The topic of baffle geometry, and its influence on the playback quality of loudspeakers, has already been described in the article "About baffle design, edge diffraction, secondary sound sources, ..."Detailed. If you are still in the belief that designing a baffle and the arrangement of the drivers on Selbiger is primarily a question of optics, may the article like (again) read.
In the process of finding a suitable baffle, I often ended up with the variant "slanted chamfering". Again and again, the question arose how steep these really have to be to be effective. So far I have relied on secondary knowledge in this regard ... Read in books, forums, magazines. When the question arose recently, I decided to investigate the matter metrologically.
To accomplish this, a baffle with adjustable, beveled chamfers had to be made. After a lot of headache, how to build this, I finally came to the following conclusion:
It was milled from behind a V-groove, just less deep than plate thickness, the whole thing with tape from the front "secured" and a 2mm strong copper wire festgetackert from above, which serves as an adjustment mechanism. The "wings" can be set as desired by 0 ° -55 °.
As a tweeter, I chose one that by itself shows no disturbances in the straightening behavior, in the area relevant for edge diffraction, which could falsify the results. The choice a lot on the very much appreciated by me XT25TG-30 / 04.
In 55 ° steps for chamfer angles of o-5 °, the respective radiation behavior of 0-90 ° in 10 ° steps was measured. Altogether 120 measurements.
The initial situation: Bevel angle 0 °
See also "Worst case" scenario
The left picture shows the radiation behavior in a sonogram normalized to 0 °. The right-hand example shows the angles 0, 30 and 60 °.
In the sonogram it can be seen that the loudspeaker at 3kHz is far too broad, and too narrow at it. Axis, and angular frequency responses are unbalanced to each other. Where there is a sink on the axis, there is an elevation under angles. The speaker changes its character with the listening position. Desirable would be a steady and evenly narrowing course.
Sonogram for chamfers with a slope of 0-55 ° in 5 ° steps
0,30 and 60 ° for bevel angles of 0-50 ° in 10 ° steps
The main disturbance (widening by 3kHz along with the constriction below) decreases with larger chamfer angle. Chamfers with a steepness <20 ° do not bring satisfactory results yet.
The chamfers themselves form a new edge, which comes to bear with increasing steepness. Recognizable by the arising expansion around 5Khz. This is less pronounced because the tweeter in this area is no longer working as a half-space radiator.
This is also one of the main tasks of the oblique chamfer, in addition to the resolution of the "rectangular geometry": to bring the case edges so close to the tweeter that it no longer "sees" them or only weakens them. Less distance ⇒ Smaller wavelength ≙ higher frequency ⇒ more directivity ⇒ less influence of the edge
The concept "oblique chamfer" works particularly well when the tweeter in the corresponding frequency band already bundles clearly and thus this is practically no longer "seen". Of the Seas DXT 27TBCD / GB be mentioned here by way of example. How well this driver works with oblique chamfers is an example DXT-Mon to see.
- Bevel <20 ° are not sufficiently effective
- The distance of the chamfers to the place of sound creation must be sufficiently low, otherwise one reaches only one, at least partial, displacement of the problem
- Depending on the situation, the chamfers can also zu steep be!
The concept of slanted chamfers is a very effective approach to the problem of secondary sound sources through edge diffraction. But not unconditionally. Slope and position must be chosen carefully and according to the concept. If you asked me for a universal recommendation I would say:
"Make the chamfer 25-30 ° steep and position it as close to the tweeter as possible"
Finally, the measurements in single images, for those who want to take a closer look. The steepness of the chamfer is in the lower left corner of the picture.
Note: The picture change is also possible with the mouse wheel