I purchased a pair of 1977's Cantatas in very good shape for a very attractive price. The set had been already recapped with a Falcon kit. With this, they sounded good but strangely dull. The highest sparkle was somehow not there. Or it was there but squashed flat. I did a simple frequency check and could hear both tweeters up to 12000 Hz (above this I think my aging ears are limiting).
The treble became a bit better after I put low loss caps into the tweeter section (WIMA MKP10 0.1uF, Mundorf polys 3.3uF and 6.8uF and Mundorf E-caps plain 4.7uF) . But in comparison to the 105.1 which I also have it is still dull. It is even dull with the +2 db setting. The highest frequencies are not really there.
I don't have measuring equipment. What could be the cause of this? Aging T52s? Both with the same aging defects? The T52s in the 105.1s are still great. Same vintage. T52s don't use ferrofluid. I will check the 2R2 resistor R5 in the naB network next time I open them up and will change the 3.3uF to an even better quality cap. But according to my experiences with lots of vintage KEF speakers the Cantatas should sound fantastic even after a recap with electrolytics.
What could be amiss? I cleaned the switches and all connectors already. I did not look a the bass crossover, as I need to take out the bass driver to access it.
Any more ideas? Thank you!
crossover schematics here:
http://www.hifiloudspeakers.info/Anatomy/Crossovers/Cantata/CantataCrossoverIndex.html
The treble became a bit better after I put low loss caps into the tweeter section (WIMA MKP10 0.1uF, Mundorf polys 3.3uF and 6.8uF and Mundorf E-caps plain 4.7uF) . But in comparison to the 105.1 which I also have it is still dull. It is even dull with the +2 db setting. The highest frequencies are not really there.
I don't have measuring equipment. What could be the cause of this? Aging T52s? Both with the same aging defects? The T52s in the 105.1s are still great. Same vintage. T52s don't use ferrofluid. I will check the 2R2 resistor R5 in the naB network next time I open them up and will change the 3.3uF to an even better quality cap. But according to my experiences with lots of vintage KEF speakers the Cantatas should sound fantastic even after a recap with electrolytics.
What could be amiss? I cleaned the switches and all connectors already. I did not look a the bass crossover, as I need to take out the bass driver to access it.
Any more ideas? Thank you!
crossover schematics here:
http://www.hifiloudspeakers.info/Anatomy/Crossovers/Cantata/CantataCrossoverIndex.html
The capacitor shouldn't be made responsible for fixing your treble... How are the switch contacts?
I agree. The caps I found however (Falcon kit) where not specified low loss, or if they where, they were physically the same as the non LL versions on the same board, which makes me a bit suspicious. This could make a difference indeed, as LL was specified for the aB section caps.
I am asking myself if it would it make sense to check if one or both of the tweeters are wired in the wrong phase? awful lots of cables in there however (due to the switches).
Thanks @AllenB for your quick reply!
You should still be able to hear +2dB, otherwise you could try shorting the resistors.
Polarity could be a concern if they've been changed. A response measurement is simple for a polarity check but it can be done by ear.
Polarity could be a concern if they've been changed. A response measurement is simple for a polarity check but it can be done by ear.
Thanks.
I do hear the difference in the treble switch settings and can hear +2dB. but they sound dull even with +2dB. it is as though the middle trebble (sibilance in human spoken language for example) came through, but the higher frequency stuff (fine cymabl overtones) didn't.
I do hear the difference in the treble switch settings and can hear +2dB. but they sound dull even with +2dB. it is as though the middle trebble (sibilance in human spoken language for example) came through, but the higher frequency stuff (fine cymabl overtones) didn't.
Many many years ago I repaired 4 of these T52 tweeters .. I was shocked to see that two of them had a SECOND half-roll suspension glued UNDER the first one!
I removed the second suspension and the glue , re-soldered the coil wire etc and everything worked fine! But I checked not the freq response ....
So maybe your dull tweeter is one of this weird experimental item!
bye
I removed the second suspension and the glue , re-soldered the coil wire etc and everything worked fine! But I checked not the freq response ....
So maybe your dull tweeter is one of this weird experimental item!
bye
Thanks.... impressive. I'll try to have a look... but am not sure if I'll be able to dig this deep into the tweeter without damaging it.
I'll directly wire the chassis to the crossover at some point to check the influence of the wiring and switches.
I'll directly wire the chassis to the crossover at some point to check the influence of the wiring and switches.
Update after some experimentation and fault finding:
The tweeter network is now all film caps. I swapped the 10uF 100v shunt cap in the mid section network for a 10uF polypropylene. Images of the original crossover show a LL cap here as well... not a "lossy" one as I found, provided by the Falcon capacitor replacement kit.
This now made the treble much better (always amazing how much component character changes in the different sections of a crossover influence one another). Still something was off – two much mids. So I put 2x4R7 5 Watt resistors in parallel to the midrange input. I assume that the midrange section even in -2dB setting was a bit more prominent than treble and bass from the get-go.
The Concertos had the same mid prominence, and also there it helps to reduce mid output in order to gain clarity. See my thread on an improved KEF Concerto crossover: https://www.diyaudio.com/community/...d-dn12-crossover-with-full-ab-circuit.370666/
Now things were much better. Not perfect, but much better. I then realized that the B110 midrange chassis was wired wrongly. It must be wired out of phase according to the schematics. But the person who did the recap obviously had soldered it back as it seemed right – green wire to +, black to -. But it needs to be the other way round. With the mid drivers in the right phase orientation, a good bit of the "off" sensation is gone now.
I still feel the speakers are not totally right. They could have still more sparkle, but it is much better now. They don't seem "off" anymore. No need to set the treble knob to +2dB.
I have the midrange now at -2dB – which gives an impressive series resistance of 6.6 Ohms for the mid driver. Maybe it is slightly out of spec? Perhaps here a 1 Ohm resistor in series to the shunt cap would help as it did in the Concerto upgrade. I might try after some more listening.
Looking at the schematics of the mid section of Concerto, Cantata, CS7 and Ref 105.1 and .2, it is interesting to see the evolution in terms of how much steeper the slopes of cutting off the B110 (a or b) became with time. The series coils were also providing series resistance – in case of the CS7 2.5 Ohms.
That's the state of affairs right now. After more listening my opinion might solidify – or change. I'll report.
The tweeter network is now all film caps. I swapped the 10uF 100v shunt cap in the mid section network for a 10uF polypropylene. Images of the original crossover show a LL cap here as well... not a "lossy" one as I found, provided by the Falcon capacitor replacement kit.
This now made the treble much better (always amazing how much component character changes in the different sections of a crossover influence one another). Still something was off – two much mids. So I put 2x4R7 5 Watt resistors in parallel to the midrange input. I assume that the midrange section even in -2dB setting was a bit more prominent than treble and bass from the get-go.
The Concertos had the same mid prominence, and also there it helps to reduce mid output in order to gain clarity. See my thread on an improved KEF Concerto crossover: https://www.diyaudio.com/community/...d-dn12-crossover-with-full-ab-circuit.370666/
Now things were much better. Not perfect, but much better. I then realized that the B110 midrange chassis was wired wrongly. It must be wired out of phase according to the schematics. But the person who did the recap obviously had soldered it back as it seemed right – green wire to +, black to -. But it needs to be the other way round. With the mid drivers in the right phase orientation, a good bit of the "off" sensation is gone now.
I still feel the speakers are not totally right. They could have still more sparkle, but it is much better now. They don't seem "off" anymore. No need to set the treble knob to +2dB.
I have the midrange now at -2dB – which gives an impressive series resistance of 6.6 Ohms for the mid driver. Maybe it is slightly out of spec? Perhaps here a 1 Ohm resistor in series to the shunt cap would help as it did in the Concerto upgrade. I might try after some more listening.
Looking at the schematics of the mid section of Concerto, Cantata, CS7 and Ref 105.1 and .2, it is interesting to see the evolution in terms of how much steeper the slopes of cutting off the B110 (a or b) became with time. The series coils were also providing series resistance – in case of the CS7 2.5 Ohms.
That's the state of affairs right now. After more listening my opinion might solidify – or change. I'll report.
A little update after some time of listening and experimenting. I settled with all polypropylene film caps (mundorf standard 250V) in the tweeter network and also in the mid zobel. I resorted to putting back the 100V blue Alcaps from the Falcon kit in the mid input and the mid shunt.
I find this mid shunt with the B110 of utmost importance, it is so also in the 105 and Concerto... the ESR / resistance needs to be somewhat elevated, but the cap quality needs also to be good. see: https://www.diyaudio.com/community/...th-full-ab-circuit.370666/page-2#post-6808367
I also changed the 5R resistors in the aB network against new 4.7R ones. One 5R measured 6R, the other was fine. Although it does not really make sense, I found the highs much better after this swap.
The Cantatas need to be well away from the listening position, I'd say at least 2.5 meters, if you are sitting in an armchair. Preferably farther away. Otherwise they will be too low, the highs will be toned down and the imaging will suffer. The sweet spot is rather limited with these. Maybe useful to put them one stands, the bass might even become better (there is a LOT of it).
Now, if listened to from the right position, and after all changes made, these sound great: detailed, resolving, dynamic, fast, warm, with outstanding imaging. They sound better with a refurbished/upgraded Sansui TR-707a than with a refurbished/upgraded Kenwood KA-7300. (The KEF 105 sound better with the Kenwood.) In my overdamped room (many books) I have the treble set to +2dB.
This setting actually means bass: +-0dB, mids: -2dB, treble: +-0dB. In the "normal" setting for mids and highs there is a 2R2 resistor in series with the speakers. This – 2R2 in series with the B110 – is what I found most pleasing when I modified the Concerto crossover as well.
In still later iterations of 3-way crossovers with the B110 KEF used the internal resistance of the series coil in the mid circuit to down down the mid volume.
Comparing the Cantata, 105, and Concerto which I also own (modiefied crossover with aB section), I'd say the 105 is the most linear sounding, with the least bass hump. The Concerto sounds very close, with that particular seet treble and the mid niceness of the original B110. The Cantata is the most meaty sounding of the three, spectacular in the right listening window. Great rendering of piano music.
I find this mid shunt with the B110 of utmost importance, it is so also in the 105 and Concerto... the ESR / resistance needs to be somewhat elevated, but the cap quality needs also to be good. see: https://www.diyaudio.com/community/...th-full-ab-circuit.370666/page-2#post-6808367
I also changed the 5R resistors in the aB network against new 4.7R ones. One 5R measured 6R, the other was fine. Although it does not really make sense, I found the highs much better after this swap.
The Cantatas need to be well away from the listening position, I'd say at least 2.5 meters, if you are sitting in an armchair. Preferably farther away. Otherwise they will be too low, the highs will be toned down and the imaging will suffer. The sweet spot is rather limited with these. Maybe useful to put them one stands, the bass might even become better (there is a LOT of it).
Now, if listened to from the right position, and after all changes made, these sound great: detailed, resolving, dynamic, fast, warm, with outstanding imaging. They sound better with a refurbished/upgraded Sansui TR-707a than with a refurbished/upgraded Kenwood KA-7300. (The KEF 105 sound better with the Kenwood.) In my overdamped room (many books) I have the treble set to +2dB.
This setting actually means bass: +-0dB, mids: -2dB, treble: +-0dB. In the "normal" setting for mids and highs there is a 2R2 resistor in series with the speakers. This – 2R2 in series with the B110 – is what I found most pleasing when I modified the Concerto crossover as well.
In still later iterations of 3-way crossovers with the B110 KEF used the internal resistance of the series coil in the mid circuit to down down the mid volume.
Comparing the Cantata, 105, and Concerto which I also own (modiefied crossover with aB section), I'd say the 105 is the most linear sounding, with the least bass hump. The Concerto sounds very close, with that particular seet treble and the mid niceness of the original B110. The Cantata is the most meaty sounding of the three, spectacular in the right listening window. Great rendering of piano music.
Just answering the thread title, sounds about right. The 2” dome does not do real well up top
dave
dave
I think it does. It does so if you find the right placement for the speakers in case of the Cantata. It does stunningly so in the KEF Reference 105, with no difficulties in speaker placement. No problem with the tweeter from my point of view. That's what I wanted to say.
Reporting some results here in order to close this thread.
First, answering @planet10 above: In the KEF speakers I have (R105.1, Cantata KIT, plus the commercial CANTATAs described here), the 2" T52 dome does an admirable job. It is more directive, that's why it works best in the small, moveable "robot" heads of the 105. Those have no problem at all with treble. But in the other designs treble is great as well if one remains in the listening window.
I finally fixed the Cantatas. They now also have great treble. Their general character is dynamic, fast, meaty, powerful with shockingly great imaging and fantastic resolution. With the changes I have made now (I can imagine they're final) the KEF Cantatas reveal themselves to be brilliant speakers. It is interesting to hear how they differ from the 105.1 and the Concertos which I also own. More brutal and forward than both. They work well with low-powered amps.
The Concertos (crossover rebuilt with an aB circuit and other changes, see here: https://www.diyaudio.com/community/...d-dn12-crossover-with-full-ab-circuit.370666/) and the 105 (crossovers restored as well) sound more similar among them. The Cantata has a lot energy and slam with which one can fall in love head over heels.
What did I do in this final stage?
I changed the 3.3uF input cap of the treble circuit to Mundorf Evo Alu, replaced the 5R resistor in the aB section with a Mundorf MOX, replaced the existing 100uF 100v Alcap from the Falcon replacement kit in the bass crossover with a Mundorf 100uF 100v raw. And I replaced the remaining 6.8uF 100v Alcap Falcon replacement cap in the Zobel of the B110 with a LOWER QUALITY cap which I had handy – a 6.8uF 100v Audyn cap. Looks tiny. To avoid, one thinks.
My experience with the B110 section in classical KEF three-way designs is that it requires a lot of damping for the parallel caps. The thought "I'll replace the old Alcaps with fresh polypropylene (or even plain foil bipolars)" will yield results that sound off, make the mid range muddy and too prominent, and pull the treble down.
Very interesting. I think someone should model this at some point (@system7?)
First, answering @planet10 above: In the KEF speakers I have (R105.1, Cantata KIT, plus the commercial CANTATAs described here), the 2" T52 dome does an admirable job. It is more directive, that's why it works best in the small, moveable "robot" heads of the 105. Those have no problem at all with treble. But in the other designs treble is great as well if one remains in the listening window.
I finally fixed the Cantatas. They now also have great treble. Their general character is dynamic, fast, meaty, powerful with shockingly great imaging and fantastic resolution. With the changes I have made now (I can imagine they're final) the KEF Cantatas reveal themselves to be brilliant speakers. It is interesting to hear how they differ from the 105.1 and the Concertos which I also own. More brutal and forward than both. They work well with low-powered amps.
The Concertos (crossover rebuilt with an aB circuit and other changes, see here: https://www.diyaudio.com/community/...d-dn12-crossover-with-full-ab-circuit.370666/) and the 105 (crossovers restored as well) sound more similar among them. The Cantata has a lot energy and slam with which one can fall in love head over heels.
What did I do in this final stage?
I changed the 3.3uF input cap of the treble circuit to Mundorf Evo Alu, replaced the 5R resistor in the aB section with a Mundorf MOX, replaced the existing 100uF 100v Alcap from the Falcon replacement kit in the bass crossover with a Mundorf 100uF 100v raw. And I replaced the remaining 6.8uF 100v Alcap Falcon replacement cap in the Zobel of the B110 with a LOWER QUALITY cap which I had handy – a 6.8uF 100v Audyn cap. Looks tiny. To avoid, one thinks.
My experience with the B110 section in classical KEF three-way designs is that it requires a lot of damping for the parallel caps. The thought "I'll replace the old Alcaps with fresh polypropylene (or even plain foil bipolars)" will yield results that sound off, make the mid range muddy and too prominent, and pull the treble down.
Very interesting. I think someone should model this at some point (@system7?)
Last edited: