I don't know if there was a conclusion, but that can be an issue, depending on the woofer module's depth.
I also think the smaller version(5.5") of the illuminator 7" paper cone could be an alternative, making a narrower and shallower cylinder.
I also have to say that having 7" midbass is a very tempting option, especially if it can be in the adaptive module usable for either bass or midbass by plugging/unplugging the port. I want to list all the options after the bass module dimension is determined.
Thank you very much again for your help and efforts.
Is this for 5.5" or 7" midbass, or 2x 8" woofers? I think I want the woofers to cover from (50~70)Hz to (150~300)Hz, and the midbass from (150~300) to (700~850)Hz.
You are talking about triple woofers, possibly without midbass, right? Using 3 same drivers for a way, are 2 connected in parallel, than the parallel pair connected to the other in series? Connecting all 3 in series or parallel seems to present some extreme driving condition, but parallel/serial hybrid connection seems to cause unbalanced power distribution....
I think I prefer 2x 8" to 3x 7", but the latter is obviously worth a consideration. I guess 2x 8" has a lower Fs advantage. Can an SVS 13" sealed sub covering up to 50~70Hz below the woofer make things any easier?
Is this for 5.5" or 7" midbass, or 2x 8" woofers? I think I want the woofers to cover from (50~70)Hz to (150~300)Hz, and the midbass from (150~300) to (700~850)Hz.
You are talking about triple woofers, possibly without midbass, right? Using 3 same drivers for a way, are 2 connected in parallel, than the parallel pair connected to the other in series? Connecting all 3 in series or parallel seems to present some extreme driving condition, but parallel/serial hybrid connection seems to cause unbalanced power distribution....
I think I prefer 2x 8" to 3x 7", but the latter is obviously worth a consideration. I guess 2x 8" has a lower Fs advantage. Can an SVS 13" sealed sub covering up to 50~70Hz below the woofer make things any easier?
Last edited:
Spherical enclosures are attractive and good-looking. I am afraid they are just difficult to make and stack. Maybe half-spherical baffle(or the part of the enclosure holding/surrounding the driver frame) in front of a pipe?
How about a tapered cylinder? I copied Gradient 1.4 concept using coaxial mid+tweeter. Tube has downfire subwoofer and a Hypex FA123 dsp-amp module.
Well, sure spherical enclosures are not easy to do. And i won't commit to something myself as i definitely don't have the skills ( i'm working on this skills for some years now... but i'm just good enough for prototype shoebox... some projects on here are just gems... makes me feel ashamed and jealous sometimes!).
That said the ones ( spherical enclosure) i've heard so far sounded rather good to me besides their weakness ( Elipson's from 60's/70's are somewhat dated on some points imho and Cabasse's Baltic were satellite+woofer principle and i'm not fond of it).
About box shape, why reinvent the wheel or shout in the dark? Pioneers already done the job :
https://www.diyaudio.com/community/...pe-to-make-at-home.380342/page-5#post-6876651
The round over towards a spherical shape is what stops multiple reflections coming back.
The baffle step is a loss off efficiency versus frequency relative to the width and shape of the baffle and the diameter of the driver.
The boundary of the room lifts the response below 100 hertz.
An option to be investigated is a 6.5 in woofer instead of an 8 inch.
The VAS of the 6.5 inch woofers is lower than an 8 inch. This will reduce the length of the tubes.
Here is an example using four 6.5 inch woofers
The cone area is the equivalent of a 12 inch woofer.
Configuration
Bottom pair stacked near floor.
Top pair are separated by a gap large enough to place the 2 inch dome with crossover at 800 hertz with one inch tweeter above.
Because the diameter of woofers and the dome mid and tweeter is relatively small compared to wavelength at 800 hertz an MTM acoustic concept is possible.
This will acoustically resemble an MTM array with augmented low frequency support.
There are a number of advantages to the MTM in terms of beam steering the midrange polar response in the vertical plane to reduce floor and ceiling interference. You also get vertical scale in the sound stage with a virtual point source in the 500-800 & 800-3000 hertz region.
Each pair could be put in parallel electrically which will simplify the active crossover and EQ to linearise the baffle step. You can also tilt the vertical polar axis to focus on the listening height.
There are technical advantages of using 6.5 inch woofers. They are better in the midrange and will have lower cone mass. . As mentioned they need less enclosure volume.
There are more and better choices of 6.5 inch woofers. They are low cost..
It’s important to use four identical woofers.
Typically the low pair will have an approximate 6db per octave low pass filter in the baffle step region possibly around 300 hertz. The top pair run up to the crossover point of 800 hertz.
