Rogers have re-introduced the LS3/5A. For £1495. But the following quote interests me. John Bell one of the co-designers said:
"Getting a response curve to match the LS3/5A's was quite easy, but getting the sound was not, so a number of different drive unit combinations were tried"
Any thoughts?
"Getting a response curve to match the LS3/5A's was quite easy, but getting the sound was not, so a number of different drive unit combinations were tried"
Any thoughts?
A response curve doesn't really take into account dispersion characteristics, phase and distortion.
It's actually old news:
Rogers LS3/5A (£1495) - Loudspeakers
You can study the small monitor design at the The Unofficial LS3/5A Support Site. There's a lot to learn about it. It has an amazingly clear vocal midrange. I heard some many years ago. Everything was carefully thought out for accuracy and consistency of performance.
I wouldn't get too sentimental about an old loudspeaker with no real bass using flawed drivers, but it's an object lesson in good engineering. I use that KEF-style negative polarity filter to this day. It's a classic. 🙂
Rogers LS3/5A (£1495) - Loudspeakers
You can study the small monitor design at the The Unofficial LS3/5A Support Site. There's a lot to learn about it. It has an amazingly clear vocal midrange. I heard some many years ago. Everything was carefully thought out for accuracy and consistency of performance.
I wouldn't get too sentimental about an old loudspeaker with no real bass using flawed drivers, but it's an object lesson in good engineering. I use that KEF-style negative polarity filter to this day. It's a classic. 🙂
The usual hyperbole.
I was the engineer on the 11 Ohm redesign project at KEF back in the 80s. We changed B110 types to something that was more long term stable, then I revised the network through a number of iterations. A number of versions were evaluated by the BBC and when I got the response curve to match the chosen reference, the BBC felt that it was a very near sonic match.
I even offered a flatter version (less of the raised mound below 1k) and the BBC liked the sound of it but didn't want to muddy the waters with a new version too far removed from units already in the field.
David S.
I was going to PM you, Dave, but your box is full.
Slightly off-topic, but was the KEF acoustic butterworth any good?
I couldn't get my version sounding good, and concluded it let through a lot of bass energy causing IM distortion by simming it.
I subsequently thought that maybe I had made a mistake in using an attenuator which took the impedance a bit low for my modest Rotel amp.
Be interested to hear your take on that. 🙂
Slightly off-topic, but was the KEF acoustic butterworth any good?
I couldn't get my version sounding good, and concluded it let through a lot of bass energy causing IM distortion by simming it.
I subsequently thought that maybe I had made a mistake in using an attenuator which took the impedance a bit low for my modest Rotel amp.
Be interested to hear your take on that. 🙂
Hard to say that it was inherently good or bad. It was more a philosophical realization that when we deemed something to be "3rd order" or "4th order" then it really only made sense to be referring to the full network plus driver as a summed pair. A Butterworth electrical input was of no particular significance if the combined result was anything but.
The key design challenge was dealing with the underlying impedance curve. For example a 2nd order electrical high pass on a tweeter would be very roughly 4th order in total, but would have a distinct inflection (kink) in its output at tweeter resonance. Laurie and the others came up with some kind of bridged Tee topology that put a mild notch at tweeter resonance to counteract the impedance bump. (This is shown in the relevant KEF Topic, available at their web site.)
In later years we just went to a standard topology and let the optimizer drag the curve to the right slope. If anything this acoustic Butterworth approach should reduce the drive to the tweeter as the rise of crossover output at driver resonance would be down a bit, but that may not be universally so.
Imagine, and this was about 1980, being able to sit somebody down that knew a little about loudspeaker design, show them an ideal 4th order AB target, toss in a network topology and some starting values and then after half a minute of chugging the computer spits out values that got the network and tweeter combined to well within a dB of on-target perfection. This was quite an achievement in its day.
Regards,
David
Yes, the flatness of the curves was awesome! 😎
FWIW, here's the analysis I did that raised the problem of bass breakthrough:
http://www.diyaudio.com/forums/multi-way/97043-elsinore-project-thread-175.html#post3895980
I don't think anyone was interested. But I found it sounded better if I took out the acoustic butterworth Fs cap and "Q" resistor and did plain old 3rd order butterworth.
Thanks.
FWIW, here's the analysis I did that raised the problem of bass breakthrough:
http://www.diyaudio.com/forums/multi-way/97043-elsinore-project-thread-175.html#post3895980
I don't think anyone was interested. But I found it sounded better if I took out the acoustic butterworth Fs cap and "Q" resistor and did plain old 3rd order butterworth.
Thanks.
I can see what you mean about it looking ultimately like a first order plus notch (and third order acoustic will always be 1st order electrical below the corner), but there is nothing to keep you from achieving a higher effective order. At KEF we would always use at least 2nd order electrical, plus whatever extra was required to hit an AB shape (or more usually an LR shape).
