I have performed some measurements on my speakers and these are presented here with my comments
This is an extreme closeup of the impulse response, only 2.5 ms from the impulse start to the end of the window. The 0.25 ms ringing is typical for the Quad ESL-63 and the probable cause is the delay rings, these are delayed with 0.025 ms each. Even though there is some oscillation, this is a very good impulse response for a loudspeaker.
The impulse response (inverse polarity) shown here is more typical, and very good. Do not mind the irregularities in the frequency curve, it is very difficult to measure planar speakers and moving the microphone makes a big difference.
The waterfall diagram shows NO RINGING at all, and the waterfall is very even, this is probably the reason for the homogenic sound of the Quad ESL-63. Ringing is the worst enemy of good sound reproduction, as it increases the energy radiated from the system. This is why some systems seem to have too much treble, even though the measured (static) frequency curve is flat.
The MLS waterfall diagram shows a great similarity to the impulse measurement waterfall diagram. The difference probably occur because the used window is 5ms instead of 10ms, used for the impulse measurement.
Gradient Nearfield Frequency Measurement
The Gradient SW-63 with the new filter tweak (see 'Quad Tweaks Page') has the -3dB point around 22Hz and at 20Hz the response is only at around -6dB. The difference is easily heard with almost all program material. It is possible to set the low cut frequency even lower, but it will make life harder for the woofers. The chosen low cut frequency works fine for me, and in order to go deeper maybe stacked SW-63 could be used. This is a bit expensive though.
New Filter Tweak, Gradient Nearfield Frequency Measurement (numeric)
Low Frequency Response in Listening Position
The response for the ESL-63 with and without the SW-63 are plotted, you have to figure out yourself which one is with the SW-63. My listening room is a little more than 6.5 meter long and 3.5 meter wide, and the resonances at 25, 50 and 75Hz are all the result of standing waves in the loudspeaker direction. No resonances from the side walls or the ceiling /floor are visible, and that is one of the advantages of the dipole action. The low frequency response reaches down to the expected 22Hz with almost full level, and the reduction in bass level (or rather increase above 100Hz) seen is probably due to the microphone behaviour (I tried different microphone positions with white noise and spectrum analyzer, and the result differed a lot). The ESL-63 is ruler straight down to 50Hz and as you can see the response from the SW-63 follows that very closely. The measurement is made using sinus sweep and this enhances the room resonances and it does not look as pretty as third octave measurements often published, with real music the resonances are not as apparent other than with long lasting low frequency notes (organ, synthesizer), since it takes some time to build up the standing waves. The resonances should be quite easy to get rid of with the help of tube traps (more on that subject later), but even if it looks bad it is not too disturbing. When I listen in my room, I have the feeling that it has unusually low amount of room resonance, but it could of course be better. The 75Hz resonance can be a little annoying sometimes, but the 25Hz and 50Hz resonances does not matter much. The 50Hz resonance seems to be reduced by the fact that my speakers are positioned almost in the middle of the room. The Quad ESL-63 on its own does indeed make sound at 25Hz, when I measured this I had to check that the SW-63s were not running!
Distortion Measurement in Listening Position
The distortion was measured in my listening position (2.5 meter distance) with both speakers driven at around 85 dB. I was mainly interested in the low frequency distortion, and the SW-63 seems to have around 3% at this level. With nearfield measurement the figures would probably have been lower, but 3% is about what you could expect and at high levels it is probably around 10%. With servo control this could easily be reduced to around 1%, a very interesting project.
Distortion Measurement of Quad ESL-63 at 1 Meter
The distortion was measured at 1 meter with one speaker driven by my Jadis DA30 tube amplifier and the level was around 80-85 dB. The distortion is very low indeed and lies around 0.1 to 0.3 % with the 2nd harmonic as the dominating.
Distortion Nearfield Measurement of Gradient SW-63
As expected the distortion was reduced when nearfield measurement was used. The level was around 80dB at 1 meter, and the distortion lies around 3-5 % at the lowest frequencies and is reduced to around 0.5 % at 50-100 Hz. Not too bad, and combined with the freedom from box sound the result is a very clean sounding bass.
This site uses frames, but if you have found this page via a search engine and do not see any frames, you can use the button below to get to the main frame of the MT Audio Design site.