I just reminded myself about the KEF Constructor designs.
The CS1A was recessed B110A and negative polarity.
The CS1 was flush mounted B110B, a later driver, and positive polarity.
IIRC, BW3 has different group delay and lobing depending on polarity. John Kreskovsky knows this sort of stuff.
I wouldn't knock light ply and damping panels really. But KEF used a bigger 8L box. It works.
BTW, David Barnett's variant has a glaring error. That 4R/8R attenuator before the tweeter filter is just going to burn up bass power. I also looked at the original's impedance. It's terribly bumpy.
The CS1A was recessed B110A and negative polarity.
The CS1 was flush mounted B110B, a later driver, and positive polarity.
IIRC, BW3 has different group delay and lobing depending on polarity. John Kreskovsky knows this sort of stuff.
I wouldn't knock light ply and damping panels really. But KEF used a bigger 8L box. It works.
BTW, David Barnett's variant has a glaring error. That 4R/8R attenuator before the tweeter filter is just going to burn up bass power. I also looked at the original's impedance. It's terribly bumpy.
Does it bear repeating that the LS3/5 and subsequent sanctioned derivatives was designed for a very specific application, and that much of the very elaborate XO work was to mitigate for the known performance characteristics of the chosen available drivers, as well as to achieve a targeted near-field response curve in very confined space.
The particular construction methods chosen for the enclosure should be considered as a separate matter, and similar approaches have been used by many commercial models and DIY builds since.
Certainly modern drivers offer many advantages, that include the ability to simplify the task of the filter network.
There are numerous factors involved in the design and implementation of a loudspeaker system, but shouldn't the prime objective of that process be to define the application and venue?
The particular construction methods chosen for the enclosure should be considered as a separate matter, and similar approaches have been used by many commercial models and DIY builds since.
Certainly modern drivers offer many advantages, that include the ability to simplify the task of the filter network.
There are numerous factors involved in the design and implementation of a loudspeaker system, but shouldn't the prime objective of that process be to define the application and venue?
@ system7 : I have both sp1057 and sp1003.
drivers now mounted are sp1057 so , you are saying : I have to switch the drivers ?
I did a Q&D build so NO flushmount !!
The Original question was about making things better ?
What I did is good for bass improvement.
Maybe someone else can do XO adjusts ?
We work all together , making something better?
There is a cloning site already so .....?
drivers now mounted are sp1057 so , you are saying : I have to switch the drivers ?
I did a Q&D build so NO flushmount !!
The Original question was about making things better ?
What I did is good for bass improvement.
Maybe someone else can do XO adjusts ?
We work all together , making something better?
There is a cloning site already so .....?
As far as I know these were designed and built under licence for Outside Broadcast Monitoring.
Any sound producer could jump into any OB van with any pair of LS3/5 monitoring speakers and hear the same performance repeatedly and reliably.
And it was to achieve this consistency that the cost was so high.
I think the audiophile reviewers jumped on the Hi-cost=superlative performance formula and wrongly recommended these for a completely different duty.
They were superb as OB monitors that were consistent.
Andrew - thanks for reinforcing my point. To meet the targeted design objective, it's proably safe to say they were a great success.
I heard a pair of either Rogers or Chartwells version back in the late 70s / early 80s, and wasn't particularly gobsmacked. Only a few years later as part of my apprenticeship in the Golden Ear Club, membership to which I've long since been unable to afford the maintenance costs, I owned a pair of the little KEF 101s. They were nice enough in their own way, but for my money the Spica SC50 ate their lunch.
I heard a pair of either Rogers or Chartwells version back in the late 70s / early 80s, and wasn't particularly gobsmacked. Only a few years later as part of my apprenticeship in the Golden Ear Club, membership to which I've long since been unable to afford the maintenance costs, I owned a pair of the little KEF 101s. They were nice enough in their own way, but for my money the Spica SC50 ate their lunch.
Perhaps I should have said rear-mounted or front mounted bass/mid driver. 😱
It's a bit like trying to unlock the secrets of the ancients. IIRC, our old friend speakerDave did a lot of work on the KEF versions of this idea.
