B&W 686 S2 Bookshelf Speakers Review

The first successful test of B&W's Kevlar technology.

"Having visited the Bowers & Wilkins Research Center and been impressed by the depth of the company's engineering expertise, I am always puzzled by the fact that B&W speakers don't measure as well as they could." - John Atkinson, Stereophile

This quote sums up the brands issues with reviews based on objective measurements. It's almost a game they play with Stereophile seeing how bad of a measuring speaker they can produce while still getting put on the class A recommended list.

Last time we looked the KEF speakers and the british sound. In that little rant I lumped B&W into the mix as a source of the British sound without really giving much evidence to back things up other than my impressions. B&W is very different from KEF in how they do things, and while the sound they produce does contain many tonal similarities with other British brands they indeed deserve their own review to figure out how they get that sound.

Another pair of speakers from my friend and internet stalker Wes. He provided the Wave Crest HVL-1's up for the initial sacrifice, but I demanded more, so his prized B&W 686S2's are on the altar ready to burn.

Component Overview:


The tweeter is a metal dome loaded into some kind of Hannibal Lector face mask type enclosure, with a metal grill and a plastic trim ring that locks it to the plastic u-shaped bezel both of which are painted to look like brushed aluminum. From the limited viewing angles it looks like the driver has a slight horn loading inside of its little cage, and unlike other tweeter prisons this design was impervious to my attempts to remove the grill.

This rear shot of the tweeter shows the back of the unit has a sealed plastic tube that protrudes in a very strange way. I'm guessing this rear bit is for extra air space behind the dome. This would allow space to absorb the back wave created by the tiny driver while not being susceptible to the back pressure of the woofer.


The iconic yellow cone made of woven Kevlar fibers and fuzzy felt covered dust cap make this a flashy, but handsome driver. The trim ring is huge and has a nice round over, a negative side effect of this is that the woofer is shoved way back in the speaker in comparison to the tweeter, making time alignment something that would need to be done in the crossover.

Back of the driver is braced in a maze of MDF, and shows off B&W branding right next to Made in China on the sticker.


Golf ball textured port is another B&W thing that they do. I don't think it's doing much audibly, but it's a nice textured piece to contrast with the rest of the quality looking components on the front of the box.


The outside of the cabinet is a nice textured vinyl with an extremely dense plastic front baffle covering. The inside has lots of thick MDF and with plenty of polyfill and bracing. A knuckle wrap leaves no doubt that this is a stout cabinet.


The rear terminals are over done in my opinion for such a small speaker. The 686 S2 has two sets of binding posts, one for the tweeter, and one for the woofer. The metal bit in between bridges them so you only need to run one pair of wires to the speaker.


Initially laughable first order crossover, but on further examination these are high quality parts. The coil, square cap, and wire wound resistor are all B&W parts that look like they cost a pretty penny to manufacture. Obviously the topology of the crossover network maters more than the makeup of the components, so this is all a little silly.


Grills look great, very transparent and work well with the speaker. Here is the thing though, when you buy a speaker with a Lamborghini yellow cone you don't really use the grills.

Desktop Setup:

Initially I put the speakers up at ear level with the tweeter, but quickly noticed that the tweeter sounded way too forward, a side effect of the tweeter being physically forward on the baffle. Moving the speakers to ear level with the top of the woofer ended up being the best height for the speaker.

I tried these with both my old home theater receiver, as well as a cheap Behringer UCA222 DAC and SMSL SA50 amp. Both setups did the job just fine, driving the speakers past loud without distortion or tonal differences.

Listening Notes:

The drivers sound super clean, another low distortion speaker from the Brits.

I'm hearing all kinds of stuff I've never heard in my music... this is always a bad sign as it means the speakers are doing something wrong. This is usually a rising response issue, but instead it sound like a more complicated issue.

Hall and Oats I can't go for that sounds like a remix. The vocals are completely different, there is almost nothing resembling the original recording.

