SpeakerTalk

[speakerguru]

Capacitor inductance will not be significant within the audio range. ESR is very important (see my other posts).
0.5 Ohms in series with a 4 Ohm tweeter is around 1dB attenuation.

[proffski]

[iso]

Where this ESR matters is interesting question.

In middle section of Cantata crossover there is 0,85 mH coil with 0,7 ohm DCR, 0,25mH coil with 0,4 ohm DCR and 2,2 ohm attenuation resistor in flat position in series to 50uF cap... so 0,5 ohm DCR of cap does attenuate signal somewhat, but so do other components in circuit too. So most obvious question is do electrolytic have other ill effects compared to film caps and why?

In Cantata treble position situation is slightly different, there is only 2,2 ohm attenuation resistor in series to caps and signal. Rather modest riple current capability of electrolytics might have some ill effects, but most likely in bass section.

So it would be nice to know, if there any reason to believe that other characteristics of electrolytics hide low level information of signal and how and why.

Best Regards

[proffski]

A few points here are worth perusal: http://sound.westhost.com/articles/capacitors.htm
_________________
I contend that for a nation to try to tax itself into prosperity is like a
man standing in a bucket and trying to lift himself up by the handle.
-Winston Churchill

[audiolabtower]

Very good article, thanks for the link

[iso]

I replaced tweeter section electrolytics with film caps in Cantata Kit crossover. I noticed slight improvement in articulation and maybe slight increase in treble level. But resolution was not acceptable yet.

Then I figured that there may be some interaction between inductors due control panel. In this Cantata Kit, level switches are attached directly to crossover board and consequently there is only maybe 7-10 mm distance between inductors and controll panel. I suppose panel is manufactured from aluminium.

As I felt that there must be way to improve transparency of sound, I decided to try if ihis interaction was the propblem. I attached crossovers with rubber bands to their cavity and removed control panel. Considerable imrovement in resolution was result of this modification.

I suppose that crostalk between coils was biggest proplem with these Cantatas. Coils can be scaled for correct values close to metal panels or far away. I suppose there is no way reduce crostalk between coils if they are close to same metal plate.

Coils are still quite close to B110 and T52 magnets...

[iso]

I ordered pile of MDF to make proper enclousures for Cantatas, as my latest findings have been quite promising for upgrading Cantatas.

I am now wondering if it is possible that T52 SP1049 is ferrofluid damped tweeter. As far as I remember there was no note about it on Kef datasheets. If so... dried fluid can be one reason for mediocre low level resolution.

I have also been thinking if aluminium control panel has been acting as vavequide for B110 and T52. If so, it might be good idea to make proper quides from MDF or other nonmagnetic or non resonant material, if original crossover position is retained.

I have always admired Cantata design as measured response on axis and of axis is very controlled. There are practically no irregularities even at crossover frequencies.

If anybody have information about these two issues, please let me know.

[ColinR]

[proffski]

[iso]

I have made new cabs for these first generation Cantata kits according SC7 plans. However as i made front baffle removable, I used two self braces to stiffen cabs and I relocated hi/mid crossover well on back cover between braces to make back cover stiffer. This way, I was also able to mount both crossover boards away from magnetic fields of drive units.

Cab is made from 19mm MDF, 50% inside covered with 3mm bitumen pads. Inside walls covered by 30mm foam, lower 2/3 of back covered with 120mm foam and upper third full of BAF. Outsides are veneered by 1,5mm walnut veneer, waxed and polished. They look quite nice.

Front baffle is made of two 12mm MDF sheets glued each other. This way I was able to sink B139 in flush to baffle and T52 and B110 to roughly same plane as in original design. Layout is similar to Cantata mirror imaged L&R design.

I did not use metallic Cantata plate, but radiused B110 and T52 mounting holes roughly to 30mm radii. Front baffle is roughly 12mm higher than sides, so grille can be fixed similar way than ATC and Meridian used to fix grilles. Grille is now as close to front baffle as possible and there is minimum amount of corners to cause diffraction.

Crossovers are rebuilt with new 2% alcaps on bass and mids, and matched KT and MKT:s on HF. I measured inductors to be within 5% of spec, resistors seems to be OK. Switches have been cleaned with 1200 grit wet and dry sandpaper and WD40 was used as cutting oil. Residue was rinsed with wax and silicone cleaner. Connections are made by Monacor bi wireable terminals. Fuses are also bypassed.

However now I try to explain my problems.

Most irritaiting issue seems to be hardening of some female voices, strings and upper register of pianos. It is most obvious on pianos. Tilting speakers slightly backwards (speakers are on the floor) seems to lessen this hardening. There is also trace of honkines that Collin earlier described as B110 virtue, but I do not consider this any major fault. Most often actually no honkines exists.

I also got slight improvment on this issue when I removed original Kef wadding from mid enclousure. Then I placed 2pc 30mm sheets of foam on back cover of cavity and placed 50% of original wadding over foam.

I also tried to reduce zobel resistor values from 10ohm to 8,2 or 7,5 and increase cap values from 7 to 8,5 or 10. These mods reduced hardness but they actually made soundstage more diffuse.

Bass seems also be too dry to my taste. LF extension seems to bee very flat to 40z in room, level on 30 Hz is reduced. Tympanis and electric bass are reproduced very nicely but double bass on jazz recordings like Art Pepper Meets The Rhythm Section is thin. It feels like no fundamental are reproduced. On positive side I must add that upper bass is very articulate, say when double bass is played with bow on same recording.

Extreme Hf seems also be quite dry, but I suppose reason for this is dia of T52 and there is nothing that can change this.

Soundstage of Cantatas is somehow belivable and honest. My main speakers are Wilmsow/ATC100. They somehow have more believable frequency balance and reproduction of low level details... but somehow Cantatas seems to have more honest soundstage.

I feel also that level controls are best at mid position contrary most suggestions.

I have used to drive Cantatas now CJ Premier 11 with 0,5 output impedance, so this amp should be OK to Cantatas.

Cantatas depth, width and height of soundstage seems to be reproduced very nicely, so I hope that somebody can give me sound advice on above mentioned issues.

Best Regards

[iso]

I placed 110 mm dia felt ring with 43 mm dia hole around dome of T52, this seems to cure most of hardening of voices I mentioned. JBL has used similar foam rings around LE25 tweeters and I suppose many other drive unit makers too. This ring seems to make also upper treble better focused.The trouble seems to be around T52, not B110 as I first thought.

Metal plate on original Cantata was supposed to form crude wave guide to B110 and T52, similar that LS5/8 front baffle design was for 12" woofer. I thought that I improved this design when I mounted MF and HF units in depth of 12mm from front plane (original Cantat kit mounted them in 10mm depth) and radiused mounting holes.

Now I am wondering if this uncontrolled diffraction is from radiused edges of mounting hole or from front plate of T52. When you use foam or felt to absorb diffraction you usually lose important information too.

Martin Colloms describes shallow contoured horn, properly curved to provide termination to dome circumference in his High Performance Loudspeakers book as ideal mounting for domes. This would probably be better device than felt ring.

I now have roughly 6-7 mm depth for for this device between front face of T52 and front plane of Cantatas. Is this depth enough for horn and is is possible to fabricate this horn from MDF or hardboard with router and basic woodworking tools to decent accuracy.

If anybody have any idea about this issue, your advice is welcome.

[iso]

Colin, Audiolabtower, Proffski and Speakerguru have provided usefull information for my Cantata project. Project has been very interesting but more difficult that I expected.

Now I have sorted out most hardness on piano and female voice. It seems that when I assumed that Conrad Johnson Premier 11A was enough to drive Cantatas, I was wrong. When I switched my 70:s EAR 509 pair in action, reproduction of sound became very much smoother. It seems that audible distortion especially in middle register did become much lower. Situation seems be even better if I drive Cantatas from 4 ohm tap.

Bass quality improved when I used BAF to make 100% fill on cab. CS7 plans were probably fine for new drive units, but upper bass quality seems to be better on 100% fill. It was odd to note that only negative reaction for 100% fill is due slightly inferior articulation on low bass.

As foam rings around drive unit seems now degrade transparency, I have now removed them and try to sort upper register hardness some other way. Therefore your help would be most welcome.

First... Colin mentioned in his earlier tread that especially first generation Cantatas blew quads due problem reactance problem around 300 Hz. When I compare 1st and 2nd generation crossovers ,only difference in scematics is that 1st gen units use shared attenuation network on treble and midds. 2nd gen units do have separate attenuators for trebele and midds. This should not make 2nd gen units easier to drive.

So... did Quad 405 Mk2 cure this drive problem, or was there some mod on Cantata crossover which made it easier to drive, that I do not understand. Modified inductors or something like that?

2nd reason I have considered for cause of hardness is ageing of drive units. I have tried 2 pairs of T52 but no real improvement. I have second pair of B139, but haven´t tried them yet, as it seems so unlikely cause for slight hardness. So... is there any place in crossover to be slightly adjusted, if say T52 ageing has made upper crossover region imperfect.

In my opinion Cantata/CJ combination sounded like the amp was not 100% stable. The amp has earlier performed very well with other speakers. However it would be nice if Premier could be used with Cantatas as Premiers are very nice amps, especially if you like honest soundstage.

Any information would be most welcome.

[ColinR]

[proffski]

When I started work I lusted for a Garrard 401, damned good looking, far better on the eye than the 301.
My yearning for a 401 died when I heard my first 401 with a 12" SME arm, I forget which cartridge was being used.
My second listening session was in Swindon on a wide bandwidth system. Sadly I called them "Rumblemasters" from then on... that is when I bought my first Thorens, ah… silence!
_________________
I contend that for a nation to try to tax itself into prosperity is like a
man standing in a bucket and trying to lift himself up by the handle.
-Winston Churchill

[iso]

Thank you for your comments.

However I haven´t felt that low bass was flabby, fundamentals should be only slightly more hearable. Maybe level of low bass is slightly low.
I don´t think that low advertised D/F: s are any problem with Cantatas, as B139:s relay on mostly mechanical damping in Cantatas.

If you calculate reactance and DCR components of crossover components at say 38Hz and 200 Hz, you will note that at B139 sees roughly 4,6 ohm source from amp with 0 ohm source impedance, source impedance is 5,6ohm from amp with 1 ohm output. At 200 Hz situation is even worse. In my opinion, this clearly demonstrates that amp output impedance is not too important issue in this situation.

If we rate B139 at 8ohm D/F:s are 1,7 or 1,4 with 0 ohm amp output or 1,0 ohm amp output. Situation is actually better as impedance is considerably more at 38 Hz due resonance max at 38 Hz. In vented box situation might have been considerably worse.

Actually I believe interaction of amp and Cantatas might have had some undesirable effects via feedback loop. Even CJ has only 12dB long feedback and EAR 509 even less long feedback, as most its feedback is within output- and driver stage.

It would also be interesting to bolt my grease bearing 301 to plinth with SME V. I have´t yet sorted out plinth design, but 301 is refurbished and SME is ready to go. This is my first respected idler drive turntable since not so good BSR128. I will report what kind of rumble generator I will have.

Best Regards