Not the LS3/5A but the LS5/5X!

[Purité Audio]

Phil Ward, who recently reviewed AsciLab’s A6B for ‘SoundonSound ‘ magazine,

AsciLab A6B

With their debut speakers, South Korean company AsciLab have raised the bar for nearfield monitoring.

www.soundonsound.com

Has been working on a contemporary version of of the LS3/5A, he asked me to post his post with the hope that it might engender some useful feedback.

https://musicandmiscellany.com/2026/04/06/not-the-ls3-5a-but-the-ls3-5x/

 

[Soniclife]

Phil Ward, who recently reviewed AsciLab’s A6B for ‘SoundonSound ‘ magazine,

AsciLab A6B

With their debut speakers, South Korean company AsciLab have raised the bar for nearfield monitoring.

www.soundonsound.com

Has been working on a contemporary version of of the LS3/5A, he asked me to post his post with the hope that it might engender some useful feedback.

https://musicandmiscellany.com/2026/04/06/not-the-ls3-5a-but-the-ls3-5x/

Be interesting to do how it compares to the Fidelia when it's finished.

 

[DavidMcRoy]

The description of the proposed tweeter to be used is strikingly similar to a 30mm matched pair I have for a large DIY project under development:

Dayton Audio - High Resolution Ceramic Dome Tweeter 30mm 4 Ohm Matched Pair EC30MP-4

High Resolution Ceramic Dome Tweeter 30mm 4 Ohm Matched Pair EC30MP-4

www.daytonaudio.com

And I guess it's this:

SB26CDC-C000-4 / Ceramic - Sbacoustics

Aluminium/Ceramic dome, phase optimized diaphragm design for coherent high frequency radiation, Saturation controlled motor system for low distortion.

sbacoustics.com

 

[jhaider]

Finally, a design with grilles!

 

[kemmler3D]

Purifi + waveguide is promising in most instances. I think it's neat that he's tying it back to a vintage design / aesthetic but still taking advantage of modern tech.

 

[Puddingbuks]

Finally, a design with grilles!

Would even be better with a waveguide.

 

[DavidMcRoy]

Would even be better with a waveguide.

It appears to be molded into the baffle.

 

[Philberish]

The description of the proposed tweeter to be used is strikingly similar to a 30mm matched pair I have for a large DIY project under development:

Dayton Audio - High Resolution Ceramic Dome Tweeter 30mm 4 Ohm Matched Pair EC30MP-4

High Resolution Ceramic Dome Tweeter 30mm 4 Ohm Matched Pair EC30MP-4

www.daytonaudio.com

And I guess it's this:

SB26CDC-C000-4 / Ceramic - Sbacoustics

Aluminium/Ceramic dome, phase optimized diaphragm design for coherent high frequency radiation, Saturation controlled motor system for low distortion.

sbacoustics.com

Clearly I can't sneak an SB25CDC tweeter past you lot. Yes, it's that one. Seems to me it inhabits a cost/performance sweet spot in contemporary tweeters. Apparently I'm not the only one to conclude that.

It appears to be molded into the baffle.

That it is. There's also a bit more waveguide hidden behind the tweeter grille.

 

[Philberish]

Purifi + waveguide is promising in most instances. I think it's neat that he's tying it back to a vintage design / aesthetic but still taking advantage of modern tech.

That's kind of the basis of the idea. Whatever you think of the LS3/5A now (personally, I understand why folk still appreciate its subjective character, but I do think we can do significantly better now), it was a ground-breaking design back in the early/mid 1970s. And while the LS3/5A's driver tech has been well surpassed, its cabinet construction technique (fully battened thin-wall birch ply with damping panels) is actually still pretty valid. Anyway, I'm hoping to have a representative working pair within the next month or so. That'll be the moment of truth

 

[Matt_Holland]

Phil Ward, who recently reviewed AsciLab’s A6B for ‘SoundonSound ‘ magazine,

AsciLab A6B

With their debut speakers, South Korean company AsciLab have raised the bar for nearfield monitoring.

www.soundonsound.com

Has been working on a contemporary version of of the LS3/5A, he asked me to post his post with the hope that it might engender some useful feedback.

https://musicandmiscellany.com/2026/04/06/not-the-ls3-5a-but-the-ls3-5x/

Following this with interest!

@Philberish very keen to learn more about your design goals for things like bass alignment, directivity, boundary conditions and linearity.

 

[kharan]

What kind of sensitivity are you hoping to get?

The Radiant Acoustics 4.2 is barely smaller, and it already drops to 78 dB @2.83V.

 

[Philberish]

Following this with interest!

@Philberish very keen to learn more about your design goals for things like bass alignment, directivity, boundary conditions and linearity.

LF alignment is closed box with a Qt around 0.65 at 55Hz. I was tempted by the idea of a rear mounted ABR but the thing really wouldn't be anywhere near an LS3/5 then. Also, subjectively, the bass from my mono prototype sounds really good, so I'm leaving well alone. I'm using the 8Ω version of the Purifi so baseline mid band sensitivity, with the crossover series R taken into account, will be around 82dB/2.83V 1m (somewhat depending on how close I voice the LS3/5X towards the LS3/5A - with its recessed lower mid).

With its low(ish) Qt the system will be happiest subjectively used relatively close to a rear wall, but then, I've never really seen the point of a very compact speaker used out in the room - why not just have a bigger speaker?

The crossover frequency is around 1.5kHz so, also considering the waveguide, horizontal directivity is nicely linear. It gets slightly messy above 10kHz where the dome radiation tends to decouple from the waveguide (I'm constrained by the LS3/5 dimensional DNA to use a relatively shallow/modest waveguide - it's about 15mm deep) but I'm not too concerned about stuff above 10kHz.

Vertical directivity is biased towards optimum slightly above axis (that's really just a matter of getting the driver phase integration optimised through the crossover region). Finally, I'm hoping for THD getting towards -60dB (at say 85dB/1m) above 150Hz. That's very, very much better than an LS3/5A and primarily defined by the drivers - which as we all know, aren't too shabby.

Hope that all helps.
P

 

[Matt_Holland]

LF alignment is closed box with a Qt around 0.65 at 55Hz. I was tempted by the idea of a rear mounted ABR but the thing really wouldn't be anywhere near an LS3/5 then. Also, subjectively, the bass from my mono prototype sounds really good, so I'm leaving well alone. I'm using the 8Ω version of the Purifi so baseline mid band sensitivity, with the crossover series R taken into account, will be around 82dB/2.83V 1m (somewhat depending on how close I voice the LS3/5X towards the LS3/5A - with its recessed lower mid).

With its low(ish) Qt the system will be happiest subjectively used relatively close to a rear wall, but then, I've never really seen the point of a very compact speaker used out in the room - why not just have a bigger speaker?

The crossover frequency is around 1.5kHz so, also considering the waveguide, horizontal directivity is nicely linear. It gets slightly messy above 10kHz where the dome radiation tends to decouple from the waveguide (I'm constrained by the LS3/5 dimensional DNA to use a relatively shallow/modest waveguide - it's about 15mm deep) but I'm not too concerned about stuff above 10kHz.

Vertical directivity is biased towards optimum slightly above axis (that's really just a matter of getting the driver phase integration optimised through the crossover region). Finally, I'm hoping for THD getting towards -60dB (at say 85dB/1m) above 150Hz. That's very, very much better than an LS3/5A and primarily defined by the drivers - which as we all know, aren't too shabby.

Hope that all helps.
P

82dB average sensitivity for an 8ohm small speaker seems perfectly decent. As you say will depend on midrange voicing. My take would be to not push too much towards the original’s upper mid peakiness only because this can be achieved using EQ. Go more for linearity with excellent directivity.
Agree regarding bass alignment too. Better to optimise for use against a wall. What’s the predicted F3 of the driver/box combo you’re using?

 

[Lars Risbo]

LF alignment is closed box with a Qt around 0.65 at 55Hz. I was tempted by the idea of a rear mounted ABR but the thing really wouldn't be anywhere near an LS3/5 then. Also, subjectively, the bass from my mono prototype sounds really good, so I'm leaving well alone. I'm using the 8Ω version of the Purifi so baseline mid band sensitivity, with the crossover series R taken into account, will be around 82dB/2.83V 1m (somewhat depending on how close I voice the LS3/5X towards the LS3/5A - with its recessed lower mid).

With its low(ish) Qt the system will be happiest subjectively used relatively close to a rear wall, but then, I've never really seen the point of a very compact speaker used out in the room - why not just have a bigger speaker?

The crossover frequency is around 1.5kHz so, also considering the waveguide, horizontal directivity is nicely linear. It gets slightly messy above 10kHz where the dome radiation tends to decouple from the waveguide (I'm constrained by the LS3/5 dimensional DNA to use a relatively shallow/modest waveguide - it's about 15mm deep) but I'm not too concerned about stuff above 10kHz.

Vertical directivity is biased towards optimum slightly above axis (that's really just a matter of getting the driver phase integration optimised through the crossover region). Finally, I'm hoping for THD getting towards -60dB (at say 85dB/1m) above 150Hz. That's very, very much better than an LS3/5A and primarily defined by the drivers - which as we all know, aren't too shabby.

Hope that all helps.
P

hi Phil, very interesting! regarding the sensitivity. The 82dB/8 ohms seem quite high. Did you include the baffle step loss?

cheers

Lars

 

[Matt_Holland]

hi Phil, very interesting! regarding the sensitivity. The 82dB/8 ohms seem quite high. Did you include the baffle step loss?

cheers

Lars

Can the speaker be optimised for against wall use and offset baffle step loss to some degree?

