Hello all,
As the proud owner of a pair of 104/2s for many years, having never felt the need to upgrade, I was very disappointed to have been struck down by fate earlier and suffered one's failure.
Let me explain the problem: Playing music very loud although with no audible distortion, Roland 909 snare very loud in one track and suddenly one side loses both mid and treble. Bass drivers continue to work but the B110s and T33 don't. Taken midrange enclosure off, looked inside to see any obvious crossover component failure and can't see any. Played music through speakers and when a certain volume is reached, there's a relay-type device inside on the rear-mounted crossover that starts clicking away like crazy and is very audible. Now very distorted signal plays through the midrange and treble units, and relay device still clicks.
I was not aware that relays were used in crossovers, it says '12v coil' on top but as it clicks and is in a plastic square case, I'm confident it must be a relay? Is this some sort of protection relay that has failed, causing distortion? I was surprised a crossover component has failed before the drivers, but on feeling the amount of heat outputted from the crossover components, I'm now not that confused.
Has anyone else suffered similar failure on their 104/2s or other speakers? Is this actually a relay or am I confused? They're very early examples of the speaker with the black port surround, wooden midrange enclosure, crossovers mounted in the top and no easy terminals for the midrange enclosure removal and so I'm also worried that a schematic for mine will be impossible to find. Any advice appreciated!
As the proud owner of a pair of 104/2s for many years, having never felt the need to upgrade, I was very disappointed to have been struck down by fate earlier and suffered one's failure.
Let me explain the problem: Playing music very loud although with no audible distortion, Roland 909 snare very loud in one track and suddenly one side loses both mid and treble. Bass drivers continue to work but the B110s and T33 don't. Taken midrange enclosure off, looked inside to see any obvious crossover component failure and can't see any. Played music through speakers and when a certain volume is reached, there's a relay-type device inside on the rear-mounted crossover that starts clicking away like crazy and is very audible. Now very distorted signal plays through the midrange and treble units, and relay device still clicks.
I was not aware that relays were used in crossovers, it says '12v coil' on top but as it clicks and is in a plastic square case, I'm confident it must be a relay? Is this some sort of protection relay that has failed, causing distortion? I was surprised a crossover component has failed before the drivers, but on feeling the amount of heat outputted from the crossover components, I'm now not that confused.
Has anyone else suffered similar failure on their 104/2s or other speakers? Is this actually a relay or am I confused? They're very early examples of the speaker with the black port surround, wooden midrange enclosure, crossovers mounted in the top and no easy terminals for the midrange enclosure removal and so I'm also worried that a schematic for mine will be impossible to find. Any advice appreciated!
I've not seen any relays in KEF crossovers. Try having a search here, SpeakerTalk :: Index great forum for all things KEF.
They had the "S-STOP" things in a bunch of speakers, but those were labeled. Tried a quick google, found a photo in this thread, which explains nothing but you could compare to your unidentified thing.
Sounds like you have the same crossovers that I have. While I had no problems with the crossovers I did find a few failed capacitors when I disassemble them. Note the corrosion from the electrolyte on the foil side of the boards...
I will have to check the date of my 104.2s but there is a specific range of serial numbers that have this setup.
KC
I will have to check the date of my 104.2s but there is a specific range of serial numbers that have this setup.
KC
Right, I'm going to remove the low-pass filter / bass crossover board now and see what I need to pick up to replace parts. Should I be looking at replacing that blue relay (as seen in first picture, which is the one that's clicking) also, or just recapping the board?
Any suggestions as to what brand or specific caps I should be using?
Any suggestions as to what brand or specific caps I should be using?
Looks like the same cap (or two) has failed on mine! Is this likely the cause of the sudden loss of sound?
Furthermore does anyone know the value of the caps?
Furthermore does anyone know the value of the caps?
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
Last edited:
Relay protection of high-end speakers is common enough, especially in old monitors designed to deal with fast-forward tape winds which had extreme high frequency energy. B&W employed it too. It's easier to reset than replacing a fuse in front of the crossover. 🙂
How it SHOULD work is if you overdrive your tweeter (and possibly midrange) section, the relay cuts out the top section. Bass being fairly indestructable. Then when you reduce volume, the top section should start working again.
These sort of circuits draw the power to work by filtering off a bit of the high end usually through a resistor. They then rectify it through a diode bridge and store the power on a leaking capacitor connected to the energising coil of the relay.
There is a certain amount of skill and judgement in setting the trip point. It's at a certain high level of sound output.
So is it resetting satisfactorally?
Or is there now a distortion problem?
Those three wet electrolytic capacitors I see may be past their best days, and need replacing. 10-20 years is the likely lifetime of electrolytics. It's also possible that the circuit board is corroding due to leaking capacitor electrolyte.
These speakers are 30 years old, which is a long time in reliability terms. We could be looking at corrosion of soldering and circuit tracks too. It's possible the failures might be in diodes or resistors or the relay too.
But I would start by replacing the relay electrolytic capacitors first, which should be marked for value and voltage. This might be quite a tricky soldering job, since fine tips don't work well with modern lead-free high temperature solder.
You could, of course, just bypass the top-end relay protection altogether. But then you might be looking at expensive and unobtainium tweeter replacement. And to end on a gloomy note, you might have fried the tweeters altogether, despite KEF's best efforts to save you from Saturday Night Party levels of sound. We've all been there. 😀
How it SHOULD work is if you overdrive your tweeter (and possibly midrange) section, the relay cuts out the top section. Bass being fairly indestructable. Then when you reduce volume, the top section should start working again.
These sort of circuits draw the power to work by filtering off a bit of the high end usually through a resistor. They then rectify it through a diode bridge and store the power on a leaking capacitor connected to the energising coil of the relay.
There is a certain amount of skill and judgement in setting the trip point. It's at a certain high level of sound output.
So is it resetting satisfactorally?
Or is there now a distortion problem?
Those three wet electrolytic capacitors I see may be past their best days, and need replacing. 10-20 years is the likely lifetime of electrolytics. It's also possible that the circuit board is corroding due to leaking capacitor electrolyte.
These speakers are 30 years old, which is a long time in reliability terms. We could be looking at corrosion of soldering and circuit tracks too. It's possible the failures might be in diodes or resistors or the relay too.
But I would start by replacing the relay electrolytic capacitors first, which should be marked for value and voltage. This might be quite a tricky soldering job, since fine tips don't work well with modern lead-free high temperature solder.
You could, of course, just bypass the top-end relay protection altogether. But then you might be looking at expensive and unobtainium tweeter replacement. And to end on a gloomy note, you might have fried the tweeters altogether, despite KEF's best efforts to save you from Saturday Night Party levels of sound. We've all been there. 😀
Last edited:
I repaired a friends 104.2,s some years back. He had fried both tweeters in his pair despite the protection circuit. I was able to buy two replacement T33,s from the KEF agent in Aus at the time and the result was very satisfactory until he had another party and did the job properly this time. Both tweeters again open circuit and several exploded electrolytics and burnt up resistors in the crossovers. My point is the protection circuit doesn't seem to stop damage to the tweeters. In your case I would definitely replace the caps associated with the protection circuit and clean up and repair any damage to the pcb. The leakage from electrolytics is very corrosive and will eat the copper completely away if left. If you are not confident about soldering then find someone who is as once the pcb tracks are damaged it's a much bigger repair job. Use lead solder as this is what was used when those speakers were made. You should still be able to obtain some. Much easier to work with. I would also make sure you identify which cap was the culprit of the leakage and replace it. If it was one of the ones associated with the protection circuit then replace it with a 105 degree low leakage type. If it's one of the electrolytic crossover filter caps then it should be replaced with a good quality bipolar electrolytic type specifically designed for crossover use i.e. Bennic or similar. Replacing those electrolytics with film caps will alter the crossovers characteristics and KEF went to a lot of trouble in designing them around those components. Nice speakers when they are in good nick.
Wow, what a superb set of responses! Hurley appreciated.
After close examination of the board I found that the leaking cap's electrolyte had indeed (as said above!) corroded part of the circuit leaving a break. A bit of solder later (yes I should've taken the lead free advice as my Maplin's special lead-free solder didn't want to play ball) and my mids and tweeter have sprung back into action. All good. 🙂
Of course, the leaking caps point to the fact I do need to recap the crossovers. My only concern being the supposed very low tolerances used by KEF in their original construction and that by replacing them with higher tolerance components, I'll be changing the crossover points slightly. Apparently even the inductors were changed from speaker to speaker to ensure everything was matched as close as possible. Should I proceed or simply replace the two leaking caps per side and leave the rest as is?
After close examination of the board I found that the leaking cap's electrolyte had indeed (as said above!) corroded part of the circuit leaving a break. A bit of solder later (yes I should've taken the lead free advice as my Maplin's special lead-free solder didn't want to play ball) and my mids and tweeter have sprung back into action. All good. 🙂
Of course, the leaking caps point to the fact I do need to recap the crossovers. My only concern being the supposed very low tolerances used by KEF in their original construction and that by replacing them with higher tolerance components, I'll be changing the crossover points slightly. Apparently even the inductors were changed from speaker to speaker to ensure everything was matched as close as possible. Should I proceed or simply replace the two leaking caps per side and leave the rest as is?
Hi there,
Recap both sets of crossover in entirety.
I have restored many Kefs from this era - Kef 104/2s, Kef 103/3s and Kef 107s. In all of them I recap the crossovers as a measure to maintain the factory sound.
First of all you are correct, they component matched the crossovers to the speakers. However, do you think this component matching is maintained after 30 some years? It isn't. The capacitors are now degraded and are not performing near where they used to. In fact, in their current state, the whole matching concept between the speakers is likely completely gone now and they will measure more than a couple of dB from the Reference model as Kef did in the factory.
I was apprehensive myself of recapping the crossovers in these speakers because they put so much effort into them. But based on what I said above, I decided to go for it. I'm glad I did - the recap makes a substantial improvement. By the way, I am not talking about recapping with higher grade or polypropylene capacitors, I am speaking of recapping with Alcaps (which are still available, albeit in limited quantities now). Go to falconacoustics.co.uk and give Jerry Bloomfield a shout. Write down all the values of the capacitors you have, and make sure you note which ones are marked LL (which stands for low loss) as those ones will need to be replaced with a low loss equivalent. Ensure you let him know you are replacing capacitors in a pair of Kef reference speakers because he will hand match a pair for you. I recommend going with Alcaps because they are exactly the same as what was in there before and will maintain the voicing of the crossover so they will restore their factory sound. No, they won't be exactly component matched like at the factory, but they will be a step above where they are now and will be very, very close to their original factory sound.
If you need any pointers, please let me know. By the way, make sure you also replace the ferro fluid in the tweeters as that dries out and lowers the overall SPL of the tweeter which messes with the balance of the sound. And finally, check on the internal B200 drivers to make sure the foam is not rotting. If it is replace it as it will restore the bass response.
How can be the original values found again as the two speakers are matched very closely in relation to the self and the reference pair in factory ? Does the impedance of such capacitors move a lot or just its ohmic values (Q factor ?) ?
Maybe with MKT at 2% to be near from the electrolytic sound ?
Are the Alcaps maid today trully the same ? Or maybe off the shelf stocked caps (dry as well at Falcon ?)
Be prudent with foam kit : I ried to refoam one month ago a old speaker... foam size were not exactly the same (Dutch store on Ebay !) 😡 !
Maybe with MKT at 2% to be near from the electrolytic sound ?
Are the Alcaps maid today trully the same ? Or maybe off the shelf stocked caps (dry as well at Falcon ?)
Be prudent with foam kit : I ried to refoam one month ago a old speaker... foam size were not exactly the same (Dutch store on Ebay !) 😡 !
Every capacitor made varies slightly. Kef at the factory would create crossovers, and then make measurements of them and they would hand match pairs together to be within 1-2 dB of each other to obtain a freqnecy response and other measurements within a certain specification of the Reference model. Rogers did the same thing with the LS3/5a. It's something usually found in high end loudspeakers.
Do not recap the crossovers with anything other than electrolytics, unless you are prepared to spend a bunch of money and potentially end up with something that does not sound good. The crossovers were voiced and designed with electrolytics in mind - thus, replacing them with electrolytics is the smartest (and most cost-effective) method to go with rather than attempting to "upgrade" or "modify" them. Ask yourself - did the speakers sound good 30 years ago? Do they really require upgrading? I say this because the ESR component in capacitors (equivalent series resistance) is different amongst different types of capacitors and changing an electrolytic to a polypropylene will significantly alter the balance of the speaker because the ESR is very different. Thus, you will end up with a different and potentially worse sounding speaker.
Yes, Alcaps today are the same as they were then. Same factory, formula, etc.
As for the foam kit many foam kits are available for the 104/2. Just make sure you buy it from a reputable source and you will be fine.
Bit of a big job, replacing the capacitors, but Falcon Acoustics do some pretty good and cheap replacements:
Alcap 50V Standard Electrolytic capacitors for all audio and hi-fi loudspeaker crossover applications.
I really don't know what is going to work best. 50V or 100V. Low-loss or standard.
Nothing wrong with lead-free solder except it needs a very hot iron. I use a 40W jobbie with a big tip and get in and out in 5 seconds or so. Don't forget to tin the new iron tip first time it warms up too. Not like the days of copper tips (and solder with flux, which excluded the oxygen, which is why it made good connections to fresh metal) that you could sandpaper back to goodness. The new type of iron is ruined once it blackens and oxidises.
Alcap 50V Standard Electrolytic capacitors for all audio and hi-fi loudspeaker crossover applications.
I really don't know what is going to work best. 50V or 100V. Low-loss or standard.
Nothing wrong with lead-free solder except it needs a very hot iron. I use a 40W jobbie with a big tip and get in and out in 5 seconds or so. Don't forget to tin the new iron tip first time it warms up too. Not like the days of copper tips (and solder with flux, which excluded the oxygen, which is why it made good connections to fresh metal) that you could sandpaper back to goodness. The new type of iron is ruined once it blackens and oxidises.
Jerry @ Falcon has always recommended to go with the 100V Alcaps, and then of course Low Loss to replace the low loss capacitors (there are a few on the crossovers as I recall - they are marked LL). So, Alcap 100V for most of them, and then the low loss for the marked LL caps.
Even if a lot of caps are buyed to match them for the Left and right speaker I can't understand how it's possible to find the original values of the caps if they have moved ? (I mean this is not exactly the values marked on the body cap as the crossovers were matched at 0.5% to the original factory model !). E.G. on another speaker I have a standard 100 uF cap is exactly measured at 110 uF on each speaker... it's not by chance 🙂 ! the same for an other cap : 11 uF instead the standard 10 uF marked (the both on each speaker which is a high End Boston Acoustic speaker...)
Is there a viscosity advised for the T33 Ferrofluid ?
I can try to measure my Kef 104/2pcbs but don't see where the pcbs are ??? Have the shematic somewhere in pdf but with the standard values of the caps (not the factory model ones) if interested.
Is there a viscosity advised for the T33 Ferrofluid ?
I can try to measure my Kef 104/2pcbs but don't see where the pcbs are ??? Have the shematic somewhere in pdf but with the standard values of the caps (not the factory model ones) if interested.
Last edited:
That is an S Stop module and its relay. The red things are the thick film circuit that drives it. It should be normally closed (conducts unless triggered), but the relay contacts may be bad. It is a very well implemented thermal prediction protection circuit.
David
David