Because you are using an active system to have some options such as cascading the filters to create the output of the lower woofers. We can look at this in more detail later on.
The Seas range is excellent in Nextel, Satori and Scanspeak.
Because you have four woofers the work on each driver is divided by four. The motional displacement is 1/16 that of a single 6.5 inch woofer. This means very low distortion compared to a single driver.
The baffle step is a loss off efficiency versus frequency relative to the width and shape of the baffle and the diameter of the driver.
The boundary of the room lifts the response below 100 hertz.
An option to be investigated is a 6.5 in woofer instead of an 8 inch.
The VAS of the 6.5 inch woofers is lower than an 8 inch. This will reduce the length of the tubes.
Here is an example using four 6.5 inch woofers
The cone area is the equivalent of a 12 inch woofer.
Configuration
Bottom pair stacked near floor.
Top pair are separated by a gap large enough to place the 2 inch dome with crossover at 800 hertz with one inch tweeter above.
Because the diameter of woofers and the dome mid and tweeter is relatively small compared to wavelength at 800 hertz an MTM acoustic concept is possible.
This will acoustically resemble an MTM array with augmented low frequency support.
There are a number of advantages to the MTM in terms of beam steering the midrange polar response in the vertical plane to reduce floor and ceiling interference. You also get vertical scale in the sound stage with a virtual point source in the 500-800 & 800-3000 hertz region.
Each pair could be put in parallel electrically which will simplify the active crossover and EQ to linearise the baffle step. You can also tilt the vertical polar axis to focus on the listening height.
There are technical advantages of using 6.5 inch woofers. They are better in the midrange and will have lower cone mass. . As mentioned they need less enclosure volume.
There are more and better choices of 6.5 inch woofers. They are low cost..
It’s important to use four identical woofers.
Typically the low pair will have an approximate 6db per octave low pass filter in the baffle step region possibly around 300 hertz. The top pair run up to the crossover point of 800 hertz.
Because you are using an active system to have some options such as cascading the filters to create the output of the lower woofers. We can look at this in more detail later on.
The Seas range is excellent in Nextel, Satori and Scanspeak.
Because you have four woofers the work on each driver is divided by four. The motional displacement is 1/16 that of a single 6.5 inch woofer. This means very low distortion compared to a single driver.
Attachments
Sorry Juhazi, I can't stack the tapered cylinder. It also looks difficult if you meant to mound 2 or 3 drivers with it standing up like the example you showed...
What I'd like most is the cone + cylinder type as I tried to show with the steel milk can. Hopefully cone + cylinder is just as good as the cone...
Do you have an issue with cabinet depth? You an aways drop a subwoofer in a sub enclosure of the main cabinet and set it up facing sideways and run sealed a port or passive radiation on the other side driver depth permitting
Rob 🙂
Rob 🙂
Why do you think a cone (or cylinder) is good? The B&W for example with their sphere shaped cabinets creates an identifiable characteristic sound rather than a neutral one. It is sufficiently strong they tone it down a bit by putting significant lumps and dips into the on-axis frequency response to compensate. Or is an audiophile/characteristic sound fine?
Not trying to put you off stacking cylinders if that makes for a fun project just pointing out there is a cost compared to a cabinet+crossover that smoothly controls the off-axis radiation to achieve a neutral sound. Do you intend to put the 2 x 8" woofers in a single oval baffled cabinet or a pair of cylindrical ones?
First of all, I am sorry if I haven't been clear. I don't want to build 2 ~ 3 cylinders for the woofers. I have been planning to use a plain narrow box for the bass module. I may use trapezoid shape or some rounding, but still a box.
Hopefully linear slope has a similar effect to that.
This is also very interesting. I haven't been thinking of using multiple drivers for anything other than woofers, worried about interferences from multiple drivers. To me, vertical directivity control seems a bit tricky and too much dependent on the environment.
Are they still in the same 'way' of woofer band, so driven by the same electrical signal?
Do you mean different frequency ranges for some of the woofers?
I will focus on those drivers. I was misunderstanding and a bit disappointed by Seas nextel ones, but need to look at them again. Satori papyrus ones and Scanspeak revelators(maybe too expensive?) are also interesting.
So, I guess there is no way to evenly drive 3 drivers without making the load impedance too low or too high.
Hopefully linear slope has a similar effect to that.
This is also very interesting. I haven't been thinking of using multiple drivers for anything other than woofers, worried about interferences from multiple drivers. To me, vertical directivity control seems a bit tricky and too much dependent on the environment.
