Date: Wed, 27 May 1998 15:12:23 -0700
To: bass@mcfeeley.cc.utexas.edu
Continuing with the free air data:
20"W One-sided Baffle, RD75 BG S/N..54 in free air:
The one-sided baffle is 2" wider than the two-sided baffle, and was kludged together from bits and pieces of other baffles. It is as close an approximation to the size of the two-sided baffle as could be created from existing materials. It is a 1.5" thick baffle. The baffle edge is treated with a 1/2 round section, and the speaker edge is left untreated with the mounting flange exposed.
Unfortunately, only the high-resolution low frequency measurement was
recorded at the 3m distance, and hence that data is only shown to
~650Hz. It is interesting to note the undesired smoothing that occurs
in the merged plots below 650Hz by comparison with the 3m data. When
the 0.5Hz resolution data is transfered into the 15Hz resolution data
domain significant detail is lost. What is gained is dramatically
improved accuracy in the low frequencies. Full-range plots (61KHz
sampling rate) with the IMP are rarely useful below about 100-150Hz
except when long duration sample times are analyzed!
20"W One-Sided baffle in free air at 1m, 2m, and 3m - unsmoothed responses.
20"W One-Sided baffle in free air at 1m, 2m, and 3m - 1/12 octave smoothed responses.
Analysis:
The 1m response shows a large rise in amplitude below 500Hz, which is seen to rapidly flatten as the distance from the baffle increases. The notable feature, the reduction in the low frequency rise compared with the two-sided baffle is best seen in the comparsion of the responses:
A most unusual feature shows itself in the comparison plots. Whereas the primary low frequency support provided by the baffles is seen to cause an upward 'hinge' in the frequency response from ~850Hz on down, the baffles are also seen to cause a downward 'hinge' in the frequency response from about 3000Hz to 850Hz.
There is no clear mechanism that I know of to explain this downward trend. Why would a rectangular baffle of the size of these baffles depress the output from 850 - 3000Hz? A notch also appears in the range 1-1.2Khz with the rectangular baffles at the distances measured.
The free air data convincingly demonstrates the when the RD75 driver is used in rectangular baffles of approximately the same surface area that a two-sided baffle with the RD75 centrally mounted produces an extremely large low frequency rise below 1000Hz. That rise is subject to the Proximity Effect, but is still a prominent feature at 3m - a common listening distance.
That the 11 two-sided rectangular and trapezoid baffles I built and tested last June/July all showed to various degrees this large low frequency rise led Rudi Blondia to immediate action in confirming the effect. His insightful analysis of the cause of the phenomenon was published at his website on September 7, 1997 as "Day 22".
Now that it is understood what to look for, the Virtual Diffraction Line Source signals Rudi first identified are readily seen in the impulse data from RD75 baffle measurements. The purpose of the labor intensive free air measurements was to dramatically demonstrate that "John's Bump" is a result of the baffle type and not 'floor mirroring' or some mixed effect.
In conclusion, maximizing the linear frequency response capability of the RD75 driver would seem to preclude the use of rectangular or trapezoid two-sided baffles.
Ground plane measurements of the rectangular baffles beyond 3m distance must await the next availability of the Gym measuring venue. Next time, ground plane comparisons to the free air data.
John Whittaker
Dipole Baffle Study Report#18.
Dipole Baffle Study Report#20.
The RD75 Dipole Baffle Study - Table
of Contents
Acoustic Line Source Research -
Table of Contents.
