writes also:
Thus as said several times, high stiffness on its own is not enough, damping is as important.
Similar threads