David
David
Speaker Dave, I applaud the efforts you and Laurie expended back then. I was a dealer for KEF, and still use my 101s and 103.2... Happily.
103.2 was a very good speaker. I gave a pair to a friend and he still uses them happily, as well.
Cheers!
Cheers!
I used to have two pair of Chartwell LS3/5 A's which I was told were better sounding than the Rogers version. I finally sold one off to a friend who was begging me to sell it. Once he got it, he promptly blew it up. The pair I still have, I completely redid the xover capacitors with polypropylene where I matched all the original cap values within 1% - significantly improved inner detail, btw and I recall I had to replace the woofers once. At $1495 for the new ones I wish I still had that second pair to sell of (it was bone stock).
I believe part of the original LS3/5A sound was claimed to be due to the type and thickness of the plywood used, and perhaps bitumen dampening. The woofer mounting behind the panel and the dampening ring around the tweeter were also supposed to be part of it. Even the grillcloth was supposed to be part of the equation, made out of some material they stopped manufacturing a while back, although I always liked the sound of my Chartwells better without the grillcloth on.
I believe part of the original LS3/5A sound was claimed to be due to the type and thickness of the plywood used, and perhaps bitumen dampening. The woofer mounting behind the panel and the dampening ring around the tweeter were also supposed to be part of it. Even the grillcloth was supposed to be part of the equation, made out of some material they stopped manufacturing a while back, although I always liked the sound of my Chartwells better without the grillcloth on.
Last edited:
I visited Chartwell's factory in Mitcham, Surrey many years ago for a speaker repair. Very small outfit actually. 3 or 4 people IIRC. The carpenter's workshop took up most of the space.
It's surprising if you look up the B110 SP1003 bass used that it was a Qts around 0.31. So that box was acoustic suspension. I'd agree that everything about it was critically balanced on the lossy damped birch plywood construction with separate front baffle secured by screws, the "cracked bell" that doesn't ring.
The crossover is interesting. It has huge 4mH bafflestep which makes it bassy. The tweeter is wired negative polarity as is common with many BBC designs. That is why the woofer is recessed and mounted on the back of the baffle approximately half a wavelength at 3.5kHz crossover.
The 1.5kHz LCR midrange notch is the only real complication, the autotransformer on the treble being a level adjustment. It might as well be a coil. The RC Zobel on the tweeter shouldn't actually do much, but corrects impedance so the filter is more accurate.
I must have another go at simming this one. Jeff Bagby studied it very deeply before attempting the Continuum.
It's surprising if you look up the B110 SP1003 bass used that it was a Qts around 0.31. So that box was acoustic suspension. I'd agree that everything about it was critically balanced on the lossy damped birch plywood construction with separate front baffle secured by screws, the "cracked bell" that doesn't ring.
The crossover is interesting. It has huge 4mH bafflestep which makes it bassy. The tweeter is wired negative polarity as is common with many BBC designs. That is why the woofer is recessed and mounted on the back of the baffle approximately half a wavelength at 3.5kHz crossover.
The 1.5kHz LCR midrange notch is the only real complication, the autotransformer on the treble being a level adjustment. It might as well be a coil. The RC Zobel on the tweeter shouldn't actually do much, but corrects impedance so the filter is more accurate.
I must have another go at simming this one. Jeff Bagby studied it very deeply before attempting the Continuum.
Attachments
I probably created a few stripped screw holes on my LS3/5As with the times I've had them open.
In a way, it's fascinating that such a small outfit as Chartwell could be licensed to produce such a well -designed and -regarded audio product.
In a way, it's fascinating that such a small outfit as Chartwell could be licensed to produce such a well -designed and -regarded audio product.
Last edited:
Well, it's been a few years since I heard them, Mr. Thoriated, but i remember they sounded like a much bigger speaker than they were. And a clear and detailed midrange that I had only previously heard on Quad electrostatics. 🙂
How do you find them compared to modern small speakers? Can you still say they are better? I always find the peaky bass of minis a bit annoying.
Can you put your finger on what is good here?
How do you find them compared to modern small speakers? Can you still say they are better? I always find the peaky bass of minis a bit annoying.
Can you put your finger on what is good here?
Darn. I had written a fairly long response, and it got deleted. Basically, I mentioned that I felt like the LS3/5a generally played like a somewhat larger speaker than it appeared to be wthin its dynamic limits and the first sign of significant distress would typically be a loudly bottoming woofer which would send me rushing to cut the volume some. that's probably better than some other speakers that would noticeably distort or compress first.
The little 80 hz bass hump was probably just enough to keep the LS3/5A from sounding notably bass deficient. I wonder if the baffle step inductor would also begin to saturate at high levels somewhat which might temporarily boost the midrange slightly, almost adding a hint of 'loudness compensation' and 'dynamic expansion'. I believe I read somewhere that the xover also would attenuate very high levels to the tweeter to help protect it during tape rewinds, etc. which would probably reduce the most obnoxious HF peaks a bit.
I believe the B110 (and B200) were relatively rugged drivers. I once created subwoofer cabs with B200's at the bottom to work with my LS3/5A's, and they actually were much more reliable being driven by a Hafler DH220 than a set of much beefier looking Audax woofers with much larger voice coils, two sets of which I burned out the voice coils of in the subwoofer position in about three months.
The B110 may have good detailing partly because of all the damping material added to suppress the major bextrene resonance also. By comparison, I was disappointed by a Dynaudio MW-175 reproducer I used in a home-built design which couldn't seem to get microdetails right but dynamically was much better than a B110 in most other ways - I convinced myself finally that it was not a FR issue but related to the MW-175 cone construction.
I'm convinced that the matched xover and driver parameters of pairs of LS3/5A's significantly aids in its soundstage and imaging reproduction, considering the above factors.
The enclosure likely adds just a bit of 'dimension' to the sound without having enough coloration to stand out as a distinct audio source.
The little 80 hz bass hump was probably just enough to keep the LS3/5A from sounding notably bass deficient. I wonder if the baffle step inductor would also begin to saturate at high levels somewhat which might temporarily boost the midrange slightly, almost adding a hint of 'loudness compensation' and 'dynamic expansion'. I believe I read somewhere that the xover also would attenuate very high levels to the tweeter to help protect it during tape rewinds, etc. which would probably reduce the most obnoxious HF peaks a bit.
I believe the B110 (and B200) were relatively rugged drivers. I once created subwoofer cabs with B200's at the bottom to work with my LS3/5A's, and they actually were much more reliable being driven by a Hafler DH220 than a set of much beefier looking Audax woofers with much larger voice coils, two sets of which I burned out the voice coils of in the subwoofer position in about three months.
The B110 may have good detailing partly because of all the damping material added to suppress the major bextrene resonance also. By comparison, I was disappointed by a Dynaudio MW-175 reproducer I used in a home-built design which couldn't seem to get microdetails right but dynamically was much better than a B110 in most other ways - I convinced myself finally that it was not a FR issue but related to the MW-175 cone construction.
I'm convinced that the matched xover and driver parameters of pairs of LS3/5A's significantly aids in its soundstage and imaging reproduction, considering the above factors.
The enclosure likely adds just a bit of 'dimension' to the sound without having enough coloration to stand out as a distinct audio source.
Last edited:
The most significant design decision was the negative polarity filter IMO. It would have worked better on paper with a positive polarity. In fact, all that depends on is how much you recess the bass, the filter itself is unaffected.
Modelling it now with modern drivers, the big 4mH bafflestep makes for a peak in the bass. I got something more balanced with about 2.5mH overall. I kept the midrange notch which suits many 5" drivers, and it was very flat indeed for a 3.5kHz crossover.
The way the BBC aligned phase, it is good at crossover, but wanders off away from there. There may be something to be said for that. The cone breakup tends to null, rather than reinforce. It's a thought!
I might as well give you the LS3/5A sim I did. It's interesting for crossover enthusiasts. 🙂
I used paper bass and a 8L volume, which is neither here nor there for the essential features. I played with the time alignment too.
You can see the negative polarity filter works well. Recessing the bass is problematic in practise though. With the lower bafflestep and the shallow chassis I used, it needs near on 5cm recess for negative polarity. That's not going to be doable unless we have a brainwave. 😀
You can see the cone breakup at around 6kHz behaves 4dB differently with the two designs. That's a lot.
The result is very similar to the original, especially on the double hump midrange impedance, the top end being just a question of filter damping:
BBC LS3/5a loudspeaker 1989 Measurements | Stereophile.com
I've gotta conclude I don't see how to recreate this BBC LS3/5A classic right now. Very hard to find a suitable modern driver too. But it's always good to understand the problem anyway. 😎
I used paper bass and a 8L volume, which is neither here nor there for the essential features. I played with the time alignment too.
You can see the negative polarity filter works well. Recessing the bass is problematic in practise though. With the lower bafflestep and the shallow chassis I used, it needs near on 5cm recess for negative polarity. That's not going to be doable unless we have a brainwave. 😀
You can see the cone breakup at around 6kHz behaves 4dB differently with the two designs. That's a lot.
The result is very similar to the original, especially on the double hump midrange impedance, the top end being just a question of filter damping:
BBC LS3/5a loudspeaker 1989 Measurements | Stereophile.com
I've gotta conclude I don't see how to recreate this BBC LS3/5A classic right now. Very hard to find a suitable modern driver too. But it's always good to understand the problem anyway. 😎
Attachments
- Status
- Not open for further replies.