"mutual coupling" is just in-phase summation, and it gives you +6db.
If you double the volume of air moved you get a 6dB boost. And to do so you can double the excursion of single woofer (that means 4 times the power, because P=U^2/R) or double the woofer area (that means two woofers or a bigger one) with the same excursion.
If you double the volume of air moved you get a 6dB boost. And to do so you can double the excursion of single woofer (that means 4 times the power, because P=U^2/R) or double the woofer area (that means two woofers or a bigger one) with the same excursion.
It's been since H.S. for me ( along w a slide-rule
).I figured most of these were demos. of principles...
The most useful one I've found is:
http://www.falstad.com/interference/
Discussing aspects of sound via text is clumsy so I also have referred to these demos:
http://www.santafevisions.com/csf/demos/video/index.htm
http://www.santafevisions.com/csf/demos/audio/index.htm
http://www.santafevisions.com/csf/html/math/001_acoustic_math.htm
http://www.santafevisions.com/csf/html/index.htm
Glad to see this topic about spherical enclosures being raised again. It is a known fact that the diffraction benefit is the major positive to this baffle shape. However, there still seems to be a great debate as whether the internal sphere shape is the best type of enclosure. Rather than discussing 'step baffle' effects, what about the internal standing waves. I've seen on this forum that the spherical enclosure has been rubbished, as it exhibits excessive standing waves and that a normal straight-sided enclosure with non-parallel walls would be far superior to the spherical enclosure.
What are everyone’s views with regards the spherical enclosure itself? Is it the best shape as an enclosure?
What are everyone’s views with regards the spherical enclosure itself? Is it the best shape as an enclosure?
Here's how to build a spherical enclosure the hard way:
http://www.objectreality.co.uk/DeathStarPlans/
http://www.objectreality.co.uk/DeathStarPlans/
Which is also the WRONG way. These spheres do very little or nothing to reduce diffraction. The sharp edges around the entire tweeter will make these not much better than a flat baffle. The important aspect of any enclosure, be it spherical, elliptical, coffinish, etc., is the transition from tweeter faceplate to cabinet walls. I think this is what Geddes was saying so many threads back when he stated an oblate spheroid is better than a sphere.
By using a sphere with a flat area cut onto it for driver mounting, you have done very little to reduce the edge where diffraction occurs. A sphere is not a magical shape that will restore what has already been destroyed by waves being diffracted by a sharp edge. You will have far less diffraction in a rectangular enclosure with a large radius (not merely rounded corners) between tweeter edges and flat walls than you will have with a sphere with sharp edges around the tweeter.
A sphere for woofers is wasted effort, in my opinion, as diffraction is not very audible at lower frequencies, and there aren't many other important advantages to curved enclosure shapes, other than an increase in wall stiffness and strength for similar materials. For the amount of work required, it's a lot simpler to build a strong box.
Peace,
Tom E
By using a sphere with a flat area cut onto it for driver mounting, you have done very little to reduce the edge where diffraction occurs. A sphere is not a magical shape that will restore what has already been destroyed by waves being diffracted by a sharp edge. You will have far less diffraction in a rectangular enclosure with a large radius (not merely rounded corners) between tweeter edges and flat walls than you will have with a sphere with sharp edges around the tweeter.
A sphere for woofers is wasted effort, in my opinion, as diffraction is not very audible at lower frequencies, and there aren't many other important advantages to curved enclosure shapes, other than an increase in wall stiffness and strength for similar materials. For the amount of work required, it's a lot simpler to build a strong box.
Peace,
Tom E
Since it costs very little to fabricate, I would suggest that individuals hear for themselves...
I found an article by R. C. Kral "Diffraction the True Story" in which he throughly recreates the tests performed by Olson in 1951.
He also went on to determine how diffraction affected the response of a "best selling 2 way speaker".
Each cause of diffraction was isolated and removed, until he was left with "an approximate 4db dip @ 2.5Khz, neatly attributable to diffraction around the edges of the enclosure". "Rounding the edges with a file won't help".
"If a driver is not mounted flush with the front baffle, diffraction might add or subtract up to 5db to or from the system's frequency response"
Even .25" height increase of a tweeter baffle board caused problems @ 2.5K.
