Apologies for length!
Leak Sandwich construction.
“The following text is taken from company literature of the day:
"To understand the impact of this invention one must remember that to obtain low distortion the movement of a loudspeaker diaphragm or cone should follow as accurately as possible the waveform of the signals present applied to its speech coil. The ability of the cone to follow these impulses depends on the lightness as stiffness of the cone assembly.
"Conventional cones (paper, impregnated fabric, plastics, aluminium) suffer from low stiffness. This results in significantly large areas of the cone vibrating in motions uncontrolled by the speech coil. The uncontrolled motions produce transient "hangover", amplitude distortion, and intermodulation distortion. These shortcomings are inherent in all direct radiator loudspeaker systems using conventional cones, irrespective of cost.
"All of the above shortcomings of conventional cone loudspeakers are overcome by the Leak Piston-Action Sandwich construction. Using engineering principles as applied to airframe construction, the "Sandwich" cone comprises stiff aluminium skins for the outer surface where stresses are greatest, bonded to a thick core of featherweight expanded aluminium where stresses are lowest, giving far greater stiffness than the same total weight of either material used separately. This use of the most suitable material in optimum proportions and dispositions gives immense stiffness and rigid piston-action over a range of more than 6 octaves.
The original Sandwich loudspeaker was a two-way system with a 13-inch woofer mounted in a damped, sealed cabinet of approximately 2 cubic feet. The free-air resonance of the woofer was a low 19Hz, but when mounted the system resonance was 58Hz. The suspension of the woofer featured a cambric roll-surround which allowed for a low resonance and long throw.
The system crossover frequency was 900 Hz. The treble unit was approximately 3 inches in diameter and also featured a sandwich diaphragm (although early units appear to have had a fibre-cone tweeter).
The loudspeaker cabinet for the Sandwich system featured some novel design features also. The cabinet walls were only 12mm thick, but were very heavily damped with a similar thickness of bituminous lining. The magnet of the woofer was rigidly attached to the rear wall of the cabinet, thus increasing cabinet rigidity.
Leak had considerable success with the Sandwich loudspeaker, with it being used extensively in the professional audio industry and in homes both in Britain and overseas.
The marketing of the Sandwich system produced some interesting photographs. In one Harold Leak is shown standing on a plate glass sheet which in turn is sitting on top of the apex of a Sandwich woofer cone. In another, the immense stiffness of the Sandwich cone material is demonstrated by placing a column of 100 coins on it without causing a deflection, and comparing it to conventional cone material of the same weight where just one coin cause the material to bend.
The following notes are taken from the review in Gramaphone magazine : -[http://www.gramophone.net/
Note: this text is converted from the original magazine using OCR (optical character recognition) which clearly has problems in some places!
Leak "Sandwich" Fufi Range Loud- speaker. H. J. Leak & Co. Ltd., Brunel Road, London, W.3.
Both loudspeaker units show considerable thought in design. The bass unit has a massive aluminium frame casting, and a 4 in. diameter ring permanent magnet is mounted on a -, in. thick front plate. The back centring device is permanently stuck to the aluminium frame, and centring of the speech coil is achieved by moving the magnet assembly relative to the frame. The diaphragm, so revolutionary in conception, consists of a closed cell expanded polystyrene-moulded diaphragm J in. thick with a hemispherical bung at the apex. Both front and rear surfaces are covered with very thin aluminium foil which Construction gives extreme stiffness-to-mass ratio, and the name to the unit of the Leak "Sandwich". The surround is made from a treated cambric type of material completely sealing the front from the back of the diaphragm. The multi-corrugated rear centring device is acoustically transparent and allows the diaphragm an unusual freedom of travel of about if in. With the unit operating in free air without any baffle the low frequency resonance was measured to be 22 c/s—an unusually low figure due to the very free suspension, and notwithstanding the light mass.