 

[Philberish]

hi Phil, very interesting! regarding the sensitivity. The 82dB/8 ohms seem quite high. Did you include the baffle step loss?

cheers

Lars

It's a kind of yes and no answer to that Lars. The 2pi/2.83V sensitivity specified for the driver is 83.1dB (as you well know . I loose say, a dB, of that due to the DCR of the series crossover inductor(s) which puts me nominally at 82dB (above the baffle step). However, I think the speaker is also almost certainly going to find itself in use very close to a wall, so thinking generally in the 2pi world seems a reasonable assumption. Hence 82dB seemed reasonable to me.

The complexity here though is that in doing a modern interpretation of the LS3/5A I have to decide how close to voice the LS3/5X to the (somewhat idiosyncratic) original, and a lot of the action in that respect falls either side of the baffle step region. I'm still messing with the voicing (not least 'cause I only have a single speaker working at the moment and thing are different in stereo) but the effective sensitivity will be somewhat defined by how the voicing ends up.

 

[Philberish]

82dB average sensitivity for an 8ohm small speaker seems perfectly decent. As you say will depend on midrange voicing. My take would be to not push too much towards the original’s upper mid peakiness only because this can be achieved using EQ. Go more for linearity with excellent directivity.
Agree regarding bass alignment too. Better to optimise for use against a wall. What’s the predicted F3 of the driver/box combo you’re using?

That upper mid peakiness of which you speak Matt is I think what's traditionally known as the "Bextrene quack". It's quite audible on the LS3/5A and I think for some folks is part of the appeal - despite the fact that in objective terms it's not great. The interesting part though is how it's masked somewhat (and maybe even rendered seductive) by the warmth that comes from the LS3/5A's underdamped LF alignment and, I suspect, some quite high levels of upper bass/low mid 2nd harmonic distortion.

The odd thing also about the Bextrene quack (well, I've always thought it was odd) is that it's probably not really about the Bextrene. I'm pretty sure it's a surround mis-termination problem (no least 'cause I experienced a very similar thing with a very similar tonal characteristic back in my Naim days on a midrange driver that had a paper diaphragm). I wonder if the buckets-full of plastiflex poured on to the KEFB110 diaphragm perhaps missed the source of the problem?

 

[Lars Risbo]

It's a kind of yes and no answer to that Lars. The 2pi/2.83V sensitivity specified for the driver is 83.1dB (as you well know . I loose say, a dB, of that due to the DCR of the series crossover inductor(s) which puts me nominally at 82dB (above the baffle step). However, I think the speaker is also almost certainly going to find itself in use very close to a wall, so thinking generally in the 2pi world seems a reasonable assumption. Hence 82dB seemed reasonable to me.

The complexity here though is that in doing a modern interpretation of the LS3/5A I have to decide how close to voice the LS3/5X to the (somewhat idiosyncratic) original, and a lot of the action in that respect falls either side of the baffle step region. I'm still messing with the voicing (not least 'cause I only have a single speaker working at the moment and thing are different in stereo) but the effective sensitivity will be somewhat defined by how the voicing ends up.

makes sense! in a way it is quite inefficient to do the baffle step compensation in the passive filter. requires much copper and burns power whilst eating you amp voltage budget. could be done upstream by DSP and then adjusted to rear wall positions etc.

 

[Matt_Holland]

@Philberish what about optimising the passive crossover for directivity only and then supplying customers with a set of DSP tunings to optimise the frequency response for different preferences and room configurations. Customers can use a DSP of their choice to apply the tunings.
Perhaps just a passive pad for the tweeter if its sensitivity is way higher than the woofer.

 

[Philberish]

@Philberish what about optimising the passive crossover for directivity only and then supplying customers with a set of DSP tunings to optimise the frequency response for different preferences and room configurations. Customers can use a DSP of their choice to apply the tunings.
Perhaps just a passive pad for the tweeter if its sensitivity is way higher than the woofer.

@Matt_Holland, you and @Lars Risbo are edging towards reinventing the KEF P60 and K300 of the mid/late 1980s
. You probably don't remember it (sadly, I do) but the P60 was a smallish, passive closed-box two-way intended primarily for nearfield monitoring (although it wasn't called that back then) and the K300 was its accompanying power amp and (analogue) EQ unit. The P60 needed EQ because its passive crossover did just the driver integration. KEF realised that managing a driver's rising response and the baffle step with passive EQ just burns through midrange sensitivity (and is expensive). So they used the crossover just to knit the drivers together then used active EQ before the power amp to correct the system response.

This is effectively what you're proposing Matt with your LS3/5X DSP EQ suggestion and in theory the suggestion is a good one. Only snag is, the way most potential customers think (and the way hi-fi retail works) isn't really compatible with such complexity. I guess I could do an optional driver integration only crossover and publish some EQ coefficients (which, ironically, is pretty much how I've been working on voicing), but I think in terms of likely sales, that would constitute a niche product within an already very small niche. We can do it for your pair though Matt