Pochago - The whole point about ESR in electrolytics is that it is NOT a single value so can't be simulated by a fixed resistor. The amount of ESR varies with Current, Frequancy and even temperature. Also the disipation in an electrolytic acts as a damping factor. Even the original Elcaps had a slightly resonant sound, but having the properties mentioned above and being slightly less transparent than a Polyprops meant it was not as noticeable. Stick a good quality PP in, either with or without ESR simulating resistors and it sounds awfull.
This mainly effects any Mid range substitutions. You can change the Tweeter caps to PP no problem as they were film caps in the first place.
I have isolated the three elements T/M/B in the cross over and got it mounted externally and tri-wired. The resonance is generated in the mid range band pass filter but at a freqancy at around the x-over point to the tweeter. A small amount of resonance is heard through the mid driver but most of it comes through the tweeter as the elements are still electrically connected, all be it through about 4 meters of cable back and forth to the amp outlets. If I disconect the signal to the mid filter the tweeter is as sweet as you could ever want. Listen to just the Mid driver and it is fairly cleen with only the very smallest amount of audible resonance. As soon as the Mid and tweeter filters operate together the tweeter resonates.
I have tried many makes of polyprop in the mid range and many values. I have also tried various electrolytics, but none seem to have the same ESR properties as the original Elcaps, even the Alcaps which are supposed to be similar don't work for the main series value in the mids.
I have not been able to find any combination that I could happily live with, or that is even as good as my old original Elcaps. I regularly pop them back in to get a reference point, all major parts are on terminal blocks so changing them takes a couple of minutes max on the external x-over.
People have suggested that the close proximity and orientation of the coils may be leading to interaction that could be causing the resonance. To this end I am in the process of making a much bigger board. 320mm x 260mm which is the largest that will fit through the driver aperture. This will be mounted behind the ABR unit to get it away from the big bass driver magnet which is normaly mounted directly over the coils. I will space the coils appart and not have their axis crossing other coils, to eliminate interaction. It may have an effect but I think it is all down to the generic band pass filter where a coil and cap are in series. Everything I have read, and I repeat I am no expert, suggests this will cause electrical resonance.
My next move is to tri-amp, the drivers each getting their own channel and isolating the mid driver and tweeter circuits compleatly. I think this might do the trick, if not I am going active DSP and doing away with the passive x-over completely. I've spent a massive amount and months of work trying to get the passive to work, but I'm not giving up on the 66's as I like the overall presentation so much.
I will let you know how it pans out.
This mainly effects any Mid range substitutions. You can change the Tweeter caps to PP no problem as they were film caps in the first place.
I have isolated the three elements T/M/B in the cross over and got it mounted externally and tri-wired. The resonance is generated in the mid range band pass filter but at a freqancy at around the x-over point to the tweeter. A small amount of resonance is heard through the mid driver but most of it comes through the tweeter as the elements are still electrically connected, all be it through about 4 meters of cable back and forth to the amp outlets. If I disconect the signal to the mid filter the tweeter is as sweet as you could ever want. Listen to just the Mid driver and it is fairly cleen with only the very smallest amount of audible resonance. As soon as the Mid and tweeter filters operate together the tweeter resonates.
I have tried many makes of polyprop in the mid range and many values. I have also tried various electrolytics, but none seem to have the same ESR properties as the original Elcaps, even the Alcaps which are supposed to be similar don't work for the main series value in the mids.
I have not been able to find any combination that I could happily live with, or that is even as good as my old original Elcaps. I regularly pop them back in to get a reference point, all major parts are on terminal blocks so changing them takes a couple of minutes max on the external x-over.
People have suggested that the close proximity and orientation of the coils may be leading to interaction that could be causing the resonance. To this end I am in the process of making a much bigger board. 320mm x 260mm which is the largest that will fit through the driver aperture. This will be mounted behind the ABR unit to get it away from the big bass driver magnet which is normaly mounted directly over the coils. I will space the coils appart and not have their axis crossing other coils, to eliminate interaction. It may have an effect but I think it is all down to the generic band pass filter where a coil and cap are in series. Everything I have read, and I repeat I am no expert, suggests this will cause electrical resonance.
