Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10 5/3/83; site spanky.UUCP
Path: utzoo!linus!decvax!harpo!gummo!whuxlb!pyuxll!eisx!npoiv!npois!hogpc!houxm!hocda!spanky!spj
From: spj@spanky.UUCP
Newsgroups: net.audio
Subject: Loudspeakers
Message-ID: <428@spanky.UUCP>
Date: Wed, 10-Aug-83 14:02:23 EDT
Article-I.D.: spanky.428
Posted: Wed Aug 10 14:02:23 1983
Date-Received: Thu, 11-Aug-83 01:59:45 EDT
Organization: Bell Labs, Holmdel, N. J.
Lines: 18

Recently I decided to attempt to design and build a pair of speakers
using commercially available components. I did a bit of research
regarding crossovers and discovered what appears to be a controversy
as to which passive crossover is best.
   Most sources cite the 6db per octave design as best due to desirable
phase effects, but warn against its use in the general case because
most drivers don't have sufficient bandwidth to handle the gentle
crossover slope.
   Although Linkwitz has advocated even order all pass crossovers, and
even provides design information for passive crossovers, Thiele warns
against using them because the input impedance is not constant, apparently
this would cause problems in cascading  sections for a mid-range filter.
Meanwhile, Linkwitz warns against using third order Butterworth filters
because of a tilt in the polar response pattern.
   Can anyone out there help me sort this out? I am not an electrical
engineer, so I don't expect my layman's analysis to be perfect.
   Also, can anyone recommend sources for drivers? I know about
Speakerlab, but their prices seem kind of high.....