Because of the 1000/1 size differential between a 20Hz and a 20Khz wavelength; To the large wavelengths of bass the relative dimensions of a cab are of less significance than to the 1.13' or smaller wavelengths of frequencies greater than 1khz.
Which neatly supports my point: the ultimate shape of the enclosure is of little or no consequence to high frequencies, and at low frequencies, it matters not at all. It is the edge nearest the tweeter that makes all the difference. The importance of flush mounting tweeter and even midrange is well documented. If the sharp edge is that of the tweeter faceplate, you're already in trouble, and anything else you do will not improve it. If the tweeter is flush mounted and the first sharp edge is the corner of an enclosure or any other boundary, it's almost as bad, and whatever the overall shape of the enclosure matters not a bit. Diffraction occurs primarily at the first boundary, and anything beyond is almost inconsequential. You can't put the diffracted signal back together with a cool looking sphere.
Don't waste your time building a spherical or any other curved shape enclosure if there are ANY sharp edges that high frequencies pass over. The example spheres do very little to reduce diffraction, and any sphere with a flat cut into it for mounting drivers will be equally bad. If you insist on building an enclosure that way, at least make the flat for the driver irregular in shape or mount the driver eccentrically.
If you must have sharp edges (and by sharp, I mean anything smaller than a 2" radius) on your enclosure, make them non-equidistant from the tweeter so that diffraction effects are spread out. Part of the reason tweeter faceplates cause such problems is that they're symmetrical with the tweeter placed in the center, which is the worst possible configuration. If tweeter faceplates were irregular or off center, the effects would be reduced. Of course, flush mounting eliminates those concerns. Mounting holes are a different story. In fact, I often wonder why every faceplate I've ever seen is perfectly symmetrical, even the mounting holes. Ease of manufacture is certainly a factor, but these engineers must understand that any anomalies caused by one countersunk screw hole is multiplied by however many holes there are.
Peace,
Tom E
Don't waste your time building a spherical or any other curved shape enclosure if there are ANY sharp edges that high frequencies pass over. The example spheres do very little to reduce diffraction, and any sphere with a flat cut into it for mounting drivers will be equally bad. If you insist on building an enclosure that way, at least make the flat for the driver irregular in shape or mount the driver eccentrically.
If you must have sharp edges (and by sharp, I mean anything smaller than a 2" radius) on your enclosure, make them non-equidistant from the tweeter so that diffraction effects are spread out. Part of the reason tweeter faceplates cause such problems is that they're symmetrical with the tweeter placed in the center, which is the worst possible configuration. If tweeter faceplates were irregular or off center, the effects would be reduced. Of course, flush mounting eliminates those concerns. Mounting holes are a different story. In fact, I often wonder why every faceplate I've ever seen is perfectly symmetrical, even the mounting holes. Ease of manufacture is certainly a factor, but these engineers must understand that any anomalies caused by one countersunk screw hole is multiplied by however many holes there are.
Peace,
Tom E
From the same article:
"If the wavelength is small compared to the size of an object, the wave will reflect, If large it will flow around.
Diffraction is contingent upon frequency: The effects of diffraction will vary depending on the relation of the object's size, to the wavelength of the incident wave, Secondly, diffraction is contingent upon the angle at which the incident wave strikes the object ( angle of incidence )."
You missed the:
"Until he was left with "an approximate 4db dip @ 2.5Khz, neatly attributable to diffraction around the edges of the enclosure"
Eliminating the 2 other small diffraction causes, the big one was still left - the cab edge.
Also in Loudspeakers By Philip Newell, Philip Richard Newell, Keith Holland
page 297/298 ( figures 9.23 & 9.24 ) Shows a CEPSTRUM analysis that clearly shows the effect of the edge to tweeter spacing and the resulting diffraction - which corresponds to wavelength length.
But, Again I invite those to make some and hear and decide for themselves.
All of which supports my point even more neatly. You're very good at citing meaningful excerpts from various articles, but do you understand what they're writing?
Yes, diffraction depends on wavelength and angle of incidence. Therefore, diffraction effects are minimal at low frequencies, but crucial at high frequencies. Therefore, diffraction effects are minimal at a radius large in proportion their wavelength, but very pronounced at any small radius (or zero radius) boundary.
I didn't miss a thing. The cabinet edge is important because it's the first boundary between a flush mounted tweeter and free space. If there is any edge, be it the sharp corner of an enclosure or any other edge such as a tweeter faceplate or mounting surface, it will cause diffraction. Whatever shape an enclosure is means absolutely nothing if diffraction has already taken place before the wave encounters it. A sphere improves nothing if the wave is already diffracted.
If reduction of diffraction effects is your goal, building a sphere with any sharp edges anywhere near the tweeter is a waste of time and effort. You might as well invite people to build a speaker with all the low frequencies sent to the tweeter and all the highs sent to the woofer as tell them to try building a sphere such as those shown above, just to hear and decide for themselves.
Peace,
Tom E
Yes, diffraction depends on wavelength and angle of incidence. Therefore, diffraction effects are minimal at low frequencies, but crucial at high frequencies. Therefore, diffraction effects are minimal at a radius large in proportion their wavelength, but very pronounced at any small radius (or zero radius) boundary.
I didn't miss a thing. The cabinet edge is important because it's the first boundary between a flush mounted tweeter and free space. If there is any edge, be it the sharp corner of an enclosure or any other edge such as a tweeter faceplate or mounting surface, it will cause diffraction. Whatever shape an enclosure is means absolutely nothing if diffraction has already taken place before the wave encounters it. A sphere improves nothing if the wave is already diffracted.
If reduction of diffraction effects is your goal, building a sphere with any sharp edges anywhere near the tweeter is a waste of time and effort. You might as well invite people to build a speaker with all the low frequencies sent to the tweeter and all the highs sent to the woofer as tell them to try building a sphere such as those shown above, just to hear and decide for themselves.
Peace,
Tom E
madisonears said:
That is false.
Diffraction effects extend well down in frequency. How far depends on how wide the "baffle" is. The nature of the diffraction at different frequencies is sufficiently different on a gross level that we do tend to pigeon hole LF diffraction as baffle step and HF diffraction as edge diffraction, but it is really a continuum of effect. Olson, and those that followed have clearly shown that cabinet shape has a distinct affect on the magnitude and nature of the ripple as we pass thru the zone of diffraction nominally called baffle step. And that the sphere has a very low impact. That many of our diy attempts at getting a driver into a sphere exacerbate the HF edge diffraction issues is a different subject. I still hold the B&W nautilus up as an example of a very good execution.
dave
Thanks for clarifying that. Of course diffraction occurs at all frequencies as a wave encounters a boundary; it doesn't magically stop at the highs or even midrange. What we're discussing here, however, is a small spherical enclosure which would have minimal audible effects on low frequencies because the wavelength is large compared to the enclosure. I doubt that anyone would be able to or care to construct a sphere large enough to affect low frequencies.
On the other hand, this size sphere can have significant audible diffraction components if the tweeter or midrange is mounted on a flat cut into the sphere as shown, and even worse with the type of mounting plate edges shown in the example. My point was that going to the trouble of building a sphere and then using that type of mounting will not produce the desired effect: reducing diffraction of frequencies significant to this size enclosure, which would be higher frequencies. IF that's what you're after, the sphere is really not much better than a properly constructed box.
There are ways to compensate for the effects of diffraction of lower frequencies, but there's no way to correct diffraction of high frequencies once it's there. The best way to avoid it is by following a few simple guidelines. I would also postulate that the effects of high frequency diffraction are more disruptive to what most listeners would consider "good" sound: ragged FR on and off axis and poor imaging.
Peace,
Tom E
On the other hand, this size sphere can have significant audible diffraction components if the tweeter or midrange is mounted on a flat cut into the sphere as shown, and even worse with the type of mounting plate edges shown in the example. My point was that going to the trouble of building a sphere and then using that type of mounting will not produce the desired effect: reducing diffraction of frequencies significant to this size enclosure, which would be higher frequencies. IF that's what you're after, the sphere is really not much better than a properly constructed box.
There are ways to compensate for the effects of diffraction of lower frequencies, but there's no way to correct diffraction of high frequencies once it's there. The best way to avoid it is by following a few simple guidelines. I would also postulate that the effects of high frequency diffraction are more disruptive to what most listeners would consider "good" sound: ragged FR on and off axis and poor imaging.
Peace,
Tom E
I'm glad you mentioned the Nautilus again because it really is an excellent design for a number of reasons, especially the way it deals with reducing diffraction of highs and mids.
Below are some pictures of my own experiment with reducing diffraction. I built these a few years before the Nautilus was introduced. In fact, the experiment was so successful that I was considering tooling up to produce these commercially until I saw the first ad for the Nautilus. I even considered contacting B&W to discuss compensation. Their idea was publicized before mine, however, and I quickly came to my senses.
Besides, I like B&W speakers a lot. I purchased one of the first 50 pairs of B&W speakers to be imported to the US in 1976, the DM6. I sold them to a good friend ten years ago and he still enjoys them. I currently own a pair of 802's, but I've replaced the stock crossovers with outboard units. B&W is very innovative at driver design and implementation, but they don't seem to put much importance in crossovers. They function, but they don't excel.
The speakers pictured here contain a 6-1/2" Peerless CC Line woofer and a Vifa tweeter. All four drivers cost under $100. The cabs are laminated, plywood and particle board layers mixed. Even mounted flat against the wall, they image like crazy and produce a deep 3D soundstage. If the FR weren't quite so rough (if I had used better drivers), these would be in my main system with a sub. As it happens, they were once used in the kitchen, but now reside in my workshop (hence the coating of sawdust!).
Peace,
Tom E
Below are some pictures of my own experiment with reducing diffraction. I built these a few years before the Nautilus was introduced. In fact, the experiment was so successful that I was considering tooling up to produce these commercially until I saw the first ad for the Nautilus. I even considered contacting B&W to discuss compensation. Their idea was publicized before mine, however, and I quickly came to my senses.
Besides, I like B&W speakers a lot. I purchased one of the first 50 pairs of B&W speakers to be imported to the US in 1976, the DM6. I sold them to a good friend ten years ago and he still enjoys them. I currently own a pair of 802's, but I've replaced the stock crossovers with outboard units. B&W is very innovative at driver design and implementation, but they don't seem to put much importance in crossovers. They function, but they don't excel.
The speakers pictured here contain a 6-1/2" Peerless CC Line woofer and a Vifa tweeter. All four drivers cost under $100. The cabs are laminated, plywood and particle board layers mixed. Even mounted flat against the wall, they image like crazy and produce a deep 3D soundstage. If the FR weren't quite so rough (if I had used better drivers), these would be in my main system with a sub. As it happens, they were once used in the kitchen, but now reside in my workshop (hence the coating of sawdust!).
Peace,
Tom E
Attachments
madisonears said:
For the purposes of this discussion, what defines LF is relative to the size of the baffle. The worst shapes can have on the order of 10 dB ripple as we transition from 2 pi steriradians to 4 pi steriradians... and it has nothing to do with edge diffraction. The sphere has no ripple, just a smooth transition to a shelf 6 dB down.
It is still important to pay attention to edge diffraction, either by ensuring there is a minimum number of "edges" to diffract off of, or that nothing gets to the edges to diffract.
Nice boxes you have there.
dave
planet10 said:
R. C. Kral was able to confirm Olson's results.
( For me it was the purchase of the Focal "Egg" that led me to build non-boxes. )
Unfortunately I don't have access to an "802" to lift a pattern from.
I have found it much easier to copy, or utilize an existing curved shape than to have to resort to "Death Star" techniques.
R. C. Kral was able to confirm Olson's results.
( For me it was the purchase of the Focal "Egg" that led me to build non-boxes. )
Unfortunately I don't have access to an "802" to lift a pattern from.
I have found it much easier to copy, or utilize an existing curved shape than to have to resort to "Death Star" techniques.
If you are referring to a large radius baffle that is easy to do... I will show pics of my current build shortly...
as long as you can work with fiberglass...
as long as you can work with fiberglass...
HK26147 said:
- Status
- Not open for further replies.