|
Speaker Asylum General speaker questions for audio and home theater. |
|
|
|
Speaker Asylum General speaker questions for audio and home theater. |
|
In Reply to: Deliberate or not. posted by klausDK on April 3, 2008 at 01:44:35:
> To me it seems like someone inside Celestion believed the
> Celestion Sl 600 to be a better speaker than the supposedly
> technically superior SL 700. Why would they otherwise have
> produced these two speakers side by side, year by year?
> To the best of my knowledge the 700 newer replaced the 600,
> strange?! Isn’t it?
Hi Klaus
I chanced upon your original article and the various followups today and thought you might appreciate some facts to set against your speculations. I suppose I need to introduce myself first; I'm Ed Form and I was one of the two Technical Directors of Celestion duing the period when the events you spoke of were happening.
The first, and most important comment is that the 600 and 700 products did not run in parallel at any time after the stock of 600 tweeter domes was exhausted - perhaps 6 weeks or so from launch.
As for the comment that someone inside Celestion thought the 600 a superior product: I can assure you that no such person existed! The development team was completely at one in this matter, and, remarkably, the marketing boys all agreed with us: the 700 was in every way a superior loudspeaker. The extra wall stiffness of the cabinet, due to the internal bracing shelf, notably cleaned up the upper midrange and the much lower moving mass of the aluminium tweeter allowed the system to be rebalanced to a higher overal efficiency. The elevation of the first break-up mode of the tweeter to about 24KHz from about 18KHz allowed the removal of the hand tuned equalisation network in the crossover and distinctly opened up the treble. Finally the much lower mechanical Q of the bass unit, due to the use of different materials for the inner and outer diameters of the surround, gave a new rhythm and freedom to the bass that the 600 could not even approach.
> Another point is that the state of the art dipole subwoofer
> system developed for the SL line was designed for the 600
> but not for the 700. A strange decision if the good people
> at Celestion truly believed the SL 700 to be superior.
What on earth made you make such a suggestion? The 6000 subwoofer system was developed before the 700 ever existed and was, therefore, recommended for use with the only existing loudspeakers, the SL6, SL6S, and SL600. I am aware of members of the development team who felt that the 6000 subwoofer was not suited to the SL700 system but their reason was that the crossover between the subwoofer and the top-box had a detrimental effect on the upper bass performance of the 700: in other words, the 700 was markedly superior to the electronics associated with the 6000 subwoofer and they did not feel the two items should be used together. I, on the other hand, have since used a very high performance digital crossover and equalisation system to meld 6000 and 700 units together with stunning results.
> When Celestion at a later stage decided to marked
> a digital correction module (the DLP) to flatten
> small wrinkles in the frequency response and to
> make it phase coherent, then it was the SL 600
> and not the SL 700 which got this deluxe option.
> Again a strange choice, if they really in their
> hearts, believed the SL 700 to be the better design.
The model you're referring to here wasn't the original 600! It had the stiffer, braced cabinet, the twin-material surround on the bass unit and a tweeter with a pressed copper dome, not the electroformed copper dome of the original. To put it in a nutshell it was an SL700 with a copper dome, and the reason the copper dome was reintroduced was that the marketplace, and most reviewers, in those days were convinced that metal dome tweeters sound metallic but that, somehow, the copper dome in the original SL6 and SL600 didn't. Let me explain why some tweeters sound metallic and others do not so that you can get to grips with the situation the Celestion people faced in those days...
The output of a dome tweeter consists of the sum of the outputs of its dome and its surround. The surrounds of *ALL* such tweeters, up to the point where the SL6 was introduced, exhibited a series of resonances at regular intervals where the surround was a number of half-wavelengths in width. In isolation therefore, the output from the surround itself appears to have been passed through a comb-filter, and, as any electric guitarists will tell you, a comb-filter gives the signal a metallic sound. What is worse, in 25mm dome tweeters, the normal kind in those days, the surround is larger in area than the dome itself, so the output from the tweeter is the sum of a resonance free metal dome and a dominant, metallic sounding, surround. In the case of the SL6 tweeter, the surround was absolutely tiny - less than 1mm across, so any comb-filter artifacts were pushed out of the top of the audio band and were much lower in level than the output from the dome which was increased in diameter to 32mm to further benefit the ratio between area of dome and area of surround.
In the case of soft dome units the wobbling bag of jelly called the dome fills in the comb filtering effect of the surround with a multitude of uncontrolled resonances, and sounds so naff anyway, that the overall effect is a smooth response curve and lousy dynamics with the comb-filter effects submerged in the mush.
The ultimate example of a metal dome tweeter that really sounds metallic is the 25mm beryllium-dome unit in the Yamaha NS1000 where the dome is completely devoid of resonances below 50KHz, and had no sonic chararacter whatsoever, but the surround is 3 times bigger in area than the dome and has *VERY* marked comb-filter artifacts in its output because it is so wide. It sounds, to be blunt, dreadful, but a modified version I made, with a 0.7mm wide surround, sounds quite wonderful - the boys at Focal and at Usher have demonstrated rather nice examples in recent years, although they haven't fully grasped the surround resonance problem yet.
