|
|
UPDATED 19-01-06
The beginnings
I guess my loudspeaker construction started as early as 16 and in those days a single drive unit was used with no crossover - drivers I used were from makers like Rola (Melbourne) and Wharfdale and Goodman's of England.
Important Events
In the early 70's two organizations changed loudspeaker design markedly
1. the BBC in England with there own in house studio monitors
2. a small company KEF (Kent engineering foundry) Raymond Cooke founder and pioneer in plastic cone drive units, in liaison with the BBC. From this work KEF created a new series of consistent low colouration drive units - Bass / MID / Tweeter all of advanced design and construction.
KEF produced several Systems based on these drive units the "Concerto" was one and for the BBC the famous LS3/5a mini-monitor (I still use a pair to-day for casual listening).
Later KEF turned to computers and impulse testing using Fourier transform and the concept of a "Target function" to enable complete system design ie. drive unit(s) together with the cabinet in conjunction the chosen crossover filter (Initially 3rd order 18bB /octave) and with the use of crossover optimization were able to synthesize very accurate results. Later with the KEF R105 speaker system and the adoption of the Linkwitz Riley 4th order crossover filter a very advanced design was created.
My Later Ambitious Projects
At that time I had long abandoned the single driver in a box in favour of my ambitious 4 way passive crossover speaker system using all KEF drive units. B139 / B110 / T27 / STC 4001 super tweeter. At the same time I built a BBC specified LS3/5a for home use and this system amazed me it sounded better in the treble midrange than my huge 4 way system, why well it was all in the crossover design, the LS3/5a simply sounded non coloured and well integrated - photo shows one hung by Nylon fishing line, no stand needed. The LS3/5a's are used in the kitchen/family room for casual listening. I additionally equalise the LS35a from a fc of ~67Hz Qc ~0.6 to fc of 30Hz Qc 0.5 so response extends to 30Hz but with the control of a Q of 0.5 [using a Linkwitz transform] perfect little circuit. :)
Cutting Edge
My attention turned two articles on crossover design "Passive crossovers for non-coincident drivers" and "An Active Satellite Sub woofer system" both by Siegfried Linkwitz in California. After much reading and head scratching everything about loudspeaker crossovers and system design just fell into place like pieces to a jigsaw - how active filters could implement the desired filter without driver interaction - how individual power amplifiers for each drive unit could provide sound that no passive filter system could ever achieve - clearly an Active Loudspeaker was light years ahead of anything else.
Advanced Active 3 way System
Some time later after many hours building the 4th Order Active crossover [principles here] the Power amplifiers and the Cabinets I had a new System, two 160 Litre woofers (23Hz-100Hz) and two 11 litre satellite units (100Hz - 20Khz) hung by Nylon fishing line from the ceiling, how could I measure and verify the results, well I used the Linkwitz Tone Burst generator which I built - this enabled in room (image1 image2 image3) measurement of frequency response without an Anechoic chamber - some measured room anomalies will still occur even with the tone burst system however. After some level adjustment I had a system that was incredibly smooth and flat and sounded incredibly dynamic could achieve 100dB SPL at the listening position with absolute ease and no strain - voice was uncoloured and stereo imaging precise. Amplification is 8 100W modules I built to the Prof. Cherry [Monash Univ.] Design, the famous nested differential amplifier.
The Future
If you were to study all loudspeaker systems in the market to-day - whilst there are similarities there is clear disparities on what constitutes an ideal system, some of it is over emphasis on the wrong priorities in the individual designs and or slavish attention to details that overall contribute little to ultimate realism or truth to the source be it CD or other music source. Indeed research has shown that the room interaction with the loudspeaker is very important, the reverberant sound added to the direct sound that the listener's ears receive effects sound quality markedly, consequently the typical closed box speaker is bad in this respect unless placed carefully in the room bass sounding boomy and poor imaging, the Dipole loudspeaker (no box) radiation from the front and rear and with the figure eight radiation pattern excites the room much less resulting in almost no bass boom and more precise direct sound field to the listeners ears. Dipole examples QUAD of England ELS electrostatic speakers the current thinking from Siegfried Linkwitz Audio Artistry and Linkwitz Lab, there are other panel dipole speakers on the market but they don't quite make the exacting requirements for various reasons.
Sound in the Home and realism
Realistic sound pressure levels in the home ~100db are easy to achieve but whether its realistic ie. like live sound is something else. I can easily tell recorded sound - why well for a start the CD is mostly a studio mix ( but not in the case of classical orchestra ) however most music is studio derived and limited dynamic range (deliberate on the part of the engineer) highly processed - to get a desired result, so how can it be real, added to that you have the playback loudspeaker and room limitations. To me live music is obvious even coming from a neighbours house ie a Piano or Sax where as recorded music from this same neighbour is also very obvious, given that the live Piano or Sax may well be by an amateur musician the live instrument still stands out. Listen to real live music at a venue without PA (amplification) or a street band its different, compared to recorded music it has real dynamics and dynamic contrasts that does not exist in home playback systems even allowing for differing venue ambiences. Listening at home at reduced levels that suit most of the other occupants simply robs the real listener of achieving close to real sound therefore it makes no sense buying and installing a "High End Audiophile System" if the listening level is always low for domestic reasons, so too placement loudspeakers simply must be positioned correctly even at the expense of domestic harmony, or alternatively a dedicated listening room (example). However some small scale music can sound very nice at home even at reduced listening levels ie classical guitar is one or vocals.
This section is to help anyone wishing to clone the legendary LS3/5a.
Over the years I have Built 2 pair's of LS3/5 clones, the first pair have been in use for more than 10 years, I needed a second pair as PC multimedia speakers and as I had B110 and T27 drive units on hand and 2 crossovers all that was required were 2 cabinets.
The KEF CAD crossover Crossover is very good, the high pass section to the T27 is a true acoustic 3rd order Butterworth response, and corrects for the T27s under damped 1100Hz dome resonance with a little null circuit, as well as a lift above 10Khz the result is very smooth, the B110 is also handled very well, the resulting driver integration is excellent.
For this pair, like the original pair, I went for an ULTRA rigid cabinet not a BBC damped thin wall enclosure, in this instance I used 19mm MDF all the 5 walls have 19mm square battens running midway along their length bisecting the panel area and additionally cross bracing on the larger panels, the result is high rigidity and in practice no wall flex at any sound levels while playing music, MDF by nature is lossy having internal damping, wall resonances will be in the Khz region because of enclosure rigidity and will in practice not be excited.
Cabinet depth overall is 325mm this gives a larger volume and inherently extended bass compared to the standard LS3/5 enclosure, the rigid box also enhances bass quality, other dimensions were adhered to to ensure a smooth mid treble region and compliance with the crossover design.
Exterior was veneered in African Teak in a special glue press and polyurethane / oiled hand rubbed matt finish
High grade 12mm ply was used for the baffle and the enclosure interior damped with Australian long haired wool, the B110 was seated on a lossy gasket to damp the driver chassis.
My measurement of the acoustic response is limited to above 100Hz in this instance as the measuring room was small. The linkwitz Tone Burst generator was used which provides near anechoic accuracy but high resolution as frequency is continuously variable without steps the result is a response that shows deviations that will be audible but avoids small deviations that in practice are not significant, it also has the ability to show driver resonances with an oscilloscope. Looking at the T27 shows quite good control of resonances for its vintage even up to 20Khz it is however a hard dome with little damping so has some sonic signature, by contrast the B110 [sp1003] shows the classic peak and ringing at about 4.6Khz the result of the voice coil becoming unstable relative to the cone is not being in synch, the B110 [SP1057] is better in this respect [and subsequent derivatives], vintage drive units yes but due praise should be given to Raymond Cooke founder of KEF [Kent Engineering Foundry] who changed drive unit design for the better.
Active equalization is used with this second pair of clones, using the Linkwitz Active transform circuit, [using a Linkwitz transform] what this does is takes the close box resonance of ~60Hz Q 0.6 and transforms it to an equivalent resonance of ~19Hz with a Q of 0.5 [this means 30Hz is -3dB] but you have the advantage of group delay for a sealed box of 19Hz resonance with optimum damping [Q = 0.5] resulting in extended bass from a box not physically realizable but with all the advantages, you will not shake the room at 30Hz however the B110 is too small, on normal music overload does not occur only extended bass close to a sub-woofer but with no phase problems.
To print the crossover do a save as on the image file and print it out.
Crossover, acoustic response, final ls3/5a clone, grille off, grille on
Active bass equalizer for ls35a speaker and clones.
Active equalising of the bass of an LS35a and a clone results in extended bass without a SubWoofer [using a Linkwitz transform] for more details and in depth explanation of the theory.
Active circuit Graph response before and after equalising, download Excel design sheet for equaliser ls35a.zip.
The equaliser should be placed between the preamp and power amp, use any good Audio OP amp a buffer stage is also needed ahead of the Equaliser use dual OP amps one as buffer and the other as the Equaliser.
Electronic measuring and construction skills are needed to build this equaliser also read the theory here first before attempting this project [using a Linkwitz transform]
