KEF 104.2 Restoration Project

October 14, 2018

A friend approached me a little while ago, and asked for help restoring his Kef Reference 104.2 speakers. The little foam donuts had deteriorated, and the overall sound wasn't up to what they should've been. In their day, they were a feat of loudspeaker engineering, and a well-looked-after pair can still keep up with modern speakers.

 

Running some measurements revealed unusually high distortion levels and fairly poor integration between the drive units. Clearly, something was amiss.

 

Doing some research, the following potential problems turned up:

- Deteriorated foam on the bass drivers

- Tweeter ferrofluid drying up

- Electrolytic crossover capacitors decreasing in value with age

 

So, there's something to work with there.

 

A quick explanation of each part, for those following along:

The foam on the bass drivers is needed to help keep the cone aligned as it moves, as well as provide an air-tight seal around the edge of the cone. This particular model had rubber surrounds on the cones, so it was only the small donut in the middle that had gone.

Ferrofluid is a fairly viscous liquid that's magnetic. It sits in the magnetic gap, in the same space as the voice coil, which ends up effectively surrounded by this liquid. This fulfills a couple of functions. First, as power levels rise the ferrofluid really helps the voice coil to conduct heat away into the magnet, which has much more thermal mass. Secondly, ferrofluid also provides some mechanical damping, which is most effective towards the bottom of the tweeter's range. The result there is that ferrofluid helps to give a very smooth and flat impedance curve, which the crossover is designed for. No ferrofluid = misaligned crossover.

Electrolytic capacitors (as used in these speakers) are normal capacitors, but filled with goo (electrolyte), to put it simply. The goo increases the value of the capacitor by a large amount, which means you can get very high values in fairly small parts. Using other technologies, some of the capacitors would have been the size of tin cans! The down-side is that the goo will dry up over time, and the only solution is replacement.

 

 

Fixing the bass drivers was a fairly straight-forward affair: remove any residual foam, and stick the replacement parts down.

Looking around online, there are lots of people that are struggling with the same problem - how to keep the cones aligned, while force is being applied as everything dries in place. I've seen photos of pencils and elastic bands etc etc, which might work, but they're incredibly fiddly. Plus, you're still at the mercy of how consistent the elastic bands are.

 

So, here's how to do it:

- Apply glue to both edges of the replacement foam donut

- Position on the cone and central plastic column

- Place a small weight on top of the central column, holding the foam in position. I used a socket from a ratchet set.

- Get an AA battery and connect it to the speaker terminals*, + to +, - to -. A TV remote makes a convenient battery holder.

 

* Yes, I know, DC kills speakers etc etc etc etc. No it doesn't. Too much power kills speakers. An AA battery sits at 1.5V. Into an 8ohm speaker, that's 0.28W. The voice coil is perfectly capable of dissipating that power indefinitely with no ill effects. Better still, the cone is being propelled by the voice coil, so it's moving straight up and down, not being pulled to the sides by elastic bands etc.

 

 

Carefully dismantling the tweeters (not for the faint of heart) revealed the ferrofluid had definitely dried up, leaving a residue on the magnet and voice coils. Of course, that had to be cleaned up. Using cotton buds for the voice coils, and a combination of business cards and kitchen roll, the residue was eventually removed. It's possible to help the process with cleaning products, but be very careful here - the wrong thing could dissolve the glue that's holding everything together.

When it comes to re-filling the ferrofluid, you don't need much. All that's required is to fill the magnetic gap, which is the top millimetre or two as you're looking at the open tweeter. So, go slowly with the syringe. Put a few drops in, and let it run round the gap a bit. Put a few more in. If it's not going all the way around of it's own volition, apply some ferrofluid at the opposite end of the circle. Keep going and filling the gaps, and then stop. There's enough ferrofluid when there aren't any spaces - it's just one continuous circle. Putting more in after that will result in leakages when the tweeter's diaphragm is replaced.

 

If you're still reading this, I'll assume you know how to replace capacitors. Take your time, and double-check values before soldering them in, and then check again before you cut the leads down to length. Having a photo of the crossovers before you start work is useful as a reference.

 

With all the work completed and the glue dried, the measurements were much improved. They show a slightly-declining response towards the top end, which gives a relaxed presentation. Distortion levels are nice and low, and driver integration is seamless. They can go to high levels without complaints, too.

 

If you have a set of these speakers and are thinking of working on them, feel free to get in touch with any questions or comments about this guide. If you'd rather leave this sort of work to the professionals, I'd be happy to come to you to work on them for a day. Please do get in touch.

 

Good luck!

Please reload

Recent Posts

November 3, 2019

September 28, 2019

Please reload

Categories