The RD75 Dipole Baffle Study began on May 17, 1997 shortly after I purchased a pair of RD75 planar-magnetic, line-source drivers from the Bohlender-Graebener Corporation (BG). BG had offered it's line of planar-magnetic drivers, 'ribbons', to members of the Bass/DIY Internet ListServe at substantial discount. A number of List members also took advantage of the generous offer.
The RD75 planar-magnetic, line-source loudspeaker driver is one of, if not, the largest of its type, measuring 75.6 inches tall and weighing-in at about 37 pounds. It's 'ribbon' is 2" x 72" in size; and, the driver is rated for use with a high pass crossover as low as 125Hz. The RD75 is as close to a full-range 'ribbon' as has been commericially offered to date with it's upper frequency range (-3dB) extending beyond 16,000Hz.
Early in May of 1997, Rudi Blondia wrote to the Bass List asking if there were any individuals in the Los Angeles area who had measuring equipment, as he wanted to measure his new RD75 and Carver 60" ribbons. I owned a Liberty Instruments IMP/MLS Fast Fourier Transform measurement system and had a large open backyard, so I responded to him and we began.
My interest in developing an optimal dipole baffle for the RD75 - a baffle with the most linear frequency response - pre-existed our work together. Though with Rudi's passion, productivity and expertise as a model, things have been more 'X-treme' and gone much farther than I would have predicted.
Being under the influence of then-existing dipole baffle theory in the limited literature available, and being unfamiliar with the characteristics of the RD75, my first foray into formal baffle testing took place over a three weekend series in late June and early July 1997. I'd converted my listening room into a test chamber and measured at 2m distance and 1m height 14 different baffle varients based upon a 'two-sided' symmetrical and/or trapezoidal concept. Data from those dates can be seen under the Listening Room Measurements section, the 'Two-Sided Baffles' sub-section.
The outcome of that event was troublesome in that all of the baffles appeared to have a large low frequency 'bump' in the frequency response centered at approximately 300Hz or so. Sharing my data with Rudi Blondia led to a sequence of events quite extraordinary. Although we had reference to a similiar phenomena described by Newman (see the References), Rudi soon discovered that the 'bump' was in existance even with the dipole baffles suspended in air. This discounted the theory that the 'bump' was resulted from a 'floor mirror' effect.
Rudi's analysis of the source of the 'bump', and his "elegant solution" to mitigating the problem is on his website at: Rudi's website: 'Day 22'. (Please remember that if you leave this page and go to Rudi's website, that you can return here using the 'Go Back' command in your browser). He identifies the source of the energy for the 'bump' as coming from "Virtual Diffraction Line Sources" at the dipole baffle's edges.
His 'Day 22' analysis led me to cannabalize my existing baffle prototyping system to create a series of 5 new 'One-Sided' dipole baffles in September 1997. One-sided baffles are those in which the RD75 line-source driver is mounted at the extreme edge of a baffle - the unused mounting flange on the 'free edge' of the driver is then finished as a half-round section. A 'one-sided' baffle reduces the energy from the Virtual Diffraction Line Sources and reduces non-linearity in the frequency response.
Having the new baffles to test, a new test measurement setup was developed for the Listening Room which allowed for the acquistion of relatively clean, semi-anechoic, data at distances of 1m, 2m, 3m, and 3.5m. This was important to more closely match typical listening distances. The data acquired is presented in the 'Listening Room Measurements' section, under the sub-section 'One-Sided Baffles'.
One of the 'One-Sided' baffles created from the earlier construction proved to be very good, and was used as a demonstration piece at the 1997 SoCal Bass List Social. It was clear though that any future research was going to require a new baffle test system. A metal frame design was conceived and efforts were made to obtain its construction to no avail. At about the same time, a local high school's gymnasium became available for testing use though the good services of my good friend John Nipps. This was as close to an ideal venue for testing as might be hoped for - much superior to my backyard, with it's potential wind problems.
Since getting the new metal baffle test jig built turned into a bust and the crash program to built a wooden alternative was still in progress when the planned first date for measurement in the gymnasium arrived on January 2, 1998, I went with what I had on hand - two of the older 'one-sided' baffles. Data from that date can be seen in the 'Gymnasium Measurements' section.
The gym's size proved to be a great boon to obtaining clean semi-anechoic measurements. For the first time measurements out to 7m distance and heights of 0.75m, 1.0m and 1.25m were obtained with great resolution and little, if any, reflection problems. As a result of obtaining this type of data, a new phenomena was identified, that of the 'Moving Dips' in the frequency response of a baffle. Apparently the sound energy from the Virtual Diffraction Line Sources at the baffle's edges will combine with the direct sound at the various heights and distances to produce dips in the frequency response. The Moving Dips are seen to have a very regular pattern across distance and height.
A second trip back to the gym was soon accomplished on March 1, 1998. This time Rudi Blondia came, with a RD75 baffle and some woofers. Much of the time spent during that session was devoted to woofer integration issues with RD75 dipole baffles, but a new baffle was also tested using the newly constructed baffle test jig.
An added goal now, given the continued availability of the gym, is to acquire a database of measurements on 'Equal Area' dipole baffles. To date, three baffles (two 'equal area') have been measured at distances to 8m and heights of 0.75m, 1.0m and 1.25m. The project continues: new baffles are in construction; active filter electronics have been upgraded. It is hoped that a complete design of active filter/equalization for the RD75 can be offered in the future. Computer modeling of the dipole baffles frequency responses is also be explored.
John Whittaker
March 25, 1998