Zero Baffle Width, Zero Edges...Sort Of...

[bluewizard]

I'm wondering if any do-it-yourselfers has ever tried a design like this.





These are from the Bower & Wilkins Signature series.

As I was reading several discussion about how critical baffle width is to a speaker design, and how problematic baffle edges are, the concept of the B&W Signature series came into my mind, though at the time I thought I had come up with something original.

Essentially, you have a curved front cabinet that has no width or edges; more or less. Into that you make small round vertical tunnels just wide enough to mount the speakers. In my mind I had envisioned a multi-woofer design in a WWMT configuration.

Conceptually it seems like a good idea, but there must be some drawbacks besides the complexity of construction. Which seem fiddly but not all that complicated.

So, to your knowledge, has anyone tried this, and can anyone tell me what the advantages and disadvantages of this design would be?

Just curious. It seemed like a good idea, and I'm surprised lots of people haven't thought of it and tried it.

Steve/bluewizard

[Mike B]

I am in the process of trying it.

https://www.htguide.com/forum/showthread.php?t=31204

[Curt C]

Quick model:

Edge prediction of diffraction effects of no baffle compared to a narrow rectangular baffle. As expected, the BSC frequency is higher for the 'no baffle' design, but ripple above that frequency is not significantly different. If crossed at or above the baffle step, it would be a viable design.

C

[norcad]

From John Murphys article:

"A spherical enclosure exhibits this transition clearly with a very smooth diffraction loss curve"

And

"The worst case appears to be placing the driver at the center of a circular baffle so that it is the same distance from all diffracting edges"

For me this looks like the worst case. It doesnt matter if its a circular baffle or no baffle at all. The driver itself is are circular with just a smaller diameter. You only push the diffraction into higher frequencies.

[KevinB]

I thought these looked great when i stumbled on them. However after doing some reading into circular baffles everything i found agreed with norcads comments.

http://www.miwis-bastelbu.de/Galerien/Leistungsdruck/index.html

Cant speak German (?) so not sure what the writing talks about.

[dlr]

Driver characteristics change things

No one is recognizing the waveguide nature of the B&W midrange unit. They have used this concept for years. Being a well-damped Kevlar diaphragm coupled with the geometry, there is likely very little if any diffraction effects from this unit other than step alone and any internal diffraction/reflection of any non-perfect nature of the waveguide itself. I would bet a significant amount of money that this is the case. The directionality makes anything other than step a moot point, I believe.

Dave

[dlneubec]

Dave,

If that were the case, then it would not matter what shape the baffle or cabinet is, so what is the benefit, if any, to the curved cabinet they use?

Taking away the waveguide issue, a conventional mount in this manner would have to be one of the worst, in terms of diffraction, it seems to me. It is obviously diificult to mount a driver directly on a curved surface. I know Curt did a version of this in his Exclamations! with a sphere for a cabinet shape, a design I have yet to hear, but will at Iowa, I believe.

I tried to do something similar to mount a small midrange in a curved cabinet, though it is a much different application, in my SoundRounds design.

[dlr]

It's very driver dependent

Dave,

If that were the case, then it would not matter what shape the baffle or cabinet is, so what is the benefit, if any, to the curved cabinet they use?

Taking away the waveguide issue, a conventional mount in this manner would have to be one of the worst, in terms of diffraction, it seems to me. It is obviously diificult to mount a driver directly on a curved surface. I know Curt did a version of this in his Exclamations! with a sphere for a cabinet shape, a design I have yet to hear, but will at Iowa, I believe.

I tried to do something similar to mount a small midrange in a curved cabinet, though it is a much different application, in my SoundRounds design.

I have to disagree. I left out a detail. When mounted on a largish baffle, there will be increasing diffraction effects at the lower end. However, the step for that driver mounting is essentially that of the driver diameter, above that it depends on the dispersion of the driver. For a flat driver that size there would be limited diffraction, as noted by Curt's post. However, this is a much more directional driver than dictated purely by diameter. This is not taken into account by the BDS nor The Edge, I believe. In actual practice, diffraction above the step is often much less significant for that size driver and larger . It's highly driver dependent. That driver appears to be about 17cm and it's very deep. It's not going to "illuminate" the edge nearly as much as some other drivers might, especially smallish true midranges or midwoofers.

I have found significant diffraction in small drivers. The 12/4631, 13m/8636, 13/8640, RS125, RS150 and others. Each has it's own signature that is dependent on the 90 off-axis signal. 6.5" and up drivers just are almost universally too directional for diffraction other than step to be significant. What is almost always more detrimental is the diffraction of the tweeter output caused by adjacent drivers. The configuration of that B&W system should be relatively good insofar as most diffraction is concerned.

Dave

Note: Edits in bold

[dlneubec]

Dave,

What about the opposit situation? What is your opinion of larger midrange drivers, say 6.5" and above, that are rear mounted or recessed, say 1/2" deep with a 1/2" roundover, crossed above 400hz or so? Would you typically expect them to have any siginificant diffraction signature above the step?

