Bds

[rinx]

Just playing right now with Baffle difr. Simul. and got some "results", but i dont know how to read those results, are they "good" or "not so good". So i will leave link, where is printed screen of results.

Is there everything OK?!



http://www.album.ee/node/11597968/50044117
http://www.album.ee/node/11597954/50044117

[Shaun]

You need to optimise for the smoothest response in the pass band. With a rectangular square-edged baffle you can get good on-axis reponse, but peaking and dipping of-axis. The best is to use a cabinet that has minimal diffraction (large-radius edges, 45 degree beveled edges, spherical enclosures), so off-axis response is also much smoother...

[Ron E]

On a narrow baffle like you have there aren't many options, and narrow baffles tend to minimize diffraction problems.. On a wider baffle you can optimize tweeter location with the BDS to smooth the response.

The graph simply shows the predicted response for each driver. The baffle step is a change from 4pi to 2pi radiation as you go up in frequency. The reradiation of sound by the baffle edges causes the step and the ripples in the response, usually characterized by a peak and then a dip at a frequency related to the baffle dimension.

[Shaun]

Baffle step and edge diffraction show up on the same plot, here and in acoustic measurements. But edge diffraction is caused by radiation off the edge, while baffle step is as a result of the 4pi to 2pi transition. For instance, mounting your drivers in a sperical enclosure will yield very little edge diffraction, but the baffle step effect will still occur.

A horn-loaded tweeter will be less affected by diffraction off the box's edge.

[Ron E]

Originally posted by Shaun
For instance, mounting your drivers in a sperical enclosure will yield very little edge diffraction, but the baffle step effect will still occur.

Your description is not correct, and not just the quote above.

The 4pi to 2pi transition step is related to baffle size (distance from source to various edges), and the ripples above the step are caused by the same thing, but the high frequency region shows the effects of various edge treatments. A curved edge acts like a continuous edge and thus spreads the "edge" diffraction over a wider time (frequency) range, smoothing the response. A spherical (sic) enclosure also acts like a "continuous edge".

An infinite baffle is better than a sphere, because it has no edges (curved or otherwise).

--------------------------------------------------------

Interesting that Mr. Verdone is billing the BDS as solely his effort, and appears to be offering to sell licenses to FRD software....

[Shaun]

Originally posted by Ron E
An infinite baffle is better than a sphere, because it has no edges (curved or otherwise).

I don't understand why you say that the description is not correct. I merely meant to point out that diffraction and baffle step have two different causes. Yes, they are both affected by baffle size, but for different reasons.

In the case of edge diffraction it really is the proximity of the edge (a discontinuity in the surface) to the transducer, and the wavelength of the signal that interacts.

In an infinite baffle there is only radiation into 2pi space, so no transition occurs, thus no step. Basically, in the 2pi region the transducer's radiated energy is limited to the forward direction. In the 4 pi region that same amount of energy (relatively) is also radiated (shared/lost) into the rearward hemisphere. So there is a 6dB loss. This is baffle step. It is the transitional area between the two that is affected by baffle shape (including that of the "edge"). The discontinuity in the surface causes sound to be re-radiated, thus causing those wavelength-dependent ripples.

I agree with how you described the effect of the shape of the edge. But then again, there's not much to disagree on...

[Ron E]

The way a baffle diffraction simulator works is to calculate the distance to an edge, calculate the geometric time delay of the reradiation of sound from the "edge" and sum it with the direct sound. That is all baffle diffraction is. The "step" and the "ripples from the edge treatment" all come from the same thing. It may make more sense to you to separate the effects, but they are not separate.

The BDS uses ~24-36 rays IIRC (I forget how many edges a roundover is broken into) - although many more rays don't get you much for the extra computation time, Excel is slow. It's a neat application, but a rather unintuitive GUI, IMO.

[john k...]

Originally posted by Ron E




Interesting that Mr. Verdone is billing the BDS as solely his effort, and appears to be offering to sell licenses to FRD software....

Hi Ron,
Not to worry;

"A very special thanks to three: again to Ron Ennenga (who also provided his troubleshooting, a great deal of insight and a detailed Pattern Lobbing Spreadsheet Example he authored based on the Derivations of) Dr. Arthur Ludwig for the mathematical basis of the Directional Models, and to John Kreskovsky who provided the underlying conceptual approach to solving Curved Edge Approximations. John specifically lead me through most of the unknown and rarely understood parts of diffraction, freeing me from earlier incorrect pre-conceptions, and focusing me toward the solving of the difficult solutions with easy to comprehend examples and clear methodology. John has always supported and fueled creative process in Acoustic Design with a precise approach atop an unconcealable enthusiasm to advance the deductive process, which makes an extremely inspiring teacher, indeed."

This is just one of the extensive acknowledgements in the BDS manual.

Additionally, Paul placed the licensing comments on all the tools at the FRDC to help protect the interests of the many developers. It was not intended for personal gain.

[jnb]

rinx,

The first thing I'd do to get a handle on this procedure is to move the tweeter sideways in several steps and watch its response above 1k. Making it flat indicates a reduction in interference from the baffle (or more accurately a desirable combination of interactions, which is what using this software is all about). I happen to think the tweeter in this case is in the greatest need of optimising.

Consider the response of the mids and woofers compared to the crossover you wish to put in place. The slight hump at around 1kHz might not necessarily be a bad thing, if you happen to like a contoured response. You may prefer the result of moving the upper mid away from the top of the enclosure just a little, or extending the enclosure a little, or just fixing it with your crossover, but that's another story.

[Ron E]

Originally posted by john k...
Hi Ron,
Not to worry;

Well, that appears to be settled then. For the record, YOUR apparent exclusion, John, (among others) was as irksome as any other to me, however self-motivated my comments may have seemed. You had a LOT to do with many of the tools at the FRD Consortium. I regret not contacting Paul about the licensing "issue" before posting something. I enjoyed contributing and never expected any gain, but I hope folks can understand the annoyance at the appearance that someone was seeking personal gain from a group effort.

rinx,
Sorry for stealing part of your thread, it was not my original intent. "jnb" has a good suggestion. Aside from his suggestion about experimenting with tweeter placement, there are a lot of documents available showing you how to use the FRD tools, one of which is the Speaker Workshop manual at http://www.audiodiycentral.com/.

A hobbyist named Roman Bednarek also has a guide, IIRC, although I have never read it.