[upload=gif]uploadImages/200311316302577492.gif[/upload][upload=jpg]uploadImages/200311316304352037.jpg[/upload]
On this page I discuss the Quad ESL speakers and especially the Quad ESL-63 combined with the Gradient SW-63 subwoofers. There are several other pages in this Quad ESL section, press the buttons below and these will pop up. My intention is to provide information on the Quad ESL loudspeakers and also inform you about the improvements that can be made to these excellent products.
The 'Quadfather'
The original Quad ESL, known in these days as the Quad ESL-57 has been elected as the greatest HiFi product of all time by HiFi News (January, 2000). It is hard to argue with such a decision, any HiFi product that after 45 years must be considered to be among the very best earns such praise. The original Quad ESL is still one of the cleanest sounding loudspeakers you can find.
Ken Kessler wrote:
Listening to the original today merely confirms what many suspected: despite its limitations, the Quad ESL delivers some of the most natural, open-sounding mid-band (especially on vocals) you're ever likely to hear. It is a speaker which addresses certain needs, appealing to those who want their ears caressed, not battered. The speaker disappears into the room, leaving you with nothing but... music. Hook up a pair to some small single-ended triode amps, a mint pair of Quad II or better still a Radford MA 15, and you'll just shake your head in wonder at the novelties of current manufacture. It is as stubbornly superior to modern designs in aesthetic, cultural and intellectual terms as a Leica M4 camera is to a polycarbonate SLR, as a 1950 Patek Philippe Calatrava is to a digital watch, as a vintage Vuitton trunk is to a nylon rucksack.
Since 1987 I have been using the newer Quad ESL-63 and even if some still prefer the ESL-57, I think the ESL-63 is an improvement with its better dispersion characteristics, increased dynamic capability and extended low bass. In a small room the older speakers can give a slightly more intimate and detailed sound, but in a larger room or at higher level with large scale music the newer speakers wins. With some tweaking of the ESL-63 the ESL-57 is left behind in almost all areas, and it is also an easier load for the amplifier. The ESL-57 can make some semiconductor amplifiers sound very bad and it is recommended to partner them with a tube amplifier or one of Quad's own amplifiers.
I will not make a long description of the Quad electrostatic speakers (a technical page is coming...), as I think there is sufficient information about them on other sites, see the links at the bottom of this page. It is my opinion that the quality of these speakers, both the old ESL and the newer ESL-63, are hard to match to ANY other speaker. They may be (are!) beaten by several other designs when it comes to quantity, but not in quality of reproduction. This is of course due to their electrostatic nature, which makes it hard to get very high sound levels with a diaphragm area suitable for ordinary rooms. Some manufacturers have tried to make really large ESL designs in order to get higher levels and deeper bass, for example Metaxas and Sound Lab. I have not heard either Metaxas or Sound Lab, and although I am sure they sound great in a very large room, I think it will be difficult to get the exact soundstage and homogenic sound typical to Quad ESL in a more ordinary room (15-30 m2). This is because you must sit quite far away from a large speaker to avoid the interferences between the different parts of the diaphragm, which causes an uneven frequency response and also gives a radiation pattern that is hard to control.
The Quad ESL-63 takes the performance advantages of electrostatic loudspeakers a stage further. The ideal loudspeaker for stereo reproduction is a single point source reproducing all frequencies. Loudspeaker engineers have been trying to develop one for more than half a century. If it is not possible to produce a point source loudspeaker, can we instead make a loudspeaker that to an observer behaves like one? Imagine a theoretically ideal point source loudspeaker radiating sound pressure waves and then imagine a plane in the air a short distance from the source and at right angles to the direction of propagation. If the air at the plane is made visible in some magical way, we will see concentric waves radiating out from the center just as they do when a stone is thrown into a still pool. If we substitute an electrostatic loudspeaker membrane for the plane in the air, make it move in exactly the same way as the air on our imaginary plane and suppress the imaginary source, the results to an observer positioned on the far side will be identical to those from the ideal source.
The Quad ESL-63 does exactly this. An ingenious arrangement of concentric electrodes fed by a sequential delay line produces a sound pressure pattern that is an exact replica of that from an ideal source placed 30cm behind the plane of the diaphragm. The ESL-63 is a totally homogenous sound source, phase true and aperiodic, with a frequency response, both on and off axis, quite free from the irregularities that are inevitable with any multi-way loudspeaker system (and also with large diaphragm planar speakers) . It has a very well controlled directivity characteristic with the result that there is no stereo hot-spot (not entirely true, but less critical than with other planar speakers). Since the loudspeakers can be placed right up to the side walls the stereo stage can be very wide. The result is a loudspeaker of unsurpassed accuracy, which with the right programme material will produce a more realistic and satisfactory illusion of a live musical performance than has been possible hitherto. Some manufacturers have chosen to make a curved membrane (for instance Martin Logan) in order to control the directivity characteristic, and in the horizontal plane it probably works, but in the vertical plane the problem still exist. Probably there will also be some distortion due to the different force needed to move the diaphragm backwards relative to the forwards move. Still Martin Logan's ESL panels sound good, but the hybrid construction of most of their speakers is a more severe problem. Even if you could make a box with no coloration (which you can't, I discuss this on my 'Opinion' page), the different radiation pattern of a monopole (bass) and a dipole (ESL) makes it impossible to get a linear frequency response on more than ONE distance. The obvious solution to this problem (if you must use a box) would be to use active filtering with the possibility to change the bass level.
Despite it's ingenious design the Quad ESL-63 is not a perfect speaker though, because you need to get lower in the bass region to produce a really convincing soundstage. Everyone with experience from high quality subwoofers can describe the effect these have on the soundstage, even when no low bass seems to be present, for instance with an acoustic guitar recorded in a church hall. The lowest frequency of a guitar is around 80 Hz, so adding a subwoofer for the 20-40 Hz region should not make a difference, but it sure does. My theory is that the transients excite room resonances with lower frequency than the 80 Hz. If you have studied Fourier analysis, you know that a perfect impulse (very short) includes ALL frequencies. The difference in soundstage is apparent even with female a Cappella singing, this is more difficult to explain, but maybe it is the interference between sounds bouncing around in the room that does it. There is a project presently going on with ultrasonic transducers creating sound with the use of interference, so my explanation is maybe not so far out. Another explanation could be the recorded background noise, which certainly adds low frequencies, and maybe helps to define the room information. As the Quad ESL-63 does not give much output below 40 Hz, there is a problem, but a few years back the Finnish company Gradient finished the Gradient SW-63 dipole subwoofers. These take over from 115 Hz and down to around 25 Hz with reasonable output and uses active filtering and an additional amplifier for the subwoofers. The filter unit also compensates for the 6 dB/octave loss (cancellation from the rear at low frequencies) due to the dipole construction. When my previous subwoofer Audio Pro B2-50 broke, I thought that I could live with the Quad ESL-63 on their own while the B2-50 was repaired. Within a week or so I realized that it was not possible, the deep bass was really missing. As the time is limited and I had been interested in the Gradient SW-63 for a while, I bought a couple of SW-63 demonstration samples at a reasonable price. At first I used them with my old Luxman SQ507x, which is not really suited for driving subwoofers of this kind, but still managed to get them in action. The first impression was that the bass was thin, no matter at what level the subwoofers were running, but quite quickly I got the feeling that it was real bass, no box-sound, no boom, and the imaging was superb. The SW-63 does not dig as deep in the bass as the B2-50, but the freedom of coloration compensated that, so I was not really missing the deep notes. The B2-50 I used from 40 Hz and down with the ESL-63s running as fullrange speakers with no filtering, since the filter supplied with the subwoofer was not up to my standard. The SW-63 I used with the supplied high-pass filter. This filter was clearly better, but I noticed almost immediately that the magic was gone in the midrange and treble area, so I tried running the ESL-63 without high-pass filter, and it worked quite nicely. The magic was back, but with less detail, as the Quads definitely benefits from removing the lowest frequencies, but the coloration from the filter was too obvious so I ended up not using the high-pass filter. This gives a 3 dB rise in the frequency response in the region 50-100 Hz (I have measured it), but the subwoofer level was easier to set. I do not recommend this solution, but you can always try it if you want a fat bass. Later I replaced the Luxman with a Rotel RB980BX, which is better suited for driving subwoofers. There are of course better alternatives, but this was what I could afford and it works alright with the 4 ohm connection. When I measured the speakers I noticed that the ESL-63 and SW-63 were 90 degrees out of phase at 115 Hz, no matter how they were connected. This was not a surprise really, as the the low-pass filter for the SW-63 is third-order, and the high-pass filter is second-order. To get them in phase the high-pass filter should be first-order or third-order. I measured a little and placed a high quality polypropylene capacitor in series with the main amplifier input, giving a first-order 115 Hz high-pass filter. The result was AMAZING! The message to all ESL-63/SW-63 users is, DO IT!! Suddenly it was very easy to find correct phase and adjust the level. The bass was a lot more homogenic and full of nuances and flavors, I actually lowered the subwoofer level quite much, with the interesting result of a more firm and heavy bass. The midrange and treble is now exceptional with no coloration from the filter and ESL-63s free of the low frequencies which improves their performance a lot. The mechanical rigidity of the Quads could be better (it shall be in the new ESL-988 and ESL-989) but with a lot of tension removed as they do not have to perform the low frequencies, this is now a minor problem. Apart from the filter tweak described above, I have also altered the value of a couple of capacitors and resitances in the Gradient low-pass filter in order to get a deeper bass. This is another success, they now deliver almost 20 Hz with reasonable level, breathing even more air into the soundstage and producing a low frequency response of an unmatched quality (maybe Audio Artistry could be a challenge). Driving the SW-63 into the 20 Hz region means of course a lower maximum SPL, but I think the quality compensates for the loss in quantity. The tweaks above and a few other tweaks will be described on the page 'Quad Tweaks'. I have also altered the input impedance of the Gradient filter from 22k to 500k and if you use tube amplification I recommend this alteration, the sound opens up considerably when the output stage is not loaded as hard as with 22k. The performance of the Quad ESL-63 with the tweaks described on the tweak page together with the Gradient SW-63 is really first class and at a reasonable price, second hand units are amazingly cheap compared to other lesser speaker systems. The midrange is of a quality that I am sure would make even the most fanatic Quad ESL-57 owner happy. One of the good things with this system is that it is an extremely neighbor friendly fullrange system. The dipole construction means that a lot less bass is radiated in the room, and it also focuses the energy on the listener, making it more sensitive than the stated 86dB/W indicates. It is not necessary to play as loud as with ordinary speakers, the clarity of the system is extraordinary and you get the sense of dynamics at much lower volume than with monopole speakers.