Are they still in the same 'way' of woofer band, so driven by the same electrical signal?
Do you mean different frequency ranges for some of the woofers?
I will focus on those drivers. I was misunderstanding and a bit disappointed by Seas nextel ones, but need to look at them again. Satori papyrus ones and Scanspeak revelators(maybe too expensive?) are also interesting.
So, I guess there is no way to evenly drive 3 drivers without making the load impedance too low or too high.
No. Cylinders are for midbass, midhigh and tweeter. I was planning to put them in a tall narrow deep box.
Is the layout of 2 8" woofers on the top part of a tall narrow deep box still problematic? I think we might have been talking past each other, so want to confirm.
Regarding midrange enclosure, round/spherical inner space is worst scenario regarding inner modes. Even dimensions will make very strong 1st mode depending on diameter, easily smack dab in the mid's passband.
Some room mode calculations allow the user to use small dimensions, even non-rectangular shape eg. https://amcoustics.com/tools/amroc
Heavy stuffing can minimize modes, but the more asymmetrical dimesions and shape are, the better. There are many ways to make outside and inside of the "box" different, allowing symmetrical, cylindrical, whatever outside shape.
Bass boxes must be quite big to have modes in passband, closed (acoustic damping) are typically smaller than resonant (BR)
Some room mode calculations allow the user to use small dimensions, even non-rectangular shape eg. https://amcoustics.com/tools/amroc
Heavy stuffing can minimize modes, but the more asymmetrical dimesions and shape are, the better. There are many ways to make outside and inside of the "box" different, allowing symmetrical, cylindrical, whatever outside shape.
Bass boxes must be quite big to have modes in passband, closed (acoustic damping) are typically smaller than resonant (BR)
For clarification the four driver are driven ad two pairs in parallel making a load of four ohms for an 8 oh driver.
I have attached a sketch for you to ponder.
My efforts are purely to assist you in evolving a solution that will work acoustically and embody your values.
Conceptually you have a solution for the baffle step using two pairs of 6 1/2 inch identical drivers. This is to ensure consistency in the timbre during the baffle step transition.
The four reasonably good 6 1/2 drivers will outperform any single hi end 8 inch woofer on displacement and distortion due to operating in a very linear range of the BL curve
The load sharing will significantly reduce voice coil temperature and as a result import transient snap. This is what suffers with power compression with relatively small voice coils.
A 6 1/2 inch driver will offer you more flexibility in the crossover design and with any future plans to try other upper midrange drivers. The thing is the 6 1/2 inch driver can work up 3000 hertz . You may want to experiment with a tweeter wave guide or a compression driver wave guide later on.
My suggestion is to think about your project as a journey and not so much an end game design at the outset. You will have more fun with the basis of the system providing flexibility to easily make change.
Where as picking out a specific driver to work with another drive can leave you with a narrow choices it you decide to change your mind.
Attachments
I didn't think much about the cylinder cavity, but you are right that it must be as bad as square cross-sectional shape. It seems that the tapered cavity is a must. If I go with a small-faced deep box, what do you think the effect of the baffle would be?
Biggest factor is baffle's frontal dimensions. Second is edge rounding/chamfering, third and very far behind side- and backside wall shape
Classic AES papers from Harry F. Olson
More about baffle diffractions:
https://www.diyaudio.com/community/threads/diffraction.196668/
https://trueaudio.com/st_diff1.htm
https://heissmann-acoustics.de/en/kantendiffraktion-sekundaerschallquellen-treiberanordnun/
Most of these papers and abstracts discuss only axial response. Off-axis effects have not been studied extensively. As well radiator membrane diameter and shape are important factors in mid-high frequencies, this is why we need test enclosures and measurements of real drivers in them - if we search for best possible compromise.
The most comprehensive design-simulation freeware-software is VituixCAD 2 by member kimmosto https://kimmosaunisto.net/
https://www.diyaudio.com/community/threads/complete-speaker-design-software.357583/
Just discussing,studying old papers and using simple simulations will only help us to find guidelines.
Very good bad examples of "smart" shape design are B&W and Wilson top models. Good examples are KEF, Estelon, Gradient, Genelec and Vivid Audio. Very good results can come from shoeboxes too like D&D and Kii3
Classic AES papers from Harry F. Olson
More about baffle diffractions:
https://www.diyaudio.com/community/threads/diffraction.196668/
https://trueaudio.com/st_diff1.htm
https://heissmann-acoustics.de/en/kantendiffraktion-sekundaerschallquellen-treiberanordnun/
Most of these papers and abstracts discuss only axial response. Off-axis effects have not been studied extensively. As well radiator membrane diameter and shape are important factors in mid-high frequencies, this is why we need test enclosures and measurements of real drivers in them - if we search for best possible compromise.
The most comprehensive design-simulation freeware-software is VituixCAD 2 by member kimmosto https://kimmosaunisto.net/
https://www.diyaudio.com/community/threads/complete-speaker-design-software.357583/
Just discussing,studying old papers and using simple simulations will only help us to find guidelines.
Very good bad examples of "smart" shape design are B&W and Wilson top models. Good examples are KEF, Estelon, Gradient, Genelec and Vivid Audio. Very good results can come from shoeboxes too like D&D and Kii3
Last edited:
Yes, I agree... A rectangular baffle has many advantages. When a circular driver is placed on a rectangular baffle, the distance from the driver to the edge varies as the angle clocks around the driver. A minimum variation of 1.44 if the baffle is square and driver is centrally located, but most rectangular baffles will have a variation of more than 1.7.
A circular baffle, on the other hand, has a constant distance from driver to edge, and this will maximize the size of the diffraction hump.
As has been pointed out earlier by others, the internal shape of a cylinder is not ideal. It will have a very strong standing wave mode based on the diameter. Heavy acoustical damping can mostly overcome this, but it would be better to avoid designing in a problem which will have to be dealt with later on.
A box shape with internal dimensions of 1 : 1.6 : 2.6 has three weak standing wave modes which are spread out, and much easier to dampen. This is why this shape is so common in speaker cabinets. It works well.
j.
This suggests the edge is not favourable and some might see that as a red flag.
What if the lateral mode is not excited?
Hehe yeah, careful there guys, all this diffraction stuff not only depends on shape and size of the baffle, but also size of the transducer. Main sin for the old Olson paper is that the transducer is only very small compared to the structure, so the diffraction related interference is as bad as can be. If the whole end cap of that cylinder in image C was the transducer, it would be second best looking graph among the set.
If baffle is big compared to transducer, it's much better to have varying distance to edge, but if baffle is no bigger than the transducer, it really doesn't matter that much assuming it works for the system it goes into. Any case, it isn't optimal to use bandwidth beyond the main diffraction hump due to many reasons, cone breakup and diffraction and eventually beaming. Tweeter has no low pass, and if it's the typical dome in flange there is no option than put it on a big sphere, or rectangle, the flange ruins the opportunity to have it a minimal baffle as it's not. Even more optimal would be to have waveguide, or the sphere, than rectangle in this regard.
But, as speakers aren't all about edge diffraction but sum of everything, rectangular baffle is above all quite handy to make, and likely just fine especially if there is some round overs or slants. If there is possibility to optimize further, then don't use rectangular baffle but optimize further. After all, all of this is readily visible and easy to experiment with the modern tools, even VituixCAD, or more elaborately with BEM.
If OP wants to use tubular enclosure for each driver it should work just fine for the mids with damping inside, but why put tube for tweeter at all? why use tweeter with flange? Why not find smallest dome tweeter without any flange and stick it end of a stick? why not use waveguide instead, there is no particular need for enclosure volume for tweeters, they are usually self contained. If it must be tube and tweeter with flange due to looks then it's fine, one just has to live with it and not care about side-effects.
If baffle is big compared to transducer, it's much better to have varying distance to edge, but if baffle is no bigger than the transducer, it really doesn't matter that much assuming it works for the system it goes into. Any case, it isn't optimal to use bandwidth beyond the main diffraction hump due to many reasons, cone breakup and diffraction and eventually beaming. Tweeter has no low pass, and if it's the typical dome in flange there is no option than put it on a big sphere, or rectangle, the flange ruins the opportunity to have it a minimal baffle as it's not. Even more optimal would be to have waveguide, or the sphere, than rectangle in this regard.
But, as speakers aren't all about edge diffraction but sum of everything, rectangular baffle is above all quite handy to make, and likely just fine especially if there is some round overs or slants. If there is possibility to optimize further, then don't use rectangular baffle but optimize further. After all, all of this is readily visible and easy to experiment with the modern tools, even VituixCAD, or more elaborately with BEM.
If OP wants to use tubular enclosure for each driver it should work just fine for the mids with damping inside, but why put tube for tweeter at all? why use tweeter with flange? Why not find smallest dome tweeter without any flange and stick it end of a stick? why not use waveguide instead, there is no particular need for enclosure volume for tweeters, they are usually self contained. If it must be tube and tweeter with flange due to looks then it's fine, one just has to live with it and not care about side-effects.
Last edited: