DIY loudspeakers series - Part I

Loudspeaker energy preservation

In which Mark Wheeler sounds off about loudspeaker sound

Perhaps energy has been observed in the wrong place or at the wrong time? Perhaps the loudspeaker cabinet walls are observed vibrating. If so, the vibration can only be present from energy derived from the expensive stuff at the input terminals, so it is not unwanted. It is energy we wanted to make beautiful music, but is now lost from that task to one of noise.

It is diverted energy, it is misplaced energy, it is misdirected energy, it is delayed energy, it might be thought of as delinquent energy. Like other kinds of delinquent behaviours, it arises from something that should have been loved once, that infant offspring of the source and amplifier, now waywardly vandalising those cabinet walls.

How may we help this newly arrived energy to develop into well-behaved grown-up sound energy? We must ensure our energy treads a true path of righteousness, towards that goal of uncoloured musical information.

Our little packets of energy will not develop unscathed by being smothered in suffocating lossy glop of one sort or another. The energy detected making nasty noises is a double problem for the listener. It is a secondary problem by generating noise that masks the music, but it was once the music itself, so its misdirection has twice the detrimental effect than its acoustic measurement might imply.

The delinquent youth is often blamed and locked away from society, much as the delinquent acoustic energy is smothered in lossy damping and eccentric fixes. It is more helpful to consider the paths the energy may take, and to make it more likely that our beloved musical energy takes the paths that will lead to music than to noise. Resolving the vectors of any energy impulse will indicate where it thinks it might want to go, and as any fule kno, it is easier to control energy by applying strength directly to the line of the vector sum than by any other creative mish-mash of planes.

Thus when a driver motor (the magnet & voice-coil system held in relationship by the spider) chucks the cone mass forward or back, the best way to keep the chassis in the same place is hardly likely to be at the centre of a big floppy sheet of particle-board, as is the case in many loudspeaker designs. Numerous more elegant solutions are possible, and many have been successfully tried & tested. Two opposed bass-drivers rigidly connected together at their magnets, as produced successfully by Kef for many years, is potentially a better idea than any single driver for the lowest decade (e.g. 30Hz-300Hz) of the audio passband (say 30Hz to 20kHz). The Kef models went further by mounting both drivers in a coupled-cavity bandpass cabinet, but it is the mechanical coupling of drivers working in opposite directions that interests us today. It could equally be applied to drivers mounted on opposite sides of a cabinet, wired in-phase.

An effective means of reference of the bass-mid driver to the floor, more direct than the connection to the bass-loading cabinet, is an alternative used by Naim in the original SBL. The SBL (Separate Box Loudspeaker) used a steel frame to support the 200mm bass-mid driver's rigid sub-cabinet, rigidly coupling the driver-basket and box via the steel tubes. The speed of sound is higher in steel than wood or chipboard. The much larger bass-loading cabinet was coupled more flexibly to the driver sub-cabinet using a gasket. My first audition of this speaker in the 80s was a revelation in low levels of cabinet colouration and high levels of information retrieval, brutally exposing the paper-coned 200mm driver. The tweeter had its own separate box too, thus minimising intermodulation of the tiny treble waveforms by the big bass reactive forces, usually transmitted through chipboard carcasses in other speakers.

Bowers & Wilkins (B&W) pioneered their Matrix system of cabinet construction that effectively supports the cabinet wall over the whole wall area. This thinking simply ignores established custom & practice; instaed of wondering how to brace an existing cabinet wall, the B&W engineers sought a solution that did not begin with the problem. These are just three commercial examples demonstrating alternative means to maintain that vulnerable electrical energy on a useful path to music as it transduces the transition from electrical to audible energy. The Naim SBL offers useful, adaptable ideas to the amateur modifier of loudspeakers. Few of our systems are likely to offer a pair of identical bass-only drivers for experiments in opposed driver coupling, and bass-mid drivers would be unsuitable in most cases.

Unless your speakers have braces connected to the driver-bolts, or have a brace to support the rear of the magnet, the most efficiently supported bass drivers are those mounted on narrow cabinets so that the outer driver srews are very close to the cabinet side walls. The worst case is the use of lossy driver mountings on lossy cabinets.

Invariably these designs sacrifice musical information on the alter of lowest colouration. Their reduced colouration has to be exceeded by reduced information. Very clever solutions to very specific problems could be better addressed by designing the problem out of the system than to mask it with lossy arrangements that vary with temperature and time.

Adding braces to existing loudspeakers is usally a great improvement, unless those speakers are both expensive and already optimised. Replacing screw-fixed separate panels (e.g. detachable rear panels on some models) with good glue joints will also be clearly audible. While that panel is out, adding vertical braces to cabinet walls will justify the effort too.

What materails to use as bracing, and what glue to use to fix them, will be discussed in future articles.

When building a loudspeaker cabinet from scratch the provision of a brace, to support the baffle directly behind the bass driver, will dramatically reduce the flexure of the baffle, even if it is already very thick and strong. The 25mm birch-ply brace shown here adds observable, audible and measurable rigidity to a baffle that is also constructed from 25mm birch-ply bonded to 12mm maple; the latter being the same wood specified for high-quality guitar necks. If bracing this strong baffle is audibly effective, imagine how much more your 12mm mdf box may be imporved!

Conclusion

So, dear reader, the message is simple. All energy in loudspeakers is valuable. It is easier to get something to behave well than to correct it when it misbehaves. It is better to keep energy well controlled at all points in the loudspeaker than to transform those lovely Watts of music into a few paltry milliwatts of heat.

Intelligent planning of all elements of loudspeaker design, excecuted with good quality construction is the effective way to go. Dissipating diverted energy using heavy lossy damping will inevitably be less musically effective than ensuring that energy is converted into musical sounds.

The next few articles in this series will explore some of the techniques amateurs may employ to maximise conversion of that musical electrical energy back into music itself.

[Go to Part II - cabinet walls materials]