Ur-ESEL
The Ur-ESEL (Ur for ancestor) marks the beginning of my ESL-diying.
The model shown participated at the first meeting of the K&T (German DIY-magazine) in Duisburg 2002.
One may notice similarities to the ESLs of Audiostatics, which comes to no surprise. The construction of a flat wire-stator-ESL is easy and the chance of success at first try are quite high. It allows for the design of fullrange-ESLs as well as hybrid-panels. Flat punched metal sheet stators discard from the choices apart maybe from very thin tall highfrequency panels. Their extreme narrow distribution character as well as the susceptibility to mechanical resonances prevents top class results.
The vertically orientated wires of a wire stator allow for an easy manipulation of the distribution character over a simple resistor array.
Also wires can be sourced easily and cheap in good quality and sufficient flashover voltage safety classes. Basically one only needs to think of a procedure of bending the wires perfectly straight.
One way might be to turn each wire back and forth under a constant strong pull. Easier, simpler and nearly as precise is it to wind the complete wire length for one stator in a meandering fashion onto a tensioning frame and to stretcher-level the wires with a strong pull.
PVC-insulated single strand wires are useable, like the H07-VU or the thinner H05-VU. Silghtly improved electrical parameters show Kynar-insulated wires which are used as wire-wrap wires. They do cost considerably more though and it might present a minor problem to find a suitable glue for this polyamide-derivative (PA). Poly-Urethane (PU) would be an alternative and with certain restrictions Poly-Ethylene (PE). PTFE, called Teflon- is inferior to thos application and very costly.
The probabely best electrical parameters would come with copper enamel wire with 2 or possibly 3 layers of insulating laquer. One needs to check though that no cracks develop in the insulation layer when tensionioning the wire. This would render the effetiveness of the insulation useless.
The tensioned wire array is glued to a flat ladder-like frame, comprised of a highly insulating material.
The ladderspokes serve as fixing points for the wires and strengthen the whole structure. They can be made from insulator material, acrylic glass and even from metal.
The upper and the lower open ends of the frame are closed with the same flat insulator material of the ladder.
The frame functions as membrane support and as spacer between diaphragm and the stator wires at the same.
The material of the frame needs to be a highly insulating and mechanically strong. The dielectric constant (epsilon) should be high (for most plastics its <10 ). Hydrophobicity is also a wishful parameter, because over time humidity forms a weak conductive, leaking film in togetherness with dust, smoke and dirt particles.
The entire stator area can be divided into several differing-in-width sections by electrical segmentation. Therefore the audio signal is fed to a thin middle section -usually about 1" wide- directly and fed via high ohmic resistors into the outer sections of the stators.
The resistors form together with the capacitance of the associated section a RC lowpass filter. High frequencies are attenuated. This way the active membrane area increases with decreasing frequency. Besides the improved, widend distribution character, the amplitude response can be manipulated.
A less prominent effect is the reduced variation of the electrical impedance and phase over frequency, which makes the life for the driving amplifier a lot easier.
A distinct disadvantage of this technique is the higher required transformation factor of the audio transformer of around 1:100.
The higher the transformation factor the more difficult and complex it becomes to wind a good audio tranny.
Paradoxically easens a ´bad´ transformer the driving amplifiers working and allows by the useage of resonance effects to manipulate the amplitude response around the upper bandwidth limit.
Contrary to the marketing promises of the manufacturers that they´d use only the best audio trannies, it was -and even is- common practise to use transormers of low, sometimes truely doubtful quality, to allow for a greater number of suitable amplifier choices.
Sonically a ESL is so superior that it can stand up against dynamic speaker even when totally crappy transformers are used.
Of course will a good audio tranny always provide for better sounds, but the number of suitable amplifiers is reduced. One needs to take a bit more care in finding the right amplifier.
Below approximately 100Hz two 10" bass drivers work together in a compact dipole casing.
Despite an active crossover and much tweaking I never manged to create a seamless transition from ESL to Bass.
Also the maximum SPL of the basses was quite limited, which might have resulted from the fact that the drivers were rather designed for small bassreflex casings.
To understand it right, the acoustic result was already more satisfactory than previous tests with Transmissionlines, Bassreflex and even closed casings. But there remained alot to be done.