Imaging is more like headphones than speakers, with vocals being so far forward they are almost behind or in the head.

These speakers are an addictive poison to the ear. Like a loudness button on a receiver that ends up always pressed. I can see why people like these, but it ain't right.

Phil Collins is a disaster, there are no words for how bad this sounds. No more 80's stuff, these just slaughter that entire era.

The speaker is off so much tonally that it's actually pushing vocals up octaves, and the 10kHz range is boosted to the point of sizzle only balanced out because of the boosted bass response.

Boosted bass on the speaker means that Deadmau5 stomps with extra umph, with the low point of the speakers output sounds like it's somewhere around 50Hz.

Unlike the KEF, B&W speakers do sound different. These sound way worse than the 800 Diamonds I heard at RMAF which were at least passable as hifi speakers.

The 686 is the kind of speaker you make when you don't respect the average consumer enough to trust them to like music the way it was recorded.


The measurement is accurate from 700Hz on up, below that my room modes show up, but that data can gives us an idea of what the speaker is doing tonally.

Yet again a +-3dB speaker that is +-2dB in most of the treble range on axis. This speaker measures better on axis than many of the larger B&W speakers tested by stereophile, so what is making it sound bad? Well, +-3dB smoothed response still gives plenty of room for error, and these speakers really take advantage of all of the ways you can spice up a speaker.

Here are some of my notes drawn on top of a 1/48 smoothed on axis measurement.

Working from left to right, we have front port cancelation as an artifact of the measurement, coupled with bass/lower midrange boost. Following that we have a large hump centered around 1k which is responsible for the imaging being a mess. The rising response from 2-7k although slight is worse off axis and gives the speaker it's over detailed sound. The dip in the presence zone helps the speakers "disappear" and is a sound signature of B&W. The whole treble range is punctuated with some nasty peaky bits at 3.5k, 5k, 7.5k that combined with 6dB treble explosion happening at 10kHz really make the speaker sizzle.

First thing I checked for when getting the speaker was comb filtering between the drivers, because I heard the tell tale signs of a first order crossover with the treble being sizzles and chunks. Turns out B&W run the tweeter inverted to avoid the issue, similar to the mod I did in the B652-AIR review. To test this I inverted the polarity on the tweeter to put both drivers in positive polarity and you can see the obvious cancelation between the drivers making a big null and affecting the response all the way down into the midrange.

The drivers cross around 4k, you can see the natural roll off of the drivers helps make the crossover a bit better than what you would expect from a first order cap and coil.

This is group of measurements is from 0-50 degrees off axis in 10 degree intervals. These are windowed to avoid any reflection data from the room, which limits the results to 300Hz and above.

This is a mess, we have a big off axis null at 2k making the hump at 1k even sharper and makes a massive 6k hump off axis creating the rising response over detailed sound, and the upper octave elevation never gets quite flat, and dies in a fire 30 degrees off axis.

DSP Correction:

I did try to make one, but with the issues of the speaker they ended up sounding bad when corrected flat on axis, and then okay when corrected with off axis averaging but not great.

Final Thoughts:

Initially when hearing these speaker I thought I liked them better than the KEF Q300's and the ELAC's, and if I was chasing pure euphorics from my music I would probably have stuck with that pecking order. The way that these change the tone of vocals really is what turned me off of them, and it's just not something I can ignore or live with.

Going back to the quote from Stereophile, why don't B&W speakers measure as well as they could? I've come to realize while reviewing these that it's because B&W doesn't want them to. These speakers have a well engineered cabinet, really nice drivers, and a crossover that while technically shit is made out of nice stuff. They have the ability to make accurate speakers, but instead choose to make something different. These speakers sound exactly how B&W wants them to sound. It's not my cup of tea, but from the rabid followers out there it is obvious that a few people like it.

Check out the system finder to see what I recommend.

Other content you may like: