Paul Marchant
pfm Member
Size of that tweeter.
That's what all the girls say...
Kef KM1 in need of repair.
- Thread starter James46
- Start date
That's what all the girls say...
Blue Belle
pfm Member
Paul Marchant
pfm Member
Right - I'm back from having my tweeters examined (ahem), and it seems that they're in good health. The recommendation was to not do anything to them at this stage. They can go back into their enclosures on Monday when I'm next at work.
At some point in the coming week, probably Friday, I'll pull one or both of the remaining midrange drivers apart and do a clean and re-ferrofluid and then that's more or less it.
I need to flip all the LF drivers 180 degrees to deal with sag in the surrounds over time.
After that, there's the not-inconsiderable challenge of getting all of the bits in one place with enough space to put them together. And then making lots of noise...
At some point in the coming week, probably Friday, I'll pull one or both of the remaining midrange drivers apart and do a clean and re-ferrofluid and then that's more or less it.
I need to flip all the LF drivers 180 degrees to deal with sag in the surrounds over time.
After that, there's the not-inconsiderable challenge of getting all of the bits in one place with enough space to put them together. And then making lots of noise...
lawrence001
Basic Member
I thought you had showed a huge commitment to this Kef project already, but to sell your house to have a room to use them in really is going the extra mile.
Paul Marchant
pfm Member
The house move was planned before I got the KM1s, but I will admit that every time I look at houses for sale, one of the criteria now is 'appropriate space for KM1 usage in living room'.
Rug Doc
pfm Member
I hadn’t seen this thread before.. always nice to have as much info as you can!
KEF KM1's - Rarest of them all? Probably!
I use to think the Maidstones were some of the rarest KEF's made until I started to look into the KM1's... custom made for the BBC and if there are even 50 I'd be surprised! I have never seen a pair for sale anywhere in the world in all the years I've been doing the D&S thing!!!! Well here's...
audiokarma.org
Last edited:
lawrence001
Basic Member
You into old rugs then?
Edit : stupid question looking at your forum handle.
Edit 2 : if that is your shop could I ask you a question about a rug I once owned?
Last edited:
Rug Doc
pfm Member
Yes go ahead, any questions welcome.
Apolgies though, the email address wasn’t meant to be there - it’s an auto paste on my phones behalf to make writing my email address easier! I have removed it from my post, could you edit it out of your quote too please?! (Or mods please!) Thanks.
lawrence001
Basic Member
Email removed
And thanks for offering to help with a question it's something I've been wondering for years and tried to get advice at the time before selling but the auction houses I tried weren't interested.
I need to go back on my back up storage to find a photo first so it won't be soon
Paul Marchant
pfm Member
There's some inaccuracy in the post. As far as I know, there were 72 production pairs of KM1s made. Of those, three pairs were KM1Ps, the ported variant which was unique to BBC Maida Vale St. 5 and St. 6 (black ash veneer), and a pair in light ash for the Golders Green TV Theatre.
05KM1Pblk-wht.jpg
drive.google.com
The KM1Ps were converted from standard KM1s at the KEF factory. The trim around the front panel ports is the port from a Reference 104/2.
They were done because the way the BBC sound guys were setting up microphones for drummers would involve a guy sitting there repeatedly rapidly hitting the kick-drum pedal for protracted periods. With the original design, at very high levels the cones would 'walk' and not return to their centre position due to asymmetry of the magnetic field away from the centre of the magnet gap. The infinite baffle cabinets were sufficiently well sealed that once this had happened, the air pressure in the cab helped keep the cones away from re-centring. The solution was to port the cabs (thus providing a very large air leak) and tuning the port resonance to the main frequency of a kick drum, thus greatly reducing cone movement under those circumstances.
The light ash pair of KM1Ps now belong to Pete Thomas (owner of PMC Loudspeakers) and are visible on PMC's website in their virtual museum.
One of the two black pairs is at the London Science Museum : https://collection.sciencemuseumgroup.org.uk/objects/co8355778/kef-km1-loudspeaker-loudspeaker
I've got the other black pair. Serial nos. 151 and 152.
I t-h-i-n-k the production pairs started from serial number 101.I'm aware of a pair of KM1s at Cork Community College, in a slightly sad condition but I've been talking to one of the lecturers there and he has plans to restore / repair them. Other than that, I don't know the confirmed whereabouts of any others.
Last edited:
Rug Doc
pfm Member
Amazing information, thanks Paul.
Found PMCs ones on their site - they look superb in that finish.
I love these speakers, but the only thing that I don’t think is very room friendly is the shape of them, and their stands.. they are obviously top heavy and I’d love to design a stand that better suits their shape.
I believe your second link is Incorrect.. pls check.
Found PMCs ones on their site - they look superb in that finish.
I love these speakers, but the only thing that I don’t think is very room friendly is the shape of them, and their stands.. they are obviously top heavy and I’d love to design a stand that better suits their shape.
I believe your second link is Incorrect.. pls check.
Paul Marchant
pfm Member
Indeed it was. I've corrected it.
Paul Marchant
pfm Member
The tweeters have been deemed healthy and I'm taking the advice to leave them well alone.
The usual Friday update has been delayed as I had to attend the work Christmas bash on Friday evening. However, I had a bit of free time this evening and have done a bit more.
There were two more midrange drivers to take apart, and some curiosity to satisfy.
So, I have (had) two MF drivers with new, runny, ferrofluid, and two with syrup-consistency ferrofluid (and some ferrosolid as I've taken to naming it) in the magnet gap.
When you manipulate the drivers by hand, there's an enormous difference in their freedom of movement. I previously mis-described it as compliance, but I think viscous drag might be a more accurate term.
I was expecting a significant difference in sensitivity between the drivers with the dried-up ferrofluid and the drivers with the new stuff.
In the name of science, I lashed-up a test setup:
There's a biro line drawn on my bench to facilitate consistent location of the two drivers. I've got a little (nasty but adequate for this) power amp, a test-tone generator and an audio analyser with a mic input.
I assembled the above, as per the picture and did test measurements at 1kHz of both drivers.
I was expecting a difference in measured output in the 6 - 10dB region.
It was less than 1dB.
I didn't have time to setup my big-boy test set and do things like impulse response or distortion measurement (nor is that really the microphone to be doing that sort of thing with, and my decent measurement mic was at home).
I'm staggered by the lack of difference in output.
I understand that, above resonance, the driver's motion is not controlled by the elasticity of the suspension, but really can't see how the frictional losses from the ferrofluid don't seem to have any tangible effect. Another moment where my lack of knowledge of loudspeaker theory is apparent...
Anyway, moving on, I've disassembled the last two midrange drivers, apparently without incident, undue stress or damage. The cones are firmly affixed to the voice-coil formers and my attempts to provoke the glued bond into failure got nowhere, so I'm mostly set on leaving that alone for now.
I'm wondering if there's any merit, whilst the drivers are apart, in adding a tiny layer of cyanoacrylate to the inside and outside of the cone / coil former bond as a reinforcement.
These two, re-coned some years later than the first pair, took more solvent and time to get the rear suspension to de-bond, but it appears to have done so without damage.
I ran out of time to clean the ferrosolid off the voice coils and the ferrofluid out of the gap, but that's a job for later in the week.
Once this is done and new ferrofluid applied, I'm almost there.
I need to fit the dust-caps. I'm unsure what adhesive to use for this. The immediately available option is the heptane / cyanoacrylate stuff which has proven successful in my previous gluing. The stuff that was there originally has a rubbery consistency. I wonder, does it matter from a damping or otherwise perspective what adhesive is used? The old stuff cleaned off with MEK quite well.
I've assembled a crude alignment jig for getting the dust-caps central, in the form of a laser cross-hair generator which is currently magneted to the ceiling. I need it a bit lower to get a sharper and brighter laser cross but that's just a case of borrowing a microphone stand from the rentals department two rooms away from my workshop. I'll deal with this in due course.
The usual Friday update has been delayed as I had to attend the work Christmas bash on Friday evening. However, I had a bit of free time this evening and have done a bit more.
There were two more midrange drivers to take apart, and some curiosity to satisfy.
So, I have (had) two MF drivers with new, runny, ferrofluid, and two with syrup-consistency ferrofluid (and some ferrosolid as I've taken to naming it) in the magnet gap.
When you manipulate the drivers by hand, there's an enormous difference in their freedom of movement. I previously mis-described it as compliance, but I think viscous drag might be a more accurate term.
I was expecting a significant difference in sensitivity between the drivers with the dried-up ferrofluid and the drivers with the new stuff.
In the name of science, I lashed-up a test setup:
There's a biro line drawn on my bench to facilitate consistent location of the two drivers. I've got a little (nasty but adequate for this) power amp, a test-tone generator and an audio analyser with a mic input.
I assembled the above, as per the picture and did test measurements at 1kHz of both drivers.
I was expecting a difference in measured output in the 6 - 10dB region.
It was less than 1dB.
I didn't have time to setup my big-boy test set and do things like impulse response or distortion measurement (nor is that really the microphone to be doing that sort of thing with, and my decent measurement mic was at home).
I'm staggered by the lack of difference in output.
I understand that, above resonance, the driver's motion is not controlled by the elasticity of the suspension, but really can't see how the frictional losses from the ferrofluid don't seem to have any tangible effect. Another moment where my lack of knowledge of loudspeaker theory is apparent...
Anyway, moving on, I've disassembled the last two midrange drivers, apparently without incident, undue stress or damage. The cones are firmly affixed to the voice-coil formers and my attempts to provoke the glued bond into failure got nowhere, so I'm mostly set on leaving that alone for now.
I'm wondering if there's any merit, whilst the drivers are apart, in adding a tiny layer of cyanoacrylate to the inside and outside of the cone / coil former bond as a reinforcement.
These two, re-coned some years later than the first pair, took more solvent and time to get the rear suspension to de-bond, but it appears to have done so without damage.
I ran out of time to clean the ferrosolid off the voice coils and the ferrofluid out of the gap, but that's a job for later in the week.
Once this is done and new ferrofluid applied, I'm almost there.
I need to fit the dust-caps. I'm unsure what adhesive to use for this. The immediately available option is the heptane / cyanoacrylate stuff which has proven successful in my previous gluing. The stuff that was there originally has a rubbery consistency. I wonder, does it matter from a damping or otherwise perspective what adhesive is used? The old stuff cleaned off with MEK quite well.
I've assembled a crude alignment jig for getting the dust-caps central, in the form of a laser cross-hair generator which is currently magneted to the ceiling. I need it a bit lower to get a sharper and brighter laser cross but that's just a case of borrowing a microphone stand from the rentals department two rooms away from my workshop. I'll deal with this in due course.
Last edited:
Paul Marchant
pfm Member
And now the last two midranges are on my bench for reassembly.
The ferrofluid has been cleaned out of the magnet gap.
The ferrosolid has been cleaned off the voice-coil.
I've killed at least 17% of my remaining brain-cells with the fumes from the solvent.
and now they're both shimmed and glued. Glue is drying as I type. I really hope I've got them right.
The coil former looks central in the gap, the shims are a fairly close fit. What could possibly go wrong....? I'll find out tomorrow.
The ferrofluid has been cleaned out of the magnet gap.
The ferrosolid has been cleaned off the voice-coil.
I've killed at least 17% of my remaining brain-cells with the fumes from the solvent.
and now they're both shimmed and glued. Glue is drying as I type. I really hope I've got them right.
The coil former looks central in the gap, the shims are a fairly close fit. What could possibly go wrong....? I'll find out tomorrow.
Paul Marchant
pfm Member
Tomorrow is now today. It was a tense hour and a half drive to get to work, waiting to discover whether or not I'd been successful.
Everything is good.
There's no coil rub on either of the mids. The ferrofluid is in. I've put test-tone through both and there's no nasty noises. My luck has held.
All that remains with this part of the task is to glue the dust-caps on and reinstall the drivers in their enclosures.
Last edited:
Paul Marchant
pfm Member
Dust-cap time....
I'd underestimated how difficult it is to stick a round thing centrally in a tapered round thing. Particularly when it's cyanoacrylate / heptane adhesive where you get exactly one go.
I have a vacuum tweezer tool, for picking up and placing small electronics parts. It's nothing special, just a vac pump, a rubber hose and a hand-piece (which is a piece of tube with a pointy end and a small rubber sucker cup). Seemed ideal for the job, but for the fact that the dust-caps are an open-pored fabric, so won't stick to the tool.
A small piece of cheap sticky tape (with very anaemic glue) on the dust-caps solved that problem. I needed to coat the dust-caps to make them airtight, but that had to happen after they're glued so as to not affect the glue's bond.
In an attempt to make life a little easier, I purchased a cheap laser-crosshair module, to give me a centre point on the cone.
I had many, many rehearsal attempts to place the dust-caps before getting the glue out.
And after an hour of very stressful poking and gluing, and painting on the sealant onto the dust-caps once the adhesive had set, it looks like everything's worked out as I'd hoped for the last five months. It's been a lot of effort, a lot of stress, an unbelievable amount of money on ferrofluid and umpteen different types of glue and solvents. It would 100% not have happened but for the assistance I received from the KEF engineer who designed the KM1's.
It's all good:
I'd underestimated how difficult it is to stick a round thing centrally in a tapered round thing. Particularly when it's cyanoacrylate / heptane adhesive where you get exactly one go.
I have a vacuum tweezer tool, for picking up and placing small electronics parts. It's nothing special, just a vac pump, a rubber hose and a hand-piece (which is a piece of tube with a pointy end and a small rubber sucker cup). Seemed ideal for the job, but for the fact that the dust-caps are an open-pored fabric, so won't stick to the tool.
A small piece of cheap sticky tape (with very anaemic glue) on the dust-caps solved that problem. I needed to coat the dust-caps to make them airtight, but that had to happen after they're glued so as to not affect the glue's bond.
In an attempt to make life a little easier, I purchased a cheap laser-crosshair module, to give me a centre point on the cone.
I had many, many rehearsal attempts to place the dust-caps before getting the glue out.
And after an hour of very stressful poking and gluing, and painting on the sealant onto the dust-caps once the adhesive had set, it looks like everything's worked out as I'd hoped for the last five months. It's been a lot of effort, a lot of stress, an unbelievable amount of money on ferrofluid and umpteen different types of glue and solvents. It would 100% not have happened but for the assistance I received from the KEF engineer who designed the KM1's.
It's all good:
lawrence001
Basic Member
When do they all go back in the cabs and you get to hear them in all their glory?
Remind me, are the crossovers ok?
Remind me, are the crossovers ok?
Paul Marchant
pfm Member
That's a difficult one.
Remaining tasks to do:
Replace all connectors (on the wires to the speaker enclosures, and the chassis mounted sockets on the amps). This isn't as bad as it sounds, because it's just the plastic housings that need replacing. You can readily de-pin the old housings (broken brittle plastic) and swap the pins / receptacles into new housings. I have the parts to do this, and a loan de-pinning tool. Trivial difficulty, just tedious.
Rotate all bass drivers 180 degrees. This is more hassle than it sounds because KEF B300's have a triangular (tri-lobal) chassis, and there aren't holes in the right place. However, on the KM1's, the upper and lower bass drivers are mounted mirror-image. What I need to do is desolder the wires, swap the upper and lower bass drivers around. Times four. Tedious, but low risk and difficulty.
It's necessitated due to sag in the surrounds over the years. I've got one driver that's just kissing the magnet pole-piece with its voice-coil former. I'm 100% certain that rotating it will cure this, probably instantly. The difficulty there is doing it whilst lying on the concrete floor of the storage place, with a battery soldering iron and little light to work by. Realistically, I just need to man-up and get on with that.
Affix two replacement dust-caps to the two bass drivers that lack them. Trivial - just a clean and glue. I have the parts.
Get everything into one place, with mains power, a suitably strong table and reassemble the five modules (amp / crossover, MF/ HF enclosure, central chassis, pair of LF enclosures) times two. This is actually the most difficult part of the project at present.
The central chassis and LF enclosures for both speakers are at the Big Yellow Storage place round the corner from work. I have a silly little car which might or might not fit an LF enclosure (only one at a time). My house is a fifty mile drive from work / Big Yellow. So far, my boss and colleagues have been remarkably tolerant of the bits of KM1s cluttering up the place and my occasional late night engineering missions. I'm a keyholder at work, and realistically, at a weekend I could get away with dragging everything into one room and making as much noise as I like. I plan to do this, hopefully relatively soon. Logistics and manpower are my main problem. There's nothing small or light about any part of a KM1. I need, realistically, someone to help, and use of a larger vehicle. This clearly isn't impossible. There's no real way of getting everything back to my current (small, mid-terrace) house, nor is there anywhere I can do high-level sound-checks. My house is on the market, and my plan is to move over to where work is and - specifically - buy a detached house with a suitably sized living room (because 'music').
The electronics, I believe, is now fixed. I had a faulty amp card on one amplifier pack (but I'm a PCB repair engineer for a living, so that got fixed relatively quickly and easily).
I've done a quick listen test on all of the amplifier channels and had a quick look with an oscilloscope and it all provisionally looks good. I own an Audio Precision ATS-1 and a Prism Dscope IIIE test set (about £15k in total when they were new) and I should probably build a dummy load and graph frequency response and THD for all the amplifier channels. I'm not expecting any problems there, to be honest. I might let that slide, and just do a proper listening evaluation. I know there's no major faults and no DC-on-output since repairing the faulty amp card. Upon arrival with me back in the summer, originally, one of the amplifier packs had the voltage selector half-way between the 240v and 120v setting. I t-h-i-n-k this is why the Ebay seller had no operation from one of the amp-packs. Without giving it any thought, I checked both and adjusted the miss-set one to 240v and it powered straight up. I'm not worried about the crossover / amp cards. If there's a fault, it's the sort of thing that I'm proficient at fixing. There's no components there that I can't efficiently test or source replacements for.
I need to get the mids and tweeters back in their enclosures and get on with it. Hopefully at some point in January, I'll get the last bits done and make them sing once again.
Last edited: