KEF 104/2 cap replacement

Hi - just wondering if there are any quick opinions on replacing caps on a pair of old Kef 104/2's.
options;
1; a dedicated cap set for 104/2 from falcon acoustics - all alcap electrolytic 'cept one of the tweeter parallel signal path caps (C.13) with polystyrene.
KEF 104/2 Replacement Capacitor Set
2; budget improvement(?) - replace mainly with alcaps though replace tweeter caps with budget MKP, and perhaps mundorf e-cap for signal path 100uf's (C6,7) on mid - or even Chinese 'Audiophiler' if space (will investigate) - these seem to have a reasonable write -up at humble home made hifi. Humble Homemade Hifi - Cap Test

The budget is low... just after a quick opinion if that is possible - please don't do to any trouble if this question is more complex than I think! - thoughts very much appreciated though.
 

Attachments

  • kef XO.jpg
    kef XO.jpg
    158.1 KB · Views: 667
Hi - just wondering if there are any quick opinions on replacing caps on a pair of old Kef 104/2's.


options;


1; a dedicated cap set for 104/2 from falcon acoustics - all alcap electrolytic 'cept one of the tweeter parallel signal path caps (C.13) with polystyrene.
KEF 104/2 Replacement Capacitor Set



2; budget improvement(?) - replace mainly with alcaps though replace tweeter caps with budget MKP, and perhaps mundorf e-cap for signal path 100uf's (C6,7) on mid - or even Chinese 'Audiophiler' if space (will investigate) - these seem to have a reasonable write -up at humble home made hifi. Humble Homemade Hifi - Cap Test


The budget is low...


just after a quick opinion if that is possible - please don't do to any trouble if this question is more complex than I think! - thoughts very much appreciated though.

I have two sets 104/2, both are single wired.
One I did the re-cap using the Falcon kit. All old caps measured spot on (tested with the Peek meter). So I guess the replacement wasn't necessary afterall.
However, it turned out that the tweeter oil had gummed up.
Cleaned out the old oil, and replaced it.
The oil replacement was by far the most important (besides from woofer donuts, which also needed replacement).

Second 104/2, is about as old, but had the tweeters replaced some ten years ago. Those 'stock' 104/2 sound really close to the re-done ones.

I would start with the tweeters, and the woofers, and wait on buying new caps.
As for oil replacement.
The tweeters is rather easily disassembled, and easy to clean using a q-tip with some IPA. The magnet gap needs some (10-15) paper strips to clean out the oil.
Re-applying is (in my experience) best done using a syring.
Lots of 'testimonials' saying that one just needs to eye-ball it and let the magnet gap pull the oil itself... in my case the result was two tweeters with significant different sensitivity.
Using the syring to control the amount of oil worked fine in my case.

As mentioned I only have experience from two sets; your case might be different.

isn't the crossover in these relatively difficult to get to for folks that would want to attempt this kind of mod?
btw, i really like this speaker. i still own the pair i bought back in maybe 1985(?).

mlloyd1

One needs to remove the midrange/tweeter enclosure (four screws), then a MDF cover (four more screws), to get access to the upper woofer.
Remove the nut in the center of the woofer, and the three screws.
Be careful to not loose the rubber washers.
Remove the woofer.
The bottom cover is held in place by some eight screws.
The bottom woofer is removed as the top woofer, leaving the connecting bar hanging inside.

The xover is mounted inside the reflex enclosure with some three/four wing nuts.
The various connection spades/jacks do benefit from being cleaned.
 
The one issue I can think of is the ESR of those old caps. They tended to be relatively high, while MKP are lower. Look at that before doing any major overhaul.



Yes, deal with any mechanical issues first. I like to use the Mundorf MKP's, cheap, detailed, if dark. Break in for 48 hours at least.
 
Thanks - yes the donuts and ferro are a given, and thanks for mundorf mkp view.


With the ESR (good point thank you) I am thinking this will probably only have a noticeable impact on the higher values(?) - the 2x100u high pass caps on mid (other high values I will alcap). So only one point of attenuation needed if problematic. the flip side being that potential gains at this point are likely to be very worthwhile?
(I know the falcon kit has a lower value changed to polystyrene...)


The one issue I can think of is the ESR of those old caps. They tended to be relatively high, while MKP are lower. Look at that before doing any major overhaul.


Yes, deal with any mechanical issues first. I like to use the Mundorf MKP's, cheap, detailed, if dark. Break in for 48 hours at least.
 
The one issue I can think of is the ESR of those old caps. They tended to be relatively high, while MKP are lower. Look at that before doing any major overhaul.



Yes, deal with any mechanical issues first. I like to use the Mundorf MKP's, cheap, detailed, if dark. Break in for 48 hours at least.

As far as the PEEK ESR meter measured, all the old caps was spot on for both value and ESR (in the range of 0.1 - 0.04ohm)
 
Member
Joined 2008
Paid Member
I spent quite a bit of time researching this same question.

I have a bi-wire pair of the 104/2 that I picked up abandoned at a repair shop after they had the woofer surrounds replaced.

My crossovers are accessed by turning the speaker on its side and removing a panel on the bottom of the speaker.

I read various discussions about the tweeters, some on here. There are substitutes that have been used successfully. One was made by Vifa of which I have a pair sitting on a shelf. There is also currently a vendor on eBay selling “replacements” that look to be virtually the same driver with its faceplate removed and some foam tape applied to the face for $45 each I think.

The thing about these crossovers that I read is that they had there components painstakingly matched to very close tolerances (2% or better I believe) at great time and expense in the KEF factory. The general opinion that was voiced was that it would be pointless to just start ordering replacement caps that have the same values as they would never match the accuracy of the original caps and it would be impossible to obtain the same crossover performance which in a fourth order design like this would be very obvious. These are not my words, just what I found on the subject.

I also found the Falcon recap kits. They originally listed that they had a limited supply of kits that approached the factory selection standard of about 2%. When I sent an inquiry I got no response. I think the kit they still list has a 5% selection standard.

My feeling about the KEFs has been tainted by my involvement with other speaker projects that I have attained some very desirable results with. I haven’t let them go. They have been retained for HT use. If they start sounding really bad or something craps out I will probably take a shot at restoring them.

I’d try the replacement tweeters I have first and then order some ferrofluid to try and restore the originals. As others have said, I’d inspect and measure the original caps for leaks/variances first before replacing them. If it ain’t broke don’t fix it.
 
thanks for your reply - very interesting
falcon told me they only have the 5%
I will go for ferrofluid replacement with tweeters, though note falcon recommend Audax TWO25A2 as an exact replacement - about $100 a pair
Its quite a job taking these speakers apart - I did it today - so while I'm in there I'm tempted to replace caps anyway, particularly as am doing the job for a friend who wants them to last 'forever' - they may still be OK now but more unlikely when they reach 50, and he may not have the free help I am giving him at the moment.
If I had access to precision cap checking equipment I might take a different approach, but I don't.
I've an inkling these speakers need careful pairing...
R





I spent quite a bit of time researching this same question.

I have a bi-wire pair of the 104/2 that I picked up abandoned at a repair shop after they had the woofer surrounds replaced.

My crossovers are accessed by turning the speaker on its side and removing a panel on the bottom of the speaker.

I read various discussions about the tweeters, some on here. There are substitutes that have been used successfully. One was made by Vifa of which I have a pair sitting on a shelf. There is also currently a vendor on eBay selling “replacements” that look to be virtually the same driver with its faceplate removed and some foam tape applied to the face for $45 each I think.

The thing about these crossovers that I read is that they had there components painstakingly matched to very close tolerances (2% or better I believe) at great time and expense in the KEF factory. The general opinion that was voiced was that it would be pointless to just start ordering replacement caps that have the same values as they would never match the accuracy of the original caps and it would be impossible to obtain the same crossover performance which in a fourth order design like this would be very obvious. These are not my words, just what I found on the subject.

I also found the Falcon recap kits. They originally listed that they had a limited supply of kits that approached the factory selection standard of about 2%. When I sent an inquiry I got no response. I think the kit they still list has a 5% selection standard.

