The dipole baffle concept tested is a stand alone trapezoid. No provision for woofer integration with the baffle structure is made or contemplated. It is believed that any extraneous structure added to the baffle would deteriorate the aerodynamic purity of the concept and thereby deteriorate the linearity of the frequency response. An external woofer will be needed with this concept.
Furthermore, in a desire to maximize the 'flooring mirroring' potential of the RD-75 driver, it will rest on the floor itself to obtain the minimum gap distance between the planar membrane and the floor. The actual distance between the bottom of the membrane and the floor is measured at 2 5/8 inches. The baffle thickness will be that of the driver, approximately 1.56 inches.
A starting rectangular width for the prototype baffle was chosen based upon previous testing of the low frequency response of the RD-75 driver with several different baffles widths. That information is available in the Outdoor Testing section. Since the test baffle would become the basis for an interim stereo system to last for approximately one year, aesthetic considerations in addition to pragmatic ones entered into the decision set.
U.S. measurement Clear Douglas Fir (CDF) boards of nominal 1" x 8" x 75.6" (actual height) became the basis for the rectangular front section of the baffle. CDF is a categorized a softwood, is very straight and strong, and easily worked. It is also relatively expensive and can be marred easily. The width and thickness of the boards was approximately 7.125" x 0.75". Combined with the 3.5" frontal width of the RD-75 when mounted to the boards gave a nominal width of 17.75" to the basic structure. Compared to typical hardwoods the CDF has a reduced resonance signature. Clearly other materials would be suitable for the baffle boards.
For aesthetic reasons the CDF was carefully sanded and bleached three times with wood bleach and then finished with four coats of water-based polyurethane. The resulting appearance is white with darker golden wood grain patterning. This contrasts dramatically with the black finish of the RD-75.
The support base was created from four 10"(L) x 10"(L) x 1.25"(W) x 0.25"(T) 'L' brackets mounted back-to-back creating a 'T' structure. These brackets were not perfect right angles when purchased and extensive work was required to straighten them - extensive! An initial attempt to glue them together with common '5-minute' epoxy failed. West Marine brand epoxy with filler material succeeded in bonding them. It is felt that with care this is a workable solution to bonding the pieces, but in all probability a large shock would result in separation of the pieces.
Each face of an L bracket had 5 screw holes. Those holes on the horizontal section (that part resting on the floor) of the 'T' were also filled with the epoxy/filler mix for aesthetic reasons. The two assemblies (for each speaker) were then painted with a satin-finish epoxy black paint. A notch, approximately 0.25" x 1.25" was cut in the outer bottom edge of each baffle board to accommodate the width and thickness of the bracket. The notch required a slant towards the rear to accommodate the right angle bend of the bracket.
The 'T' brackets were then attached to the rear of the baffle boards with 5 'sheet metal' type screws of 1" length and 3/16" diameter. It is felt that the 'sheet metal' screw was a better choice than regular wood screws for this application.
The RD-75 is of 'clam-shell' type construction, and has a flange on either side with 5 screw holes evenly spaced along the length as mounting holes. The distance from the front face of the flange to the front of the 'horn' assembly of the driver is slightly more than the width of the baffle board described. Flush mounting of the driver with the baffle boards requires further detail.
It had been noticed in previous testing that the RD-75, with it's 35.5 pound weight, and a moving membrane in the several gram range, still had vibration in it's structure during loud passages of music. Hence, a rubber gasket mounting of the driver was seen to attend to both the issue of flush mounting of the driver with the baffle face and to the absorption of vibration from the driver. A 3/16" (T) x 3/4"(W) closed-cell synthetic rubber self-adhesive (weather stripping) gasket was applied to the back inside edges of the baffle boards over the full 75.6" height.
A 1" x 4" Teflon coated floor slide designed for furniture use was obtained and cut to 3" width and affixed to the bottom of the RD-75 across the clamshell seam with a 'superglue'. It was felt that the 35.5 pound mass of the RD-75 driver resting on a 1" x 3" platform would sufficiently couple the driver to the floor to realize whatever 'mechanical diode' effect was available on a hardwood floor.
Two layers of self-adhesive black felt were cut to size and installed on the bottoms of the 'T' brackets and the bottom edges of the baffle boards to approximate the thickness of the Teflon floor slide. Pilot holes 1/16" in diameter were drilled in the baffles boards with same spacing as the RD-75 mounting holes. The RD-75 was screwed to the baffle boards using 1" 'Dry Wall' screws in such a manner as to compress the gasket material only so much as to achieve a flush mounting with the front face of the baffle boards. A straight edge was placed at many positions over the full height of the assembly to insure the flush mounting accuracy. During the mounting of the baffle boards the full weight of the RD-75 rested on the Teflon slide.
At this point in construction precision measurements were started in order to document and evaluate the potential of each design change or feature. Certain features were shown to be detrimental to frequency response linearity and were not pursued.
The baffle design at this stage is physically vulnerable, and although it is felt that it would survive a 7.1 Richter earthquake on the Inglewood-Newport fault (1/2 mile away), it surely would not survive a 3 year-old running headlong into it. After all, it is a prototype baffle.
At this point in construction all large furniture in the living room was moved to suitable wall positions and small items were moved out of the room into temporary storage. The area carpet was rolled and moved against a floor-wall junction. This produced a clear and open space with the ground-plane as a hardwood floor. More than 90 minutes was spent seeking the most optimal positions for microphone and baffle in the room. The largest distance/time achievable was 2 meters from microphone to baffle without any reflections in the MLS test signal.
Two baffles had been constructed and an initial frequency response measurement was made on both driver/baffle combinations to define a starting point for documenting the baffle construction method. From this point on the measuring environment and the speaker and microphone positions stayed unchanged for three weeks until the end of the study.
Both baffles were then disassembled and polyester 'chifon' fabric was stretched tightly over the mouth of the RD-75 'horn section' using double-sided adhesive carpet tape around the flanges and across the top and bottom front plates. This technique for adding acoustic damping to a driver is well known, and had been previously tested and judged to be beneficial in increasing the linearity of the frequency response. As Rudi Blondia has pointed out, it does no harm and will keep bugs out of the driver cavity.
It is well known that large loudspeaker cabinet panels have a tendency to produce acoustic output when excited by their driving loudspeaker. Hence there is much interest in bracing/damping cabinets. With such large surface areas involved with the dipole baffle panels an extra measure of concern over the possibility of extraneous sound being produced by the baffle panels is justified. A product had previously come to our attention in a woodworking supply catalog, a so-called 'Router Pad', that seemed to be adaptable to the cause of damping the baffle panels. This material was claimed to be high density urethane with a 'special treatment' of elastomers to insure longevity and 'grip' (for router table use). Two 24" x 72" x 3/8" sheets were ordered. The material felt greasy to the touch and seemed to leave an oil-like film on the hands, but looked appallingly like under-carpet padding.
A trip to the local Home Depot confirmed the general appearance of the router pad to under-carpet padding, but it was noted that the undercarpet padding came in 3, 5, and 8 pound weights (per some unknown unit volume?) and was definitely not greasy feeling. An estimate is that the 'router pad' material is of '10-12 pound weight'. Later, an 'eggcrate' foam bedding pad was purchased for the floor damping portion of the study, it also has a similar 'greasy' feeling and leaves residue on the hands. It's only special claim was for fire resistance.
Both router pads were glued together with 3M brand No.77 spray adhesive (for urethane); and, then cut into four 6" x 72" x 3/4" sections. Each section was glued to a baffle board flush to the outer edge, and centered on the 75.6" baffle board with Dow Corning brand 'Performance Plus' silicone sealant. A full covering coat of sealant was spread over a side of the router pad and then pressed onto the baffle board and allowed to dry.
A gap of approximately 1.25" between the edge of the router pad and the back edge of the RD-75 driver remained on either side. Furthermore, there was the problem of how to obtain a smooth surface over the router pad in addition to covering the gap between driver and foam. In regards the issue of baffle diffraction, a general rule seems to be that any discontinuity in the free expansion of sound waves from a loudspeaker can be detrimental to frequency response linearity. The dilemma: surface of the router pad is pocked with shallow holes, and there is the gap between router pad and driver, but the thickness of the baffle boards with the router pad attached now approximated the thickness of the RD-75.
A trip to the local office supply store resulted in the purchase of sufficient black poster board material to cover the back of the baffles. This material is more very thick and heavy paper than cardboard. From the local art supply store a quantity of acid-free paper artist's tape was obtained in both black and white, and in several widths. A number of 5.75" wide sections of the poster board were cut to size and glued to the router pad sections with the above mentioned silicone adhesive using full coverage, and allowed to dry. A 0.25" section of the router pad was left uncovered closest to the RD-75 driver.
All seams between the paper sections were sealed with the artist's tape and the taped areas were coated with artist's paper glue. A number of lengths of 2" wide poster board were taped to both the RD-75 and the 0.25" strip of uncovered router pad after filling the gap space with various damping materials at hand. All seams were taped; and four coats of water-based polyurethane were applied to the poster board to provide a surface that removable tape would be applied to without tearing the poster board upon removal.
The resulting surface is relatively smooth and flush with the back face of the RD-75. If the project were redone the poster board would be sealed first with the polyurethane before gluing to the router pad. There is an inconsequential 'lumpiness' in the poster board assembled as described. At this point an addition frequency response measurement on both baffles was made.
Sufficient 1.5" x 0.75" 'Half-Round' moulding stock had to be special ordered from a local lumberyard to span all anticipated baffle-edge spans of the various baffle-shape conditions. The moulding was finished with four coats of water-based polyurethane and was affixed with double-sided tape to the outer edges of the baffles to reduce the effects of edge diffraction. The dipole baffle described to this point is very non-resonant.
The rectangular baffles that resulted from applying the half-round moulding to the edges of the thickened baffles were measured for frequency response at this point.
A modification that was not kept in the final design was as follows. The top and bottom of the RD-75 driver have recessed areas compared to the front surface of their 'horn mouths'. When the horn mouths are mounted flush with the surface of the baffle boards, a top and bottom cavity approximately 0.5" deep x 3.5" x 2.5" is created by the recessed areas of the drivers. Filling those areas in, flush with the surface of the baffle boards was found to be a detrimental to the linearity of the frequency response and was not continued in the further tests.
A sufficient amount of 0.75" thick expanded styrofoam insulation material was obtained and glued together with Aline's Tacky Glue to form a composite panel 1.5" thick. Rectangular lengths of the styrofoam 75.6" long were cut into 9", 6" and 3" widths; and, then sliced on the diagonal to create right triangles with base widths of 3, 6 and 9 inches. These were the basis for creating and varying the trapezoid baffle shapes.
In the production of the various baffle shapes double-sided carpet tape was used to affix the styrofoam sections to the baffle edges and to the half-round mouldings. All exposed seams, both front and rear, were carefully covered with artist's tape; and, width measurements were taken on each baffle shape.
An acknowledged criticism of this study is that the baffle shapes obtained do not represent equal area baffles. It is easily seen that using this same technique with narrower main baffle sections, that the styrofoam test sections could be cut so as to preserve equal areas of the different baffle shapes. Unfortunately, this idea did not occur to us prior to finishing the main wood sections, and by that time it would have been too costly in time and effort to start over.
On the other hand, given the results of the study it is not clear that preserving equal area might be an issue of concern, since for the 2 meter measurements made there were relatively small differences seen below 1000Hz amongst all the baffle shapes, large or small.
How to purchase RD75 drivers.
The New Test Baffle Jig.
The RD75 Dipole Baffle Study - Table of Contents.
Acoustic Line Source Research - Table of Contents.