My next move is to tri-amp, the drivers each getting their own channel and isolating the mid driver and tweeter circuits compleatly. I think this might do the trick, if not I am going active DSP and doing away with the passive x-over completely. I've spent a massive amount and months of work trying to get the passive to work, but I'm not giving up on the 66's as I like the overall presentation so much.
I will let you know how it pans out.
I was trying to help a friend design a speaker recently. All I could think of was going active to side step the problems. Every time we solved one problem we created two new ones ( no joke, this thread says the same as I found ). This was only two drive units. What I like even more is a very simple passive crosssover that also protects the tweeter against DC. Then the bumps done actively. This means the amplifier can be made into a filter without any extra op amps. Have done it before using Texas Filter Pro freeware. Do the op amp version first as it is safer. I have my Quad 303 configured to drive the tweeter via a film cap rather than it's big electrolytic. This mildly upgrades the tweeter circuit. The Quad sits at 33.5 VDC at the other side.
A friend finds his Tannoy Lancasters dull. The graphs look weird in test reviews. We are going to bi wire them first and try bi amping. The likely outcome is that the tweeters are defective. Any thoughts? One reviewer had them working at 40 kHz. The proof being his utrasonic remote control function on something could be activated by the Tannoys with a signal generator.
I would love to try the Tannoys OB. I have some 12LTa to bolt into his cabs. Should be interesting. Poormans Lockwoods and poormans Tannoy if 12 Lta in a Lancaster cab. I think he has 12 inch?
A friend finds his Tannoy Lancasters dull. The graphs look weird in test reviews. We are going to bi wire them first and try bi amping. The likely outcome is that the tweeters are defective. Any thoughts? One reviewer had them working at 40 kHz. The proof being his utrasonic remote control function on something could be activated by the Tannoys with a signal generator.
I would love to try the Tannoys OB. I have some 12LTa to bolt into his cabs. Should be interesting. Poormans Lockwoods and poormans Tannoy if 12 Lta in a Lancaster cab. I think he has 12 inch?
You lost me a bit in your post Nigel.
Your friends speakers were two way? If so a passive crossover is much easier to implement for a two way than a three way. It's the mid range band pass that is difficult to get right, thats why so many companies produce two way floor standers, when they have the space for them to be three way.
I have looked at the ESP P09 analogue active project but I like the flexibility on the miniDSP 4x10hd to add a properly integrated sub woofer at a later stage. Also I am using the seas tweeter which would allow me to try dropping the HF crossover point from 5kHz to 4.5kHz or even 4kHz to help out the mid driver, which is pretty much on its stops, both ends of its range. Unless you have the early MF500 which is flat to about 7kHz, B&O cross this unit over at 6kHz in their implimentations. The MD500 barely makes the 5kHz cross over, but has a higher power rating, so it's swings and round abouts.
Siegfried Linkwitz has auditioned a version of his flagship LX521 active speaker, prepared by an enthusiast with a miniDSP x-over and he was so impressed he is recomending it on his site as an alternative to his own analogue active circuit. That speaks volumes to me, and was what got my interest in DSP.
Your friends speakers were two way? If so a passive crossover is much easier to implement for a two way than a three way. It's the mid range band pass that is difficult to get right, thats why so many companies produce two way floor standers, when they have the space for them to be three way.
I have looked at the ESP P09 analogue active project but I like the flexibility on the miniDSP 4x10hd to add a properly integrated sub woofer at a later stage. Also I am using the seas tweeter which would allow me to try dropping the HF crossover point from 5kHz to 4.5kHz or even 4kHz to help out the mid driver, which is pretty much on its stops, both ends of its range. Unless you have the early MF500 which is flat to about 7kHz, B&O cross this unit over at 6kHz in their implimentations. The MD500 barely makes the 5kHz cross over, but has a higher power rating, so it's swings and round abouts.
Siegfried Linkwitz has auditioned a version of his flagship LX521 active speaker, prepared by an enthusiast with a miniDSP x-over and he was so impressed he is recomending it on his site as an alternative to his own analogue active circuit. That speaks volumes to me, and was what got my interest in DSP.