The tweeter unit is also of considerable technical interest, having a massive 3 in. diameter ring magnet system, with steel end plates. The diaphragm is supported in a cast aluminium frame. As the diaphragm uses a shallow angle the frame is only - in. deep and designed to be mounted at the back of a panel of* in. thickness. When so mounted, the face of the diaphragm is flush with the surface of the baffle. In order to give the maximum space for air movement the inner circular edge of the frame is chamfered. The diaphragm is a moulded paper type with a plasticised cloth surround, central dust cover dome, and a I in. speech coilwith rear corrugated centring device.
A half-section crossover filter is used with heavy gauge air-cored inductors and high capacity A.C. operating electrolytic capacitors; and the cross over frequency is centred around 1,200 cis. The various components are mounted on a heavy wood block with a central hole. Screwed into the rear of the woofer magnet system is a threaded brass rod that passes through the wooden block. When the loudspeaker unit is in position the block acts as a spacer between the back of the magnet and the rear panel of the enclosure. A screw-head nut then puts the back panel under stress and controls the panel resonance due to pneumatic action. Connections to the unit are made with a non-reversible two-pin plug and socket.
A guide to the performance of a loudspeaker can be obtained from the impedance characteristic. The Leak "Sandwich" loudspeaker has a nominal impedance of 15 ohms and measurements made with a Muirhead impedance bridge showed the remarkable smoothness of the curve. This was confirmed in tests with a Marconi R.C. oscillator feeding a Leak "Vanslope" Stereo pre-amplifier and "Stereo 20" power amplifier. Using a calibrated A.K.G. condenser microphone and a Cossor double beam oscillograph the output as displayed maintained a sinusoidal wave form down to 35 cis, the amplitude falling slightly below 50 cis. A slight boost of the "Varislope" bass control lifted the extreme bass to give a substantially flat characteristic. From 10,000 c/S upwards the output remains at a constant level up to 15,000 cis after which it tails off, although there is still a measurable output at 20,000 c/s.
Perhaps the most devastating test one can apply to any loudspeaker is to feed it with a square wave signal, and only seldom will one ever see an oscillogram resembling a rectangular waveform. One is asking the loudspeaker diaphragm to perform an impossible task, for at one instant of time it should be at rest, then accelerated in no time to a new position where it remains stationary until the waveform returns to zero when again the diaphragm moves in zero time to its new position. A square wave can be resolved into a fundamental frequency with an infinite series of harmonics; and I made the test so that the 50th harmonic would be within the frequency range of the tweeter unit. Hence with a fundamental frequency of 200 c/s. and the "Varislope" controls set for a flat characteristic and the A.K.G. microphone two feet from the enclosure I was more than surprised to see a reasonably good rectangular waveform on the screen. The slope of the sides did not exceed 15° and the top flat only drooped about 10°. A slight cut in the bass control and introducing the rumble filter improved the rectangular waveshape and with the exception of one other loudspeaker this is the nearest to a rectangular wave response I have ever seen. You might well ask "does such a test give any information about the enclosure's musical performance?" The answer is a positive "Yes!" for it indicates the excellence of the transient response, the overall smoothness of the response without any violent peaks or troughs, and the wide frequency range.
One of the other tests that I have always enjoyed showing to students is the way in which the conventional paper diaphragm breaks up and fails to operate as a piston. This can be done simply by applying a variable low frequency to the loudspeaker via a power amplifier so as to cause the diaphragm to move a fair distance. Then one can either apply lycopodium powder to the diaphragm and see the areas of zero movement, or use a stroboscopic light source. I use a small E.M.I. Electronics hand stroboscope and adjust its flashing speed to he about I c/s. different from that of the R.C. oscillator. The diaphragm then appears to move inwards and outwards at a speed of 1 cis. and normally one finds that some parts of the diaphragm are actually stationary, other parts move outwards and other parts inwards, all at the same time. Obviously when the total area of those parts moving outwards equals that of those parts moving inwards, the radiated signal tends to cancel and there is a severe trough in the response curve. Such a test applied to the Leak "Sandwich" woofer unit shows the sound philosophy of the design. From the lowest frequency I can generate (20 cis) up to 200 cis there is no evidence of any departure whatsoever from a true piston motion. Above about 100 cis it is difficult to put sufficient power into the unit to cause the diaphragm to move a visible distance.