My feeling about the KEFs has been tainted by my involvement with other speaker projects that I have attained some very desirable results with. I haven’t let them go. They have been retained for HT use. If they start sounding really bad or something craps out I will probably take a shot at restoring them.

I’d try the replacement tweeters I have first and then order some ferrofluid to try and restore the originals. As others have said, I’d inspect and measure the original caps for leaks/variances first before replacing them. If it ain’t broke don’t fix it.
 
No specialized precision equipment, a multimeter able to measure from 0Vrms and up, one resistor, a laptop headphone out, a sine wave frequency generator software (REW) and some hook-up wire (mini mono or stereo plug). After that some calculations and you are done.
 
Thanks for everyone's input. In the spirit of sharing my findings;
Given my friend wants to keep these 'forever' I changed all the caps - some of which were showing signs of slight swelling..
Donuts and ferro done. grommets replaced. 13+14 Janzen crosscap 7.5uf. 15,16, 8, Janzen MKP, 6&7 mundorf plain ecap, all others alcap.
significant overall improvements reported, particularly detail and clarity in mid, extension and clarity in treble, all without detriment to character of the speaker. total cost inc. delivery charges just under £200. Given you can get a set of these for £3-400, surely this makes for a bargain proposition for budget high-end...?
 
I can confirm using larger values as direct replacements, MKP's elsewhere, mundorf Ecap electro for 100uf mid (should have by-passed) is a significant improvement particularly in midrange detail and clarity. whether the response curve is less accurate is secondary to the fact classical acoustic instruments have much more realism due to the textural detail (e.g. being able to clearly hear harmonics in church organ tone.)
 
I have almost zero output from my one channel. These speakers were working fine up until that morning I went to play Handel's Messiah. I checked all of the speaker cones and they are good. I took out the crossover board from behind the mid range and tweeter and then found another board at the back of the speaker cabinet at the inputs. There was corrosion on a trace on the board. This was resulting in 2000 ohms resistance on a pcb trace! I cleaned the corrosion and soldered a 1/16 piece of solid copper wire over the bad area. Also had to grab the leg of one of the very small caps that work in conjunction with the mid/tweeter cutout relay that is for safety reasons. It senses with the tweeter has been over voltage and it cuts out the signal to the mid/tweeter crossover board. My problem is I can't get the input board to work with the woofers. I have hooked up the woofers to another stereo and they work fine. Hook them up to the incoming board without the mid/tweeter board or speakers attached and the woofers are severely muddled. Any ideans? Try to find the schematic for this incoming board is a joke. I finally drew out my own....
Bottom of Incoming Circuit Board Before Cleaning.jpg
Crossover components.jpg Mid/Tweeter Crossover Board
 

Attachments

  • Input crossover board at speaker inputs.jpg
    Input crossover board at speaker inputs.jpg
    367.4 KB · Views: 116
I have almost zero output from my one channel. These speakers were working fine up until that morning I went to play Handel's Messiah. I checked all of the speaker cones and they are good. I took out the crossover board from behind the mid range and tweeter and then found another board at the back of the speaker cabinet at the inputs. There was corrosion on a trace on the board. This was resulting in 2000 ohms resistance on a pcb trace! I cleaned the corrosion and soldered a 1/16 piece of solid copper wire over the bad area. Also had to grab the leg of one of the very small caps that work in conjunction with the mid/tweeter cutout relay that is for safety reasons. It senses with the tweeter has been over voltage and it cuts out the signal to the mid/tweeter crossover board. My problem is I can't get the input board to work with the woofers. I have hooked up the woofers to another stereo and they work fine. Hook them up to the incoming board without the mid/tweeter board or speakers attached and the woofers are severely muddled. Any ideans? Try to find the schematic for this incoming board is a joke. I finally drew out my own.... View attachment 1016261 View attachment 1016262 Mid/Tweeter Crossover Board
The pair I did had the one board - as all the utube and other guides suggest too? - Was your speaker modded? - or maybe a very early example??