Its amaizing to see that this treat is still alive but just think yourself why its lasting so long, like never ending story ... ?
my opinion on that is very simple as once you start changing original "settings" you may improve one thing but at the same time unfortunately something else is going to be wrong etc etc.
I have been there. Had 3 pairs of ditton 66 in total, tried lots of different combination ,seas tweeter, OW1, ribbons... different type of caps in different variations.
To my surprise best results and "closest to original"? hard to judge on that as I have never heard dittons when they were new... so really dont know for sure how they should sound... but the fuller sound I got by, using only oryginal drivers especially Tweeters and the same value caps. The magical thing about D66 is midrange which has this fantastic ambiance voices hanging in air and there is lot of air... that's the midrange I was looking for long time and believe me I have tried many caps/types/values/combinations and never could get it right, to my surprise trick was to use ELECTROLITYC CAP! but at very low voltage - I used 25uf from Visaton 35V (10+15).
caps for tweeter may just depends on you budget as you can chose whatever you want i found again Visaton foil were better than polys and good enough, for bass I used ALCAPs. Something was telling me that the whole thing with compensating ESR may just make thing worse so I didnt go that way, and I am glad, hard to believe that low cost xover may sound better than expensive one... regarding the bass - I found that much better control you will get by using much more powered amp. tried Audiolab 8000a + power amp - worked very well / different story is to use valves those speakers loves KT88
Funny thing is that D66 really sounded great but I sold them anyway, as I wanted to try OB , bought Tang band w8 1808 + eminence 15a and will have a go. TB is amazing driver - never heard such a speed details and clarity but if project will not work out to my needs will be coming to d66 as they were special.
mixMZ:
I am certain that it is an endless road when trying to improve the sound of any driver, speaker or system since nothing can be perfect.
I have only used ESA Claritycap to recap the tweeter and mid section. I used Alcap for bass due to size and cost. I thought that using film cap rather than electrolytic should provide more improvement to the sound. So may be I will get some electrolytic to try with the mid-range. I wonder what brand of cap you have tried? It would be nice if you can list out those you feel that doesn't work well so we can avoid them. Have you tried Alcap for the mid?
For your D66, did you have the different version? I wonder if there is any different in sound between versions. If so which one you prefer?
I looking into building an OB with a full range driver so there is no crossover to due with. Still in the process of research which driver would be a good start for testing.
Quin. Being very searching in my ways of doing things I find a two way speakers difficult. Unlike active where my amplifer is directly coupled passive adds complication. It is like the parts of a motorcar that must move and yet must not move. For that we have a ball joint. In passive speakers things are not that well served. If you like they flap about. Three way takes the problem almost an impossible step forward where the best can never be had as it can never be known. My heart goes out to any who try.
The 12Lta driver is mostly a one way. The choice of how it works it taken away. When OB that almost means only one design problem. Bass EQ that purists insist on doing passively. My experiments with the two way design would cause me to run a mile from that. A 12 Lta on a 4 x 2 foot baffle using the standard EQ of an amp like NAD 3020 ( full bass boost ) is just about right. There is just about enough power to do that. The sound is very like the speakers that cost the same as small cars. Thew design problem limited to just one big one. A supertweeter could be added at 7 kHz perhaps. Personally I like the 12 kHz rolll off, it doesn't sound less detailed as there is no conventional band limiting of capacitors etc in the way.
To add complication to the story we now can have digital filters with zero phase shift. Thus when active we almost approach the ideal of my 12 Lta. This was never possible before. I am not saying the 12 Lta is perfect. Just to say it has been taken out of my hands. It is what it is.
The great value of this thread is I will use a capacitor when designing my active tweeter amp. I will take the choice very serriously. At least it will be one only. I will compare with direct and see if it is OK. I will also put it in the feedback loop to see it that is better. It will most likely be a OPA604 op amp with class A complimentary feedback pair output. The standing current about 400 mA. That circuit measures vey well without loop feedback to either the dumpers or capacitor if wanted. If I want 2 poles for the tweeter simple input filtering to the op amp ( C in ) will do fine as will C out if not inside the feedback loop . How easy is that ? I would not use true active as it has a chance of being subtly unstable if part of the amp. I would say if only the tweeter this would be a good solution here. If the amp is OPA 604 it can have a very high input impedance without noise penalty. Thus if piggy-backed on the preamp it won't load it down and will enjoy the output impedance of the preamp as a noise shunt in as much as it matters.
A word of caution. I don't like most active speakers. I suspect this simply comes down to the fact they were designed for passive use. This is why semi active is an idea. Perhaps the series inductance or simple resistance helps the damping required. Not too much nor too little. One would guess 1R should change the sound should it be required. My friend who designed valve amplifiers thought a 1 R resistor would be much of why valves sound as they do. Also 1R resolves the inductance of the drive unit to be more resistive which the amplifer will like. Like a motorcar suspension there will be a critical damping point. It might be 2R2, who knows?
One thing simple active brings is the reality of the passive design being tested. If you hear a great sound when active that never comes when passive fair chance it couldn't. The reason being the loose connections and energy sharing between drive units at the crossover points ( e.g. bass unit feeds mid with delayed energy unbraced by the amplifier ouitput impedance ). I love the BBC auto transformer. It removes one loose connection. What genius BBC.
The 12Lta driver is mostly a one way. The choice of how it works it taken away. When OB that almost means only one design problem. Bass EQ that purists insist on doing passively. My experiments with the two way design would cause me to run a mile from that. A 12 Lta on a 4 x 2 foot baffle using the standard EQ of an amp like NAD 3020 ( full bass boost ) is just about right. There is just about enough power to do that. The sound is very like the speakers that cost the same as small cars. Thew design problem limited to just one big one. A supertweeter could be added at 7 kHz perhaps. Personally I like the 12 kHz rolll off, it doesn't sound less detailed as there is no conventional band limiting of capacitors etc in the way.
To add complication to the story we now can have digital filters with zero phase shift. Thus when active we almost approach the ideal of my 12 Lta. This was never possible before. I am not saying the 12 Lta is perfect. Just to say it has been taken out of my hands. It is what it is.
The great value of this thread is I will use a capacitor when designing my active tweeter amp. I will take the choice very serriously. At least it will be one only. I will compare with direct and see if it is OK. I will also put it in the feedback loop to see it that is better. It will most likely be a OPA604 op amp with class A complimentary feedback pair output. The standing current about 400 mA. That circuit measures vey well without loop feedback to either the dumpers or capacitor if wanted. If I want 2 poles for the tweeter simple input filtering to the op amp ( C in ) will do fine as will C out if not inside the feedback loop . How easy is that ? I would not use true active as it has a chance of being subtly unstable if part of the amp. I would say if only the tweeter this would be a good solution here. If the amp is OPA 604 it can have a very high input impedance without noise penalty. Thus if piggy-backed on the preamp it won't load it down and will enjoy the output impedance of the preamp as a noise shunt in as much as it matters.
A word of caution. I don't like most active speakers. I suspect this simply comes down to the fact they were designed for passive use. This is why semi active is an idea. Perhaps the series inductance or simple resistance helps the damping required. Not too much nor too little. One would guess 1R should change the sound should it be required. My friend who designed valve amplifiers thought a 1 R resistor would be much of why valves sound as they do. Also 1R resolves the inductance of the drive unit to be more resistive which the amplifer will like. Like a motorcar suspension there will be a critical damping point. It might be 2R2, who knows?
One thing simple active brings is the reality of the passive design being tested. If you hear a great sound when active that never comes when passive fair chance it couldn't. The reason being the loose connections and energy sharing between drive units at the crossover points ( e.g. bass unit feeds mid with delayed energy unbraced by the amplifier ouitput impedance ). I love the BBC auto transformer. It removes one loose connection. What genius BBC.
pochago
elcaps 72 - replaced with standard 50v Alcaps 72uf
elcaps 6 - replaced with LL alcap 6uf
for mid 24uf or 30 uf depends on version and drivers I always used - Visaton 25 uf (part nr: 5364 + 5366)
for tweeters: 4uf (4x visaton 5319 1.0 uf) and for 6uf (4x visaton 5321 1.5 uf) you can try adding very small cap (very good quality ones) at much higher voltage sometimes it works great as well for example foil 3.8uf cap and add to it 0.3uf form Clarity Cap 630VDC or Jantzen Audio Superior Z-Cap etc it may do a trick as well. Use some good quality binding Posts and try to use the same cable in xover as you use from speaker to amp - I used qed anniversary with very good results. Thats about it really... you can also increese damping material inside to tighten bass a bit ,as additional I used some loose layer of raw sheep wool on top of the standard foam ,you can try use additional Foam Pyramid Tiles type - can reccomand AFP75. Everything more than that i think is just a case of 'form over content` or you have to got really and I mean really good quality gears to justify the whole outcome.
good luck
elcaps 72 - replaced with standard 50v Alcaps 72uf
elcaps 6 - replaced with LL alcap 6uf
for mid 24uf or 30 uf depends on version and drivers I always used - Visaton 25 uf (part nr: 5364 + 5366)
for tweeters: 4uf (4x visaton 5319 1.0 uf) and for 6uf (4x visaton 5321 1.5 uf) you can try adding very small cap (very good quality ones) at much higher voltage sometimes it works great as well for example foil 3.8uf cap and add to it 0.3uf form Clarity Cap 630VDC or Jantzen Audio Superior Z-Cap etc it may do a trick as well. Use some good quality binding Posts and try to use the same cable in xover as you use from speaker to amp - I used qed anniversary with very good results. Thats about it really... you can also increese damping material inside to tighten bass a bit ,as additional I used some loose layer of raw sheep wool on top of the standard foam ,you can try use additional Foam Pyramid Tiles type - can reccomand AFP75. Everything more than that i think is just a case of 'form over content` or you have to got really and I mean really good quality gears to justify the whole outcome.
good luck
Last edited:
Well, I've made some progress on the 66s.
First of all, I took out the ESR compensating resistors. Wow, what a difference this made - it went from sounding muffled to sounding much clearer. However, I still am not satisfied with the sound. The midrange has this nasal quality to it, meaning I think the midrange is too bright - something is being emphasized somewhere, perhaps this is the distortion/resonance at play?
So, armed with the original crossovers, I decided to take one of the original 24uF Elcaps and put it in place of the polypropylene capacitor in one of the 66s. In comparing to the other one, I notice that the midrange is a bit more subdued, not muffled, but that nasal quality is pretty much gone. Surprisingly, the old cap seems to work better than the new Solens. I'm going to do the other one and listen to the speakers in stereo and report back.
First of all, I took out the ESR compensating resistors. Wow, what a difference this made - it went from sounding muffled to sounding much clearer. However, I still am not satisfied with the sound. The midrange has this nasal quality to it, meaning I think the midrange is too bright - something is being emphasized somewhere, perhaps this is the distortion/resonance at play?
So, armed with the original crossovers, I decided to take one of the original 24uF Elcaps and put it in place of the polypropylene capacitor in one of the 66s. In comparing to the other one, I notice that the midrange is a bit more subdued, not muffled, but that nasal quality is pretty much gone. Surprisingly, the old cap seems to work better than the new Solens. I'm going to do the other one and listen to the speakers in stereo and report back.
All right, both speakers now have the original 24uF Elcap capacitors in place of the Solens.
The midrange is definitely smoother and is more compelling. I detect more richness as well. However, I still can't help but feel like it just isn't there yet. It seems to be a bit too subdued.. it's close to sounding phenomonal, but it just isn't there. I will say though, the speakers are much more enjoyable like this than they were with the Solens. At least, I'm not feeling the need to stop listening to them from hearing something I really dislike.
I think I am going to order up Alcaps for the midrange filter and call it a day. Surprisingly, with the HF filter section run the way it is (30ohm resistors with the tweeters, using the 0.14mH with modified capacitance values) it sounds good. It doesn't seem to be too low in output and I am hearing good integration. Maybe it could be better with the correct coil and cap values.
At the least, this is some progress.
Edit: I think the caps might have been sitting for a long time - the more I am listening the more I hear the midrange opening up. I could also be getting used to the sound.. but it's sounding great. This is the best they've sounded since I started working on these.