Sorry about the long diversion, but it was necessary to explain the truth behind the very real consumer opinion that metal domes sound metallic. When the copper SL6 tweeter appeared the entire universe jumped on the copper as the reason, and were certainly never told that it was the tiny surround that gave the good sound - I'm not even sure that my colleague Boaz EliEli, the actual designer of the unit, fully appreciated it! When the aluminium dome appeared in the SL6s, and then the SL700, the misconception continued and no amount of persuasion would change it - even though John Atkinson, the editor of Stereophile, adopted the 700 as his primary reference, and personal choice for listening to, and continued with it for a lot of years afterwards.
By this point I had left the company to go to Tannoy, and was shocked by the change in marketing direction that the company took: they simply gave in and went back to calling the products 6 and 600, and reintroduced copper for the domes because they thought the 'fickle' public would buy it. What actually happened was that the sales of the whole range dried up because they sounded really wimpy, and the company started on a series of truly awful excursions into marketing-lead nonsense - who remembers the dreadful 5000 series with their useless ribbon tweeter? What the company should have done was to actively promote the truth about dome tweeters. As part of my tenure as Technical Director I had the backroom boys develop a new laser system that would have allowed the separate frequency resonses of domes and surrounds to be listened to. Why the tech boys didn't use this to kill the daft idea that the copper dome was somehow magic once and for all, I cannot imagine.
> You write that they just could have adjusted the
> treble level in the crossover. This obviously is
> perfectly true. I however believe that they were
> blinded by their own measurements; they started
> out to make a frequency linear version of the SL
> 6/600 and probably mentally got stuck when this
> turned out to be to bright sounding. No one seems
> to have suggested that the entire idea of giving
> the speaker a flat frequency response was at fault.
> Their ears might have told them to lower the
> treble level, but as good engineers they felt
> obliged to follow their instruments, and
> therefore declined to do so!?
Nonsense Klaus. The 700 was in every way superior to the 600, and, in my carefully observed opinion, also superior to the 600si. Optimising all three with digital electronic crossovers and signal conditioning, the superiority of the aluminium tweeter, and the flat response approach, is glaringly obvious. If you live anywhere near East Anglia you're welcome to come and hear - although I'd need some notice because I've given away my SL600 and SL600si boxes and would need to borrow them from friends who acquired them. You should also note that the SL series of products were meant to be sited well away from walls and the floor and listened to in the quasi-nearfield so the balance of the room was not significant in their perceived sound. You should also bear in mind that the great man whose work on room sound you quoted in the original article appears to have cloth ears and never produced a good sounding domestic loudspeaker in his life.
> What they did however was to push the crossover
> frequency up from 2200 Hz to 3000 Hz, this will
> obviously reduce the dispersion in this range
> and put less lower treble into the listening
> room. This however was probably jut another nail
> in the Celestion coffin, because it did increase
> the incoherence in dispersion between mid drive
> and treble at the crossover frequency.
You know the dispersion of the one-piece cone of the SL600 bass unit at 3KHz? You know the dispersion of the 32mm dome at 3KHz? What the revised tweeter crossover frequency, and the higher tweeter crossover slope did was to cut the excursion expected from the tweeter at the lower extreme of its passband. The crossover was the work of my greatly esteemed colleague, Bob Smith, and he came as near to a set of golden ears [it's a BBC term of fairly obvious meaning] as I have ever known. The stability of the polar pattern of the loudspeaker through the crossover region was distinctly better than for the original 600 because the two units were very well matched throughout that region.
> I truly believe the Celstion SL 700 could have been
> a winner if it had had its treble level adjusted to
> the same level as the 600, and had had a it
> crossover frequency point lowered (maybe to something
> like 1700 Hz) rather than lifted.
But that would have meant increasing the width of the tweeter surround to allow sufficient excursion and would have completely ruined the sound quality by reintroducing the dreaded comb-filter problem.
> The lessons to be learned are to take something
> for granted and learn to think outside the box.
> Charles Hansen deliberately did so when he gave
> the Avalon Eclipse a crossover frequency at a
> stunningly low 1080 Hz. I salute him for that,
> GREAT!! Celestion did the same with the
> frequency response of the original SL line, but
> it wasn’t a deliberate design decision and for
> this reason they were unable to capitalise in
> their following design development. SAD!!
The choice of 2K3 for the original SL6, 600 series models was a deliberate decision based on the performance of the tweeter and bass units, as was the slightly higher frequency, but steeper slope, of the 700. The difference between 2K3 and 3K has no significance, since both lie very firmly at the maximum sensitivity of the ear and the polar pattern of the bass unit will change only marginally. Moving the crossover down to 1K, as you claim for the Avalon model, is not a major advantage because the ear is only 3 to 4 dB less sensitive there than it is at 3K. To truly make a difference that the ear can hear, the crossover would need to be set at 300Hz or so, where aural acuity has fallen 6 to 8dB and, much more importantly, sensitivity to rapid changes of polar pattern has also diminished substantially - because the wavelength is *much* more than the distance between the ears. The big problem for any design like the Avalon Ecclipse, if it's 1 inch tweeter does actually go down to 1K - I can't find that figure quoted anywhere - is that no 1 inch tweeter capable of 1K crossover has ever been made and the MB tweeter they used was certainly not capable of it.
Ed Form
This post is made possible by the generous support of people like you and our sponsors:
Follow Ups
- RE: Deliberate or not. - edform 13:53:21 06/23/09 (0)
- RE: Deliberate or not. - Karma16 04:02:45 06/24/09 (0)
Top of Page ]
[ Contact Us ]
[ Support/Wish List ]
[ Copyright © 2024, Audio Asylum, all rights reserved ]
