How important is a roundover on the top and bottom?

[spasticteapot]

I know that a nice big roundover on the sides of a speaker cabinet is important to reduce edge diffraction and the high-frequency response irregularities it causes, but what about on the top and bottom? Would some of these issues be abated by the use of a waveguide?

[charlielaub]

Re: How important is a roundover on the top and bottom?

Edge roundover (including top edge) reduces the magnitude of cabinet edge diffraction ripples, but does not eliminate them, nor does it change the baffle step. The effect is different depending on driver placement, driver diameter, and baffle dimensions. In general "more is better" in terms of the width of the roundover or chamfer, but there will likely still be some ripple to deal with (if you choose to do so).

For a picture of this, see this page:
http://www.trueaudio.com/st_diff1.htm
The main thrust of the page is the baffle step transition, but there are four figures in the middle that illustrate the response from "real" loudspeaker enclosures of various shapes and the "ripples" that I am referring to.

For the typical rectangular box, there is a usually a peak of 2-3dB in addition to the baffle step, somewhere in the vicinity of 800-2k Hz (shown at a much lower frequency, e.g. around 400 Hz, in these figures, however IMHO that is that would only result from a baffle that is quite wide, e.g. 30"). The ripples are worst when the driver has equal distance from top and sides, e.g. a 6 inch driver mounted 8" down from the top in a 10" wide baffle. Locating the driver so that it has different distances away from nearby edges is helpful. Larger drivers will have slightly less ripple than a smaller diameter driver at the same position.

It's best to model these effects using programs such as Jeff Bagby's "Baffle Diffraction and Boundary Simulator" (for single drivers) or "The Edge" (for multiple drivers of the same diameter) or the "Baffle Diffraction Simulator" (multiple drivers of different diameter). This is useful even before you build the cabinet to plan for response irregularities. Then measure after you have the drivers in the cabinet to double check the response before making up your final crossover.

Finally, keeping the width of the baffle as narrow as possible will help to keep the first "bump" of the ripples as high in frequency as possible. This may provide the following advantage: If the driver's off-axis response has decreased appreciably by the frequency of the "bump", the "bump" amplitude may be reduced significantly if there is little to no sound pressure radiated in the plane of the baffle by the driver. It depends on the driver so this is not a rule of thumb. It's best to measure the actual driver in the actual enclosure and go from there.

I would bet that a waveguide will actually illuminate the edge more strongly than a flush-mounted driver because the wavefront is "guided" through the waveguide into a gently opening "space", including along the edge of the waveguide and outward radially towards the baffle edge. This would maximize diffraction related response irregularities. Maybe someone else can comment on this?

-Charlie

[marvin]

Re: How important is a roundover on the top and bottom?

Audio-wise, others can answer, but it complicates veneering big time, not to mention designing a nice looking grill. If you paint the baffle like PE's premade enclosures, then its not an issue.

Sound quality or aesthetics

Marv

[Zilch]

Re: How important is a roundover on the top and bottom?

Just beginning some experiments with the prominent "edges" common to recessed baffles and grille frames, here in the PE KD trapezoidal enclosure kit. Orange is untreated:

http://www.audiokarma.org/forums/sho...52#post3266252

[badman]

Re: How important is a roundover on the top and bottom?

I've seen this before- the application of half-round to a surface. This actually creates additional discontinuity and diffraction, not less. It's only less if it softens the transition from 2pi to 4pi (baffle edge) or from horn contour to free-air (or baffle).

The top, in most speakers, is less important than the sides. Amongst other things, it's usually a smaller edge. It's still important to try to do the largest bevels or roundovers possible, wherever you can. Bevels are slightly worse than roundovers but not much, and much better than hard edges.

Use of foam or felt on the baffle can help a lot. And the bigger the roundover, the more effective it is. The smaller the baffle, the less important, but it's still best practice. Flush mounting of all drivers, likewise.

[spasticteapot]

Re: How important is a roundover on the top and bottom?

Huh.

I'm really very confused now - am I better off with hard edges or rounded edges? What about really big (1") roundovers?

Also, has anyone actually measured what happens if you put a driver in a spherical enclosure? B&W does it, and it seems to work well - and they don't even need any felt!

[charlielaub]

Re: How important is a roundover on the top and bottom?

Huh.

I'm really very confused now - am I better off with hard edges or rounded edges? What about really big (1") roundovers?

Also, has anyone actually measured what happens if you put a driver in a spherical enclosure? B&W does it, and it seems to work well - and they don't even need any felt!

All of the above (sharp edges, 1" roundovers, sphere) are OK and will present their own challenges and benefits. No matter how you slice it, you will need to compensate for diffraction in the crossover some how. There is no "best way" just like there is no "best driver".

-Charlie

[spasticteapot]

Re: How important is a roundover on the top and bottom?

Can you help me understand the pros and cons of using sharp edges, roundovers, and spheres? Especially with the use of of a tweeter waveguide?

[Zilch]

Re: How important is a roundover on the top and bottom?

Nobody's biting on this one? :p

Perhaps a couple of quotes will stir the pot a bit:

Incidentally, the disk circumference had a full 3/4" rounded edge. So much for rounded edges eliminating diffraction. Diffraction is caused by the rapid expansion of the wave front into the larger volume beyond the disk's edge. For rounded edges to be effective in reducing diffraction, the edge radius must be comparable to a wavelength. For a 3/4" round this corresponds to a frequency of 18 kHz.

Over the years manufacturers and amateur builders alike have produced cabinets designed to defeat edge diffraction with bevels anywhere from 3/4" roundovers to large 3 - 6" straight, compound, and curved bevel shapes.

Diffraction has always seemed to me as being touted as more of a "boogie man" than reality would indicate. I have frequently commented when asked about the importance of diffraction that "the diffraction caused by cabinet edges and baffle protrusions is probably at least as hearable as the diffraction caused by the vase your wife or girlfriend put on top of your speaker, which is to say, not at all."

[MrkCrwly]

Re: How important is a roundover on the top and bottom?

Nobody's biting on this one? :p

Perhaps a couple of quotes will stir the pot a bit:

I assume these quotes refer only to edge diffraction issues and not to be construed to also include the baffle size as being irrelevant? (Or should we just use brass screws.:D)

[Zilch]

Re: How important is a roundover on the top and bottom?

I forget which, but one of them says flush-mounting drivers doesn't do anything, either.

[Geddes might well offer up, "Clueless; HOMs RULE...!" ;) ]

[charlielaub]

Re: How important is a roundover on the top and bottom?

Nobody's biting on this one? :p

Perhaps a couple of quotes will stir the pot a bit:

Sure, bite on this. The following statement:

"the diffraction caused by cabinet edges and baffle protrusions is probably at least as hearable as the diffraction caused by the vase your wife or girlfriend put on top of your speaker, which is to say, not at all."

Is a load of crap and involves excessive generalization.

-Charlie

[spasticteapot]

Re: How important is a roundover on the top and bottom?

Sure, bite on this. The following statement:

"the diffraction caused by cabinet edges and baffle protrusions is probably at least as hearable as the diffraction caused by the vase your wife or girlfriend put on top of your speaker, which is to say, not at all."

Is a load of crap and involves excessive generalization.

-Charlie

You won't mind my asking, then, if you can back this up with some hard facts?

[dlr]

Re: How important is a roundover on the top and bottom?

I would bet that a waveguide will actually illuminate the edge more strongly than a flush-mounted driver because the wavefront is "guided" through the waveguide into a gently opening "space", including along the edge of the waveguide and outward radially towards the baffle edge. This would maximize diffraction related response irregularities. Maybe someone else can comment on this?

-Charlie

A horn (the example here, as opposed to an OS waveguide ref. Geddes) should in most cases greatly reduce the edge illumination except for whatever range it's effectiveness is compromised. That is to say, horns have significant diffraction in and of themselves, so the pattern at the edges would have to be measured and compared to a large baffle measurement. A horns reason for being is to increase the sensitivity in the front hemisphere and to lesser extent control the directivity. The secondary impact is to increase the forward radiation at the expense of the lateral radiation. The latter is what will reduce the diffraction signature, above the baffle step primarily.

What I see in the measurements here is not so much edge diffraction control, it appears to be closer to making the "roundover" an extension of the horn itself due to close proximity. The horn has a drastic change in acoustic impedance at the mouth as well as having those diffraction-inducing four "corners" on the interior that muck up the response. Increase the horn mouth radius and its response will improve. I suspect that's what we have here.

I always refer use the term diffraction control rather than elimination. The only baffle that can eliminate diffraction entirely is an infinite baffle. Any geometry that has any non-flat area (such as another driver for instance) will induce diffraction. The infinite baffle simply allows for a constant, smooth and undisturbed expansion of the wavefront.

The control of diffraction is just an attempt to reduce its impact by damping to the extent possible (.e.g. felt) and in dispersing it across the spectrum in the frequency domain (roundovers). As pointed out, the baffle step, biggest part of diffraction, cannot be eliminated nor can the full step be avoided in any box speaker. Minimizing it's impact is the limit whether with direct radiators or horns and waveguides, the latter possibly obviating the need for diffraction control.

As an aside, the single baffle/driver combo that could theoretically eliminate diffraction that is not an infinite baffle would be a bipole, not to be confused with a dipole. This would require, however, a very thin baffle and two drivers mounted back-to-back with baffle dimensions and driver location such that the diffraction signature above the step area does not occur due to the directivity of the driver at and above this transition point. The baffle edge thickness would also need to be very small, possibly almost a knife-edge in the ideal. In theory.

The audibility of any given diffraction signature will have a certain amount of uncertainty since it almost always tends toward the limits of perception at some point and the magnitude of the change depends on the amount of control attempted, i.e. roundover radius, felt qualities, thickness, driver position on the baffle, etc.

It's much easier, for instance, to test for audibility when using felt because the change can be made nearly instantaneously. Roundovers present a much more difficult test regimen, but there are certain characteristics that cannot be logically disputed, such as improved polar response that is easily measured and proved. These measured changes often fall into the area of well-documented levels of audibility, but since every case is different, baffle/driver/crossover/room, there's really no hard and fast rule that anyone can prove/dispute.

I know that I can hear differences between extensive baffle treatment and none. In between it gets increasingly difficult to make any absolute judgments. However, I can provide measurements to prove the changes occur. In the end it comes down to how far does one care to go to maximize the design response.

I believe that Dunlavy (felt treatment), Thiel (large roundovers), B&W (unique baffle shapes) and a number of others were successful in part due to their attention to detail that included diffraction control.

dlr

[charlielaub]

Re: How important is a roundover on the top and bottom?

You won't mind my asking, then, if you can back this up with some hard facts?

Read these:
http://www.linkwitzlab.com/diffraction.htm
http://www.linkwitzlab.com/faq.htm#Q8
http://www.zaphaudio.com/mtg-surface.html

-Charlie