OCR-excerpt from HifiNews
Original printed September 1981,
Reprinted Februari 1999

Quad ESL-63 by Trevor Atwell, Part 2
LAST MONTH we dealt with principles and realisation of the ESL-63, and with its subjective performance. Continued listening during the interim has proved to be absolutely compulsive, and a constant source of pleasure. Some excellent Proms relays have been available, and the sense of being physically present has never been sharper than with these units.
It is now time to look at the measurements, carried out as always by James Moir and Associates at about 7 m above ground in the open air. The sine response was taken at 1 m on axis, and is within ~2 dB from about 60Hz to 8kHz. being only 6dB down at 35Hz. In a room the ground reflection extends the bass response, and Quad have tailored the bass roll-off to avoid boom. A few peaks and troughs at the highest frequencies are partly due to reflections at the dust-cover, but our measurements are higher than those obtained by other laboratories. Since movements of the microphone produce some shifts in position and amplitude, and since we were measuring at 1 m instead of the 2 m for which the far field is designed to be flat, we suspect that this is another example of the influence of geometry on sharp features of the sine response, often produced by boundary conditions. Such features would not normally be audible, even if 'true' in a room, and were not heard. The third-octave pink noise response confirms this.
The distortion performance proved difficult to measure, because we were limited by the swept-oscillator distortion content! The distortion plot was amplified by 20 dB from 20 to 150 Hz, above which the loudspeaker distortion could not be measured by sweep methods. Single-frequency measurements at various points throughout the band showed the distortion to be generally below 0 1%, and frequently as low as 0,03%, where it was close to the residual noise level. I have never seen results like these for a fundamental output of 96 dB spl at 1 m from any other loudspeaker. Even the levels below 150 Hz are better than almost all conventional designs,  and the 3% second harmonic at 20 Hz (due to the rise in flux density in the input transformers) is quite respectable so far outside the intended passband.
The ESL-63 is unusual in having an impedance that varies with input level at low frequencies. We plotted the modulus for three constantvoltage inputs, and it will be seen that the impedance falls at very low inputs, for which the result is unimportant, the current demands on the driving amplifier being insignificant. The effect is due to decrease in lowfrequency incremental permeability of the core material at low flux densities, and must be distinguished from the quite different effect already discussed in which an excessive peak flux produces distortion at low frequencies. In all significant areas of operation the impedance is substantially resistive, giving amplifiers an easy time except under gross overload conditions. This is another improvement over its predecessor, which presented the amplifier with a highly capacitative load, falling to a very low resistive value at a resonance (about 1/2 ohm at 12 kHz. if memory serves).




Our normal sensitivity check produced 93dB spl of pink noise at 1 m on axis for one unit in an 83 m3 listening room for an input at the terminals of about 7 V which is a good figure.
Under the general heading of tests it may be relevant to describe Ouad's own routine test, which I watched at the factory. A square pulse is fed to one loudspeaker, and the audio output from a B & K microphone observed on a storage oscilloscope, which shows an incredibly good representation of the original pulse. Next, pulses of opposite polarity are fed simultaneously to two units, with the microphone in front and between them. Moving the microphone produces a null, which becomes a straight line with no more than 0,5dB adjustment to one input!
This Ouad design is a very real technological advance based on an elegant concept, followed by a painstaking practical realisation in which the basic ideas have been further refined. It may not be perfect, and I should be surprised if P.J.W is not already gathering some long-term ideas for the future, but I find it extremely difficult to envisage how any multiway moving-coil arrangement can be made to compete in terms of accuracy, precision, lack of coloration and imaging. Some m-c units do make some recordings sound more exciting, or musical, or whatever is the current buzzword, but the ESL-63 is causing at least some recording engineers to take another look at their techniques, so the number of discs that can be 'improved'  by suitable errors in loudspeakers may hopefully fall.
The listener who actually knows the sound and real excitement of live performances, and who wants to hear the true subtleties of music-making in any genre through loudspeakers, rather than the character of speakers via the music, is the natural possessor of the ESL-63, and will not easily be satisfied with less once they have been audititoned. The price is high, but it is justified, and the results are worth every penny. 'The closest approach to the original sound' has moved decidedly closer.