AFAIK, the original B110A was a much more raggedy response, but a lower mechanical loss around 6 as Qms. It needed more crossover work and a deeper notch at 1kHz to get flat. And rear mounted on the baffle and negative polarity. The KEF bass filter is specific to the bass driver, the KEF T27 filter is your Fs 1.2kHz notch and seems to be set in stone. Only polarity varies.
Back in the day, it was believed that if you equalised everything flat, you got a good sound. These days we think mechanical loss and excursion causes distortion.
For sure, a 5" bass and 1" tweeter play nicely together. In a small room. I'd like to use Monacor's software and drivers to see what I can do with BW3 and the various bass and tweeter notches. Can't be hard, can it? 🙄
Last edited:
Agree that the 4/8 ohm resistors could cause problems, but it has not been something that stood out as an issue as I listened to these for almost 40 years. That said probably an 8 ohm LPad just before the tweeter would work better. That is what I did on a large pair of monitors I built.
The drivers I have been using are non-constructor series T27 (1032) and B110 (1003). The constructor series T27 has two black wires on the front whereas the non version has a black and blue wire.
The drivers I have been using are non-constructor series T27 (1032) and B110 (1003). The constructor series T27 has two black wires on the front whereas the non version has a black and blue wire.
Last edited:
Now that I think about it. The LPad in the position I suggested will change the electrical characteristics (of the tweeter) that the filter was designed for and probably mess things up.
Last edited:
What about adding a 10uF capacitor into the tweeter line before the signal reaches the 12ohm Lpad?
How big would this DC blocking cap need to be to avoid any effect on the treble relative to the Bass/Mid?
How big would this DC blocking cap need to be to avoid any effect on the treble relative to the Bass/Mid?
I see what you are getting at - essentially a "pre-filter" to ward off the lower frequencies that will be "burned off" in the LPad. Would have to think more about the impact of such a change. Certainly something like this could be determined experimentally to see what impact if any is heard.
It is the most common method adopted by most crossover designers.
Surprising that it was omitted from this design.
And worse that it was missed by the editorial team before publication, unless it was a design decision, in which case it's rarity would have attracted a comment by the designer.
Surprising that it was omitted from this design.
And worse that it was missed by the editorial team before publication, unless it was a design decision, in which case it's rarity would have attracted a comment by the designer.
If we assume that roughly 8 ohms is reflected back by the original filter then the pre-filter would see approximately 4 + (8 in parallel with 8).
That and the intended corner frequency could be used to determine the capacitor size.
Last edited:
that would make the capacitor act as a filter.
An added filter in that position will be clearly and audibly different.
If you want it to audibly pass all the treble signal, then the capacitor will need to be bigger.
An added filter in that position will be clearly and audibly different.
If you want it to audibly pass all the treble signal, then the capacitor will need to be bigger.
At least in my example we could answer the question of "bigger than what" fairly easily.
I usually see 30 uF or bigger value used. But I never see the logic of this blocking cap. Yes, it will have audibe effect to the treble.
A high end 30 uF cap will cost a small fortune. A non-polarized electrolytic will probably do more harm than helping. 😕
I am actually not even a fan of Butter-worth filters. My latest design have used fourth order linkwitz-riley with great results. Much better behavior in the time domain. Down the road I may try such a filter with these mini-monitors.
Falcon Acoustics sell a crossover kit combining the LS3/5A low-pass with the KEF aB high-pass, one would hope they have dealt with the tweeter attenuation issue correctly. No schematic is published, although surely the documentation that comes with the kit must include one.
LS3/5a BI-WIRE CROSSOVER 15 OHM KEF B110 SP1003 T27 SP1032, aB HF SECTION
LS3/5a BI-WIRE CROSSOVER 15 OHM KEF B110 SP1003 T27 SP1032, aB HF SECTION
Motor run capacitors rated for 220/240Vac mains and made with polypropylene film make very good AC coupling capacitors and they are reasonably cheap because they are not priced to suit the audio market.
And if the coupling capacitor has no treble content voltage across it then one should not have any audible treble effects.
Any distortions of the LF content will be attenuated by the treble filter/s.
That's why one should start with a good audio duty capacitor, rather than an MKT or electrolytic.
And that leads me back to why it was omitted. Had it been included, it would have given one pole of the 3 pole Butterworth. At no extra cost this cap could just have been moved from one location to another.
Last edited: