SpeakerTalk

[habrune]

An acquaintace asked me to recap a pair of 105-II's Early serial no. 1113. I measured all the caps and all of them about 10% larger than nominal. Quite normal as I learned from very usefull information from Audiolabtower on his 105's (Thanks, Audiolabtower). And between the pair matching within 2%!

So, I have some hesitations. The set sounds perfect, pinpoint imaging, smooth and transparent. Slightly dim, but only slightly.
Good as the sets from Falcon are (and let's praise all the effort they went through in 2014 to continue their Alcap services), I once recapped a Calinda, and was not entirely happy with the result. Brighter yes. But somehow the overall coherence was slightly less perfect. Not bad, but, although brighter, the quality was not preserved as KEF intended (I felt).

Are the positive reports after recapping an example of cognitive dissonance reduction (given the effort, the money), or is a recap still fulfilling after the final thrill has faded? Too easily one wants to mark the territory by doing something on ones gear, but is it really good?
Anyone with a positive experience in the long run on the 105-II recap?

[audiolabtower]

Well the 2% matching at least means no cap is wildly drifted compared with the other channel. As far as a slight dimness, the design axis is the led full on so have this towards ear height, I assume you have zobels on the tweeters? I wonder if Guru could give an idea if altering the resistor value would reduce the high freq impedance just a little to give a little more energy in that region (amplifier stability taken for granted)?

[habrune]

Audiolabtower, thanks for the reply. LED's are functioning well and I aligned the heads. Marvelous stereo image. The T52 have only the R3 = 47 Ohm as 'zobel', C12 is a short.
In my CS9's I had exchanged the caps in the T52 section for quite expensive foil capacitors but quickly reverted the action. Was quite a relief. (nothing bad about the foils, but I didn't have the time to tune them in with R's.) The owner complies and agrees to leave the 105's as they are.

The owner asks me to shut the S-Stop circuit out, because it comes in at unpredictable moments. I'm willing to do so, because I will make it a reversible intervention (disconnect the leads to the S-Stop print in the head and shorting the relay on the XO print).
The only fault I could detect is a rubbing B300, spikes generated by the rub might trigger the S-Stop, is my hypothesis as it is the only fault I could detect. The rub has not damaged the coil, I measure a neat 7,3 Ohm. As al other units do.

[speakerguru]

[habrune]

Speakerguru, thanks. In that line of thought one could remove the 47R to increase output but I fear the 47R is also there to dampen resonance between the caps and the voicecoil inductance (as does the small 0,03mH coil in series), and one wouldn't like to increase that resonance. So, a think I leave it As Is... and explain to the owner that the slight falling tendency to highs is the voicing of this model. Unless someone with experience proclaims me stupid for NOT recapping.

[audiolabtower]

Well if the impedance can't be reduced easily, and altering cap values too difficult to do it without recourse to measurement the only thing I can think of is replacing the tweeter caps with polyester or polyprops with no extra resistance. This should be brighter, but maybe in the wrong area, could bring up the crossover region too harshly without doing much for the higher treble. As you say maybe leave well alone. If the B300 is rubbing could be worthwhile to turn it upside down if due to gravity sag, this works after some time.

[habrune]

I disconnected the S-stop circuitry totally (diconnecting the 6-strand lead from the inside of the connector on the head, black-blue-red-yellow-white-lightblue). I also bypassed the relay connections.
So, no LED, and no function on the switch on the back of the head, and no protection.
I noticed that the S-stop print takes some signal from the blue tweeter lead (apart from the overall signal before the relay, the red lead).
Maybe, a faulty the S-stop here causes a power drain from the tweeter section hence the slight dimness?
After disconnection the 105-II sounds much better now in overall balance between mids + high.

So, job done I guess, and thanks again for all the support and advice!

For the record: This pair has the 3rd order series cap made of 3 x 120uF and 1x 80uF (all original). Cabinet serial no's 1025 A+B, sp 1117; XO SP1119, serial no's 1113 and 1114.

[speakerguru]

[habrune]

That's great engineering! I love these details, thanks! So, I expect the voltage sensor is high impedance. Further, if it senses mean RMS, it won't bother about EMK spikes supposedly generated bij the B300 rub. Hmm puzzle why this 105 performs now OK when the S-stop is removed. Maybe some other fault..

For the record (2): The B300 (triangle shaped, is SP1071? no markings anywhere, but KEF-museum says so) in this 105-2 has a bigger magnet en slightly more output than my constructor series B300 SP1071 (marked): Magnet diameter 128 mm vs 123 mm, thickness 24mm vs 21 mm.

In overall sound the bigger magnet sounds just right in balance with mid+ highs (it should!), the constructor series B300 slightly too thin (in my room, that is) when mounted in the 105-2.

[speakerguru]

[habrune]

Yes that is inaudible. My experience is that differences of about 1 or 2 dB are in an absolute sense inaudible and neglectible, but when tuning e.g. a tweeter to midrange level small differences of 0.5 - 1 dB make an audible change in overall sound character. I hoped to report a difference in this sense. I agree that size doesn't have to say anything about the field strenght in the gap; my B-sensor is too large to fit in....

[T.O. Chef]

I'd like to tone in on this discussion if that's okay with everyone. Assuming all questions have been answered.

I am working on a recap of some 105's as well. (Original version) and having a mess of trouble trying to get them listenable. The mid's shriek, and the treble is hot in spots. Both mid and HF drivers are dead in spots (as indicated by use of a signal tone generator).
This set came with both an aB circuit with 0.22uF cap, and a tweeter Zobel of C7uF/R47 ohms. The second 3.3uF cap that you would expect to be there is a 2.2uF. There is also a 30uF cap in parallel with the 0.028 inductor on the Tweeter section input. (Odd?)

At this point I have removed both the 30uF cap, The ab cap, and replaced the existing 2.2uf cap on the tweeter with a 3.3 100V LL Alcap. (actual reading 3.9uF). These straightforward modifications have done little to nothing improving the sound, but the Impedance graph seems a bit tamer. But, I still have a whopping peak of 26 ohms at 2500Hz. Should that be? I'm wondering if that peak should be reduced somehow to tame the Mid's ??? How?
I should note that the B-110's have an elevated Inductance readings of 1.28mH / 1.29mH. Would this be upsetting things terribly, and how could they be so far off spec? Perhaps that mid Zobel (7uF/10R) needs revised C/R values?

I have spare B-110b's that have a near on spec .95mH L-reading, though not as well matched in DCR. Both are up near 7.5 ohms. Would substituting these drivers be helpful? Of course I'd rather leave original drivers in place if I can get the crossover right.

Any assistance will be greatly appreciated.
_________________
Music is food for the soul.

[T.O. Chef]

Sorry, I'm made a mistake above in the original layout.. I need to clarify.

The original (as found and I assume original) tweeter section was as follows:
Parallel circuit (30uF / 0.028nH) to 3.3uF series, 0.35mH to ground, 7Uf series, Zobel (2.2uF / 10R) to ground, to fuse to T-52+. There is the Ab section (.022uFspanning the 3.3uF / 7uF cap's)
I'm assuming these inductor values, I have not tested them.

The mid-range is the standard, to be expected, 1.3mH to 30uF series, 10uF to ground, 2mH to ground, 100uF series, 0.25mH series, Zobel (10R/7uF) to ground, fuse, then B-110+.

The bass is 7mH series, 80uF to ground, 33R to ground, 360uF series, 3.5mH series, to B-300+.

That tweeter section seems very complicated. As noted above I striped it down to the simpler layout removing the 30uF cap on the input and the aB cap. I replaced the 7uF with a 3.3uF and left the Zobel 2.2Uf in place.

I hear No Difference with these changes (maybe modest ones) at this point, but the shrieking mid's may be skewing my ears.
The impedance plot stays relatively the same through the 1-20 kH region. Huge peak at 2300 kH.
_________________
Music is food for the soul.

[audiolabtower]

The ~23 ohm impedance peak is normal, I have seen one review with it just below 1kHz and a slightly later one with it just below 2kHz, perhaps the difference whether or not the aB circuit was used, chosen by the computer program based on the driver characteristics? If using the simpler circuit it is important to have the larger 3,3 on the tweeter side. If you can't measure them try padding with a 0.1uF? Mine both originally were 3.5uF and 3.8uf within 1%.

If everything is correct in terms of crossover values and the drivers within spec the only thing I can think of is the positioning. Getting the bass up to match the rest of the range does depend on the room, and the head dispersion is very wide so at least 2 ft out from a back wall to balance and get that exceptional depth imaging.

Are the B300s sealed? Carefully pushed in they should take 5 seconds to return to centre.

[T.O. Chef]

Thanks again Audiolabtower.
Your speedy responses are always greatly appreciated.

I was very careful to ensure the second 3.3uF cap in the chain is of higher value than the first. I have used a supplied set of 105 cap's from Falcon Acoustics where both 3.3's were polyester types and both measured near exactly 3.3. I substituted an Alcap electrolytic for the second cap, and it measures 3.9uF.

Something though is very wrong. The usual forward presentation of the mid range is amplified. The HF seems not too bad now, and the impedance curves across the whole range seems reasonable aside from the 23 ohm peak at 2.5 K.

I'm left now to assume that the B-110 drivers "out of spec" L(e) of 1.28 mH is wreaking havoc. Is this possible?

The bass is very good, and you are correct (you and Speakerguru) that room and placement make a huge difference in that regard. The cabinets have had meticulous attention to ensure a good tight seal everywhere.
Not, my work and testing is all done in a smallish space in the near-mid field listening position, but I am very familiar with the effects of the space and can accurately assess bass response regardless.

It's fortunate that I have such ready access to the crossover from the back and can make some modifications "on the fly" using alligator clips and jumper wires.
One very interesting thing is there seems to be near zero effect from adding or subtracting a cap in the tweeter Zobel and perhaps that's why some versions (withought the ab cap) have left this Zobel out entirely or left in only the 47R resistor.
For fun and frolic though, I plugged in a 100uF cap at that spot and observed a slight leveling of the impedance slope up the 23 ohm peak at 2.5K. I did not listen to the effect however, just observed it.

Later today I hope to swap out a B-110 for one with a closer to spec L(e). Here's hoping that makes a more enjoyable listening experience! I'll report back on it.
Thanks Again. Though lots of fun at first, this is getting a bit tedious now.
_________________
Music is food for the soul.