Thanks for visiting the CODE blog! In this piece I wanted to take a look at a speaker that’s causing a bit of a stir in the Hifi world … the KEF LS50W.
It’s scooped numerous awards, received many positive reviews and if you go to any Hifi show, many of the punters will recommend a listen.
I don’t intend to give you yet another review, but instead take a look at why it’s such a good speaker.
Who we are
If you’ve not heard of CODE, we’re a small high end Hifi company based in Woking, England. My name is Ceri Thomas and I’m the founder and chief designer:
To be clear, CODE and I have no commercial relationship with KEF. The only commonality is our company also make active, DSP speakers, like our SYSTEM-2 you see below:
If you’re not sure what an active, DSP speaker is, why not check out one of our previous blogs:
What is the KEF LS50W?
It’s essentially an all-in-one Hifi system retailing at £2k here in the UK. Below is a quick overview:
And a link to HiFi Choice’s review (It was awarded ‘Editors Choice’):
Here is a great little explainer / review video by What HiFi?:
If you want some more info, check out KEF’s website below:
The LS50W uses a 1″ aluminium dome tweeter co-axially mounted within a 5.25″ mid / bass driver:
It’s an active design featuring a 30 watt class A/B amplifier on the tweeter and 200 watt class D amplifier on the mid / bass driver. The crossover is done digitally via Digital Signal Processing (DSP) and the bass loading is ported. Frequency response is quoted as 50Hz to 28kHz (+/- 3dB), with a sensitivity of 85dB at 1m.
Not to mention it can also stream high res audio through the Tidal app, sent directly from your internet router. Oh, and it has a subwoofer output and individual DSP ‘tunes’ for different room positions:
So why is it so good? Well, KEF like one or two of the other big boys (B&W spring to mind) spend a good chunk of their revenue on R&D, and it shows!
KEF have written an excellent white paper on the speaker. Well actually for the original passive LS50, but it’s close enough:
In the rest of this article we’ll reference it heavily to give you an understanding of how the speaker was developed.
Lumped Analysis Modelling
Very early on in the design process, once you know what bass driver you want to use and the physical package constraints (internal volume), you’ll be able to do a lumped analysis model like the one below for the LS50W:
This gives you an idea of the speaker’s bass response. What’s interesting here is KEF haven’t tried to boost the upper bass as many other manufacturers do with smaller speakers, like the one below:
This kind of speaker may sound instantly impressive in the show room, but it soon becomes annoying. If you have to put a speaker like this in a corner with the associated 9dB of bass boost, the upper bass will overwhelm everything.
So not only will the LS50W be well behaved near boundaries, it also has the pre-sets to ensure the sound is balanced, for it’s exact boundary conditions!
We talk about this a little more in the blog below:
KEF put a LOT of time and effort into designing their cabinets. They’re very much of the school of thought that cabinet resonance should be eliminated (and so are we for that matter).
The graphs below show the sound output of two very different cabinets. The first one is an un-braced chipboard design, while the second is the well damped and braced LS50:
You can see the chipboard design is producing a large amount of sound around 200Hz. Initially this may sound pleasant and ‘warm’, but it can also make the sound ‘tubby’ with a sense of ‘slow bass’ and poor timing.
To achieve the acoustically dead characteristics of the KEF design, you need to implement two concepts. The first is Constrained Layer Damping (CLD):
This essentially uses damping materials between the internal elements of the speaker, which stops vibrations moving through the cabinet.
The second method is cabinet bracing. To understand this, you need to utilise something called Finite Element Analysis (FEA). I’m sure you’ve seen something like this before:
It’s essentially a computer modelling tool to allow you to understand how a force applied to an object causes it to deform. Below is an early prototype of the LS50W cabinet, with limited bracing. As you can see the cabinet is not particularly stiff and will deform.
This deformation will produce cabinet vibrations, as can be see in the example of the un-braced chipboard cabinet above.
KEF use Computation Fluid Dynamics (CFD) to model the air flow through their ports. The example below shows turbulence in a straight port (left) vs. a double flared port (right), as per the LS50W:
As you can clearly see, the double flared port produces far less turbulence and will be far less likely to ‘chuff’! Yes, you heard me correctly, and no-one enjoys a chuffing speaker!
KEF is a relatively large speaker company with products ranging from a couple of hundred pounds to over a hundred thousand pounds.
Much like car companies, it allows them to develop new technology on their more expensive models, which can then trickle down over time.
One example of this is the tweeter on the LS50W, which was originally designed for their beautiful blade model, which costs in the region of £20k:
This is quite a special unit. Older, less well designed metal dome tweeters, had a nasty resonance spike that went off before 20kHz. This is in the audible range and quite unpleasant.
The below example is actually a pretty good Seas unit, but it does show the resonance behaviour of a metal dome.
*Most drive unit companies fudge their frequency plots, but to Seas’ credit, they’re very honest, which is why I often use them as an example.
The KEF LS50W manages to push this resonant spike out to 40kHz and even then, because the response can be equalised with DSP, the spike can be almost completely eliminated.
Not only that, but the tangerine waveguide seen below will increase sensitivity and aid dispursion.
Coaxial drive units
KEF are famous for their coaxial drive units (shown below), also favoured by Tannoy and now TAD. To my mind, there is just something inherently ‘right’ about the sound of a co-axial.
This could be due to the sound from the tweeter and mid / bass driver emanating from the same point (a point source). As such there is less phasing issues in the crucial 1-5kHz range where the tweeter crosses over to the mid / bass driver.
This can be seen by the well controlled polar plots:
These polar plots are likely further improved on the LS50W due to time alignment of the drive units using the DSP processor, which is able to ‘hold’ the mid / bass signal, ever so slightly, to allow the sound from the tweeter to catch up.
The graph below is lifted directly from KEF’s website, but it shows the principle:
A square wave is notoriously difficult for a speaker to reproduce. However, due to the LS50W’s coaxial point source arrangement and the time alignment, it’s phase behaviour is very good, and it has a decent stab at reproducing the input signal.
The LS50W retails at £2k. If we remove VAT (that goes direct to HMRC) you’re left with £1,666. Now I don’t know what mark-up KEF is offering dealers, but assuming it’s the standard 50 / 50 split (It may be less), KEF would get £833.
That’s a heck of a lot of speaker that KEF is giving you for the price; considering it’s got to cover the Bill of Material (BOM) costs, R&D, transportation, wages, infrastructure, advertising etc.
I suspect it might even be a break-even product and that is incredibly rare in the world of high end Hifi.
In summary, the reason the LS50W sounds so good is:
- It’s aggressively priced.
- It’s designed in the UK, yet manufactured in China.
- Vertical integration is used to trickle down tech from more expensive models.
- KEF invest heavily in R&D.
And finally; it’s just a well engineered speaker, competently utilising the very latest technologies.
I take my hat off to KEF, they’ve done a fantastic job!
Thanks for reading, and as usual, if you want to get in touch:
Founder & Chief Designer