Kef 104/2 speaker question

Hello all,
I bought a replacement capacitor kit for the crossovers in my Kef 104/2 speakers.
I purchased the kit from Falcon Acoustics, and they matched all my old capacitors for value except for two 50u/100v caps. They did not send any 50s, however they did send two 120u/100v. I'm assuming I am supposed to use these in place of the 50u caps. For those of you familiar with these crossovers, the 50u caps are in position C10 on the schematic.
Does anyone know I could plug these in where the 50s were, or that is too much of a value change?
I would/ will take this up with Falcon, however I have read here, and elsewhere, that they are not very good with email replies.
Asking for advice here, thanks in advance.
Dan.
 
Having restored a pair of these exact speakers, I can say this: replace all the caps. They go out of spec, and actually added some distortion. Perhaps they were starting to act as semiconductors.

When finished and paired with a suitable amplifier (they're not a particularly friendly load), they sound rather good. Modern designs get lower into the bass, but apart from that I wouldn't have any complaints.

Chris
 
hi chris:

just curious why you say this:

...
they're not a particularly friendly load
...
Chris

one of the appeals of them to me (I've had a pair since the 1980s) is that they present essentially a 4-ohm resistive load.
resistive-looking loads are easier on amps and I am "guessing" (yeah, I know 😉 ) that anyone who has a pair of these speakers would know better than to drive them with an amp that isn't comfortable with 4 ohms loads.

hmmm ... maybe i'm giving too much credit ....
 
Last edited:
Re-cap Kef 104/2

So what did Falcon come back with?
It took a few emails, but Falcon made good and sent me the two missing caps.
I installed them, along with all the other new caps, and the speakers sound fantastic. I really can't get over how great these speakers sound again.
Of course I also put in new dust caps, grommets, ferrofluid, and cleaned all connections. If anyone is considering recapping their aging Kefs I say go for it!
I've had these for 32 years, I'm 66, so now looking forward to listening to these babies for another 32 years-- ha ha.
 
hi chris:

just curious why you say this:



one of the appeals of them to me (I've had a pair since the 1980s) is that they present essentially a 4-ohm resistive load.
resistive-looking loads are easier on amps and I am "guessing" (yeah, I know 😉 ) that anyone who has a pair of these speakers would know better than to drive them with an amp that isn't comfortable with 4 ohms loads.

hmmm ... maybe i'm giving too much credit ....

Because I measured it as more like a 3ohm load, dropping smoothly around 2.6ohm in the treble.

A decent amp should do the job.

Chris
 
It took a few emails, but Falcon made good and sent me the two missing caps.
I installed them, along with all the other new caps, and the speakers sound fantastic. I really can't get over how great these speakers sound again.
Of course I also put in new dust caps, grommets, ferrofluid, and cleaned all connections. If anyone is considering recapping their aging Kefs I say go for it!
I've had these for 32 years, I'm 66, so now looking forward to listening to these babies for another 32 years-- ha ha.


Have you measured the capacitance of the old caps and the ones of the Falcon kit for curiosity 🙂 ?


Kef 104/2 is a strange beast, works fine 60 cm/80 cm front baffle away from the front wall, they have a great soundstage and disseapper but needs a very strong amp to work best. I never read a fine explanation of what is doing all the complex crossover at least the part // with the drivers !
 
Last edited:
What does the measured impedance look like?

dave

According to KEF Model 104.2 Speaker System Review price specs - Hi-Fi Classic

Lab Tests

We installed a pair of KEF 104.2 speakers as recommended, about 2 feet from the wall behind them and 4 feet from the side walls of the room. The room response was very smooth, varying less than ±4 dB from 150 to 20,000 Hz. The woofer response, measured at the port opening, reached its maximum at 70 Hz, falling at 18 dB per octave below 60 Hz. The output also decreased, to -5 dB, at 130 Hz and more steeply above that frequency. A close-miked measurement of the response of one of the midrange drivers showed a maximum at 250 Hz, with an 18-dB-per-octave decline below 150 Hz and about a 6-dB-per-octave rolloff above 250 Hz. Our close-miked tweeter measurements indicated that the crossover took place at about 3,000 Hz.

Splicing the close-miked bass and midrange curves to the room-response curve in an unambiguous manner was even more difficult than usual, and the composite curve showed an appreciable peak at the woofer's maximum-output frequency. This peak did not appear in our quasi-anechoic FFT response measurements, nor was it particularly obvious in listening tests. The FFT response (at 1 meter) was flat within ± 3 dB from 180 to 20,000 Hz, with the major departure from flatness being a dip of about 5 dB at 3,000 Hz. The response dip was apparently the result of an interference effect between the midrange and high-frequency drivers; not only did it fail to show up in our room measurements, but it disappeared when we measured the speaker's response 45 degrees off its reference axis.

The off-axis response diverged steeply from the axial response above 8,000 Hz, a result of the tweeter's directivity. KEF suggests that angling the speakers toward the center of the room may improve imaging under some circumstances, although in our room this was not necessary. The phase linearity of the system was excellent at middle and high frequencies, as shown by the overall group-delay variation of about 0.1 millisecond from 4,000 to 20.000 Hz and of less than 0.5 millisecond as low as 1,000 Hz.

The 104.2's sensitivity was almost exactly as rated, with a 91-dB SPL measured 1 meter from the speaker when it was driven by 2.83 volts of pink noise in the octave centered at 1.000 Hz. The impedance varied only between 3.3 and 4.6 ohms from 20 to 20,000 Hz. This was not only the most uniform speaker impedance we have ever measured, but its very uniformity attests to its almost purely resistive nature. The woofer distortion, at a constant drive level of 2.5 volts (corresponding to a midrange output of 90 dB), was less than 1 percent from 140 to 75 Hz, rising gently to 2.15 percent at 40 Hz and 4 percent at 30 Hz.

Finally, our pulse power-handling tests produced some of the most prodigious power figures we have yet come up with (in part the high figures resulted from our using a different amplifier than we usually have for these tests and operating it in the bridged mode). With a sine-wave signal on for 1 cycle, then off for 128 cycles, the 100-Hz woofer limit was 1,127 watts. At 1,000 Hz, the system withstood 1,806 watts, and at 10,000 Hz the limit was 1,670 watts. All these figures refer to the amplifier's output at its clipping point; the speaker did not clip or otherwise show signs of overload in any of these measurements.
 
Last edited: