Jepp - har prøvd og feilet med 8", 10" og 15". Forholdene rundt en vilkårlig U-baffel er akustisk uforutsigbar og direktelyden blander seg med alle former for delays, diffraksjoner og resonanser. Resultatet er oftest noe som ikke er kardioid spredning i store deler av frekvensområdet.
Jeg var tidligere i dialog med Sigfried Linkwitz og John K. i Music And Design (som har testet U-baffel i mellomtonen og gitt opp) om temaet. Nøyer meg med å gjengi utdrag fra epostdialog med Linkwitz:
SL: You raise difficult problems with trying to build a full-range cardioid. My question is what benefits do you expect from a cardioid in a reverberant environment as compared to a dipole? The lack of rear radiation only matters for the direct signal, allowing to place the speaker closer to the wall behind it. But the direct to reverberant ratio in the room is the same as with a dipole.
Meg:In your excellent articles "Issues in loudspeaker design" you mention use of sintered metal mesh and feltmetal for making frequency independent Acoustic resistors that can take high SPL. For the mids (>200Hz) my initial idea was to use damped U-baffles where the baffle would control dispersion up to a point where the driver's natural dispersion would take over. Damping the baffle internally would attenuate quarter wave resonances to an acceptable level (I believe). It was also the intention to use non-woven sintered metal mesh to terminate the back of the U-baffle (which also will be used for U-frame woofers). I am thinking 12 or 10"-er for the mid. These would meet the capacity of the woofer system.
SL:Sounds like a good idea, but you will have to find out how well the practical solution can work by making acoustic measurements. You are playing with rear wave delay, cavity resonance, attenuation, cabinet edge diffraction and vector addition in space. DSP is of limited help.
Meg: Also, do you know of any software or spreadsheet that could simulate the exact behavior of such U-baffles?
SL: No. The attenuated U-frame is a messy distributed structure. Among other things, if the absorbing material has friction losses you get non-linear distortion.
Meg: I am skeptical to implement active cardioid control, by means of a backwards radiating driver for frequencies above 4-500Hz for several reasons. I know of other solutions for cardioid dispersion like the Gradient Helsinki and Amphion's. A perforated sidewall utilizing an acoustic resistor should allow for broadband system, but again at the cost of major sensitivity/capacity loss. Therefore, do you know if there are other ways to achieve cardioid dispersion in the ~200-1.5kHz range ?
SL: No
SL avslutter med "I have never seen a respectable full range true-cardioid design. I am sure for good reasons. respectable dipoles are a lot easier to build but still rare."
Hvis du er ute etter et grunnforskningsprosjekt så er U-baffel i mellomtonene veien å gå
Det er nok løselig, men krever mye eksperimentering. Derimot, hvis du ønsker en roligere hverdag så ville jeg heller skult til kardioider med ventilert, dempede sider (eller dipol for den slags skyld). Diyaudio.com har noen hardcore entusiaster som har fått det til med beskjedne midler.