Hello all,


I only recently discovered this great site and forum and spent the last two weeks browsing through quite some threads (until my wife and kids started feeling neglected and made me stop...). Some of you guys out here are pretty serious, indeed!


Quick introduction:
I have been an avid musician for all my life (playing the piano and organ) and I also listen to a lot of recorded music of all periods and cultures, with Bruckner and Mahler having for long been the main staple of my daily diet.
I became interested in pro audio technology in the mid 1980s when I did a six-month internship at a large public broadcasting station.

In 1989 I bought a pair of 4430 monitors, cheaply, from the above mentioned station, and this eventually got me into the rewarding hobby of DIY loudspeaker building.


My approach on building loudspeakers has, for the last ten-odd years, been centered around three-way systems consisting of 15in / 10in / 1in + 90deg. CD-horn, electronically crossed at ~300Hz and hi-level XO / CD-EQ at ~ 1500Hz. (i.e. DIY 4430s with add. dedicated mid-driver and adjusted mid / hi XO)


Current (since 2006, I think) driver compliment: Beyma SM115/N in a SBB4ish 200 l. (sorry, i'm a metric guy...) 30Hz enclosure; Beyma 10Mi100 in a 20 l. sealed sub-enclosure; Beyma Cp350Ti on Beyma TD194 (bought last year) horn on top (no, I have no affiliation with Acoustica Beyma, but there's this guy, friend of mine, you know, who usually gets... ah, never mind...)


Now, to the subject of my post:


There is one thing I find particularly sexy about the Bob Smith horn (besides it beeing comparatively easy to build), and that is the narrow vertical dimension. This would allow for, given the 10in midrange and 1500Hz XO of my speakers, to get the vertical spacing of the acoustic centers (mid and hi, that is) to approach lambda/2 at f(xo), which should, taking Linkwitz' expression of [arcsin (lambda/2d)] as an approximation, result in a neat vertical lobe and an acoustical point-source behaviour around f(xo).


However... the Bob Smith design and its commercial JBL adaptation are simplified multicell exponential horns, and I am a „constant directivity“ guy and a big admirer of Keele's papers on the subject.


So, this is what I have been thinking about for the last year (without actually trying it yet): What do you think would happen, if one threw out the distributing vanes of the Smith design and instead would incorporate some constant directivity contour to the sidewalls (simply a circular arc joining the throat with 90deg conical „straight“ sections [as suggested by Hughes] and some secondary flange at the mouth (out to 150deg. or so) to reduce reflections?


I know, it's a bit weird to mate a design that tries to avoid diffraction at all costs (like Hughes') with another, that takes diffraction as an integral part of its functioning principle (Smith's), and also Keele's math doesn't really allow for a vertical expansion rate of zero, but the idea somehow intrigues me....


Maby one of the many guys here that are more knowldgeable than myself can set me straight on this one?


Anyhow, it's great to be joining this fascinating forum and I hope that time (read: wife, kids, job, thesis...) will permit for my regular participation.


Thanks for having me,
best regards,
Chris