[bluewizard]

Excellent information, thank you.

First, the example I linked to isn't exactly what I would call a spherical or circular enclosure. Tubular is more like it, but in this case, the enclosure isn't really tubular, just a small section where the speaker actually mounts. Now the front of the cabinet in which the short tubular mount is placed is curved. But again not spherical or completely tubular.

I've been lead to believe that one of the worst cabinets is a tubular cabinet with the speaker mounted in the center of the circular end. Similar to the Leistungsdruck linked to by KevinB.

To MikeB, I saw your thread earlier and thought it was pretty cool, but the way the speakers are mounted never really sunk in until you mentioned it. Even more cool.

Aesthetically, the B&W seemed like a cool idea, but I really wasn't sure what the plus and minuses of such a design would be.

Curt C,

Your graph seems to indicate that the Baffle Step ends at about 1.6khz, and you say "...If crossed at or above the baffle step, it would be a viable design. ...". Keep in mind, I'm out of my element here, but are you referring to the Bass-to-Mid crossover? Would two woofers crossed near 1.6khz be sufficient to offset the baffle step?

Again well aware that my ignorance is showing, in using a 'half' woofer, as is a 2.5 or 3.5 way system, isn't the 0.5 woofer usually crossed in the middle of the baffle step slope? And for more typical and common cabinets, doesn't that slope usually occur in the 200hz to 500hz range with the middle being roughly 300hz to 350hz?

What I'm getting at, is in at 2.5-way design, the 0.5 woofer increase the output to offset the loss of sound because of baffle step, and is set in the middle of the baffle step slope so that as the 1.0 woofer begins to increase in volume as it rises above the baffle step the 0.5 woofer fades out.

However, if the baffle step were sufficiently high as we seem to have in our hypothetical B&W case, couldn't two straight-up woofers running in parallel, without the 0.5-way aspect, compensate nicely?

I'm not sure I phrased that that well, but hopefully you get my point.

Also regarding your chart, isn't the 'Zero Baffle' the green line? While not hugely different, it does seem to be an improvement over the red line.

Again, as a point of curiosity, I wonder about the pluses and minuses of this design concept, and do appreciate all who took the time to respond.

Steve/bluewizard

[dlr]

Dave,

What about the opposite situation? What is your opinion of larger midrange drivers, say 6.5" and above, that are rear mounted or recessed, say 1/2" deep with a 1/2" roundover, crossed above 400hz or so? Would you typically expect them to have any siginificant diffraction signature above the step?

I'm not at all a fan of rear-mounts with possible exception of woofers and passive radiators. Consider that horn/waveguide drivers that have a significant half-roll surround will exhibit diffraction (localized) due to the surround and that all non-flat diaphragms are a horn/waveguide to one degree or another.

Limited true midrange drivers such as those of KEF use a flat surround. I have a pair of their midrange/tweeters coaxials. They are 6.5" and fit perfectly in the Insignia boxes. I thought that they were bad at first, because pressing the diaphragm of the mid caused almost no deflection. They have very limited excursion, but the relatively large diameter gives it the SPL with low displacement, just very limited low end extension.

A waveguide is intended to be effective to its edges. The edge termination has an impact. A rear-mount can only be detrimental above some frequency that is dependent on the geometry of the "new" termination.

Dave

[Curt C]

Curt C,

Your graph seems to indicate that the Baffle Step ends at about 1.6khz, and you say "...If crossed at or above the baffle step, it would be a viable design. ...". Keep in mind, I'm out of my element here, but are you referring to the Bass-to-Mid crossover? Would two woofers crossed near 1.6khz be sufficient to offset the baffle step?

Again well aware that my ignorance is showing, in using a 'half' woofer, as is a 2.5 or 3.5 way system, isn't the 0.5 woofer usually crossed in the middle of the baffle step slope? And for more typical and common cabinets, doesn't that slope usually occur in the 200hz to 500hz range with the middle being roughly 300hz to 350hz?

Steve,
Sorry I was unclear in my post. –Trying to work posts in between phone calls sometimes leaves my multitasking abilities wanting…

What I meant to say / should have said, is that the 6 dB shelf caused by the 4pi to 2pi transition will need to be addressed in the crossover. This could be done by crossing appropriately, or, by the addition of a .5 driver as you suggest, or by shelving down the response above the step to compensate.

What I'm getting at, is in at 2.5-way design, the 0.5 woofer increase the output to offset the loss of sound because of baffle step, and is set in the middle of the baffle step slope so that as the 1.0 woofer begins to increase in volume as it rises above the baffle step the 0.5 woofer fades out.

However, if the baffle step were sufficiently high as we seem to have in our hypothetical B&W case, couldn't two straight-up woofers running in parallel, without the 0.5-way aspect, compensate nicely?

In an ideal world, below the step, both woofers provide equal output. As we approach the step the .5 woofer is shelved down to emulate a transfer function that is equal but opposite of the baffle step, while the other woofer continues un attenuated until it reaches the crossover frequency of the tweeter, where they are both attenuated with whatever transfer function is appropriate to provide a smooth transition to the tweeter and its high pass transfer function.

Also regarding your chart, isn't the 'Zero Baffle' the green line? While not hugely different, it does seem to be an improvement over the red line.

IMO, The primary benefit of the no-baffle design would seem to be a relatively symmetrical diffraction pattern, and perhaps just as important, a more time coherent summation of the direct sound and baffle edge reflections.

C

[Curt C]

When mounted on a largish baffle, there will be increasing diffraction effects at the lower end. However, the step for that driver mounting is essentially that of the driver diameter, above that it depends on the dispersion of the driver.

The other night, I had the occasion to measure a SS 18W8545 / Seas 27TBFC/G on a conventinal 9" wide baffle. The question was raised as to where to cross the drivers. I demonstrated the upper frequency limitation by measuring the woofer at 45 degrees off-axis. The woofer response up to about 2.2 kHz was consistent with the on-axis response, but above that dropped like a rock. This validates Dave's statement that with larger drivers and narrow baffles, the only diffraction effect evident in the driver's acceptable passband would be the baffle step.

It also indicates that the old saw: 'Don't cross woofers above frequecies where their wavelengths are smaller than the driver diameter' has validity.

C

[dlr]

IMO, The primary benefit of the no-baffle design would seem to be a relatively symmetrical diffraction pattern, and perhaps just as important, a more time coherent summation of the direct sound and baffle edge reflections.

C

I've never tried this kind of setup, but if the drivers are chosen well (relating to a midrange highpass, 2-ways need not apply IMO only of course), I can see some benefits. One is the symmetrical diffraction pattern as you say. The off-axis response horizontally should be much more uniform. Edge treatment and/or baffle felt would be moot. To extend this, I would say that it might make the XO design easier, since the on-axis vs. off-axis will have fewer deviations.

There may be a downside, however. That would be possible internal reflections/restrictions behind the drivers (tweeter excepted, of course). If a smallish midwoofer or true midrange is used, the opening behind it cannot be as expansive as a larger baffle would allow. If the baffle (mounting board or whatever is) is made slightly larger to allow a chamfer on the back side, it may be enough to allow non-step diffraction to start to enter the picture. If you've ever read the B&W paper on the design considerations behind the gourd-shaped chambers it points out what can occur behind a driver. The driver frame must be very open if the cylinder used is very small in diameter.

Another aspect of which I have no idea the import is the increased power response. A driver on a cylinder will have much more of its output in the 4-pi region. This seems to work well for B&W so either it's a benefit or benign. I doubt that it's detrimental. It probably would require more distance to any reflective surfaces to the side.

Dave

[bluewizard]

A bit of a side track here -

Curt C said -

"It also indicates that the old saw: 'Don't cross woofers above frequencies where their wavelengths are smaller than the driver diameter' has validity."

Maybe my calculator is broken but 12" roughly corresponds to 90hz, so by this adage, I shouldn't cross a 12" woofer above 90hz?

That seems kind of impractical to me.

What am I missing here?

steve/bluewizard

[augerpro]

The only thing you're missing Steve is 12" is the wavelength of 1125hz :B

[bluewizard]

AURGH!!! I'm mixing up feet and inches! 90hz is 12 feet not 12 inches.

Ok now it makes sense. Sorry, I'll take a pill now and then everything will be all better.

Steve/bluewizard

[Curt C]

There may be a downside, however. That would be possible internal reflections/restrictions behind the drivers (tweeter excepted, of course). If a smallish midwoofer or true midrange is used, the opening behind it cannot be as expansive as a larger baffle would allow. If the baffle (mounting board or whatever is) is made slightly larger to allow a chamfer on the back side, it may be enough to allow non-step diffraction to start to enter the picture. If you've ever read the B&W paper on the design considerations behind the gourd-shaped chambers it points out what can occur behind a driver. The driver frame must be very open if the cylinder used is very small in diameter.

Dave

Case in point, (to exemplify your point, Dave) was the mid installation in the Statement series.

The TB W4-1337SA is has a nice open frame, although the large spider appears compromise that somewhat. The driver was fitted to a square tubular MDF enclosure with internal dimensions of 5” per side, and roughly 16” long. The through hole was 3.875” and the rear of the 1” MDF baffle was scalloped. Originally I lined the ‘enclosure’ sides full length with 1” foam, but found the sound was dull and lifeless (think the sound of an overstuffed sealed alignment) until I removed the foam from the first 2” and tapered the foam back at an angle. It appeared the foam was impeding the rear pressure wave, even though the center 3” of the enclosure was open.

Even with the scalloped baffle, the midrange operation was impeded by the proximity of the foam and enclosure walls. As you point out, the B&W enclosures (or a spherical enclosure) exhibit some positive attributes both inside and outside of the enclosure, and may be the optimal ‘minimalist baffle’.

C

[bluewizard]

For what it is worth, the example B&W Signature Diamond speakers have received "Stereophile's" Best of the Best Award.

Just passing it along.

Steve/bluewizard