The midrange is definitely smoother and is more compelling. I detect more richness as well. However, I still can't help but feel like it just isn't there yet. It seems to be a bit too subdued.. it's close to sounding phenomonal, but it just isn't there. I will say though, the speakers are much more enjoyable like this than they were with the Solens. At least, I'm not feeling the need to stop listening to them from hearing something I really dislike.
I think I am going to order up Alcaps for the midrange filter and call it a day. Surprisingly, with the HF filter section run the way it is (30ohm resistors with the tweeters, using the 0.14mH with modified capacitance values) it sounds good. It doesn't seem to be too low in output and I am hearing good integration. Maybe it could be better with the correct coil and cap values.
At the least, this is some progress.
Edit: I think the caps might have been sitting for a long time - the more I am listening the more I hear the midrange opening up. I could also be getting used to the sound.. but it's sounding great. This is the best they've sounded since I started working on these.
Last edited:
Jeffrey88 - You are going the same route as I did and trying all the same combinations. I spent a fortune on different make and value of caps and resistors, but I was never satisfied with the results.
You are wasting your time and money.
The "nazel" quality you heard in the upper mids is the resonance I have been talking about. You chase it around and by changing cap values move its frequency, but that is all. As its around the x-over frequency, the resonance can even be shifted up into the area covered by the tweeter and as the circuits are connected and interact the distortion comes out of that driver, even though created in the mid range band pass circuit.
The cleanest I have been able to achieve, was doing as mixMZ suggested and use alcaps for the bass circuit, though I found the 100v worked very slightly better than the 50v here. For the mid range parallel cap again use Alcap but this time use the 50v (not LL) as this worked slightly better re, the resonance. For the Mid series cap use a single Polyprop without any resistors, but add a fixed value L-Pad attenuation at the end of the circuit (before the driver) of approx -0.6dB. Put a 0R5 resistor in Series then a 100R accross the outputs to the driver (Parallel) to achieve this. The tweeter just needs a pair of polyprops of your choice, without any resistors. I did find that using ClarityCap ESA in these positions reduces resonance but can sound a bit reined in so using an Ansar Supersound or a Sonicap Gen1 for the smaller value brings some attack and sharpness back without going too far.
That is the best I have managed after hundreds of hours of fiddling, I am still not happy but it is better than the original circuit components. I hope this information will avoid you spending unnecessarily, but if you wan't to satisfy your own curiosity, I understand.
I am now going to split the x-over elements (HF/Mid/Bass) into seperate cicuits each driven by its own amplifier channel (Tri-Amping) and move the coils further appart and better oriented to avoid interaction, which in theory could contribute to the distortion, but I am not convinced on that one. If all else fails I will ditch the x-over all together and use the 6 channels of amplification used for Tri-Amping and go with an active x-over feeding the amps.
If you do hit on a magic combination, please let us all know, I just love the overall presentation of these speakers, more than anything I have heard at sensible money, so removing these niggling little faults would make them my keepers.
You are wasting your time and money.
The "nazel" quality you heard in the upper mids is the resonance I have been talking about. You chase it around and by changing cap values move its frequency, but that is all. As its around the x-over frequency, the resonance can even be shifted up into the area covered by the tweeter and as the circuits are connected and interact the distortion comes out of that driver, even though created in the mid range band pass circuit.
The cleanest I have been able to achieve, was doing as mixMZ suggested and use alcaps for the bass circuit, though I found the 100v worked very slightly better than the 50v here. For the mid range parallel cap again use Alcap but this time use the 50v (not LL) as this worked slightly better re, the resonance. For the Mid series cap use a single Polyprop without any resistors, but add a fixed value L-Pad attenuation at the end of the circuit (before the driver) of approx -0.6dB. Put a 0R5 resistor in Series then a 100R accross the outputs to the driver (Parallel) to achieve this. The tweeter just needs a pair of polyprops of your choice, without any resistors. I did find that using ClarityCap ESA in these positions reduces resonance but can sound a bit reined in so using an Ansar Supersound or a Sonicap Gen1 for the smaller value brings some attack and sharpness back without going too far.
That is the best I have managed after hundreds of hours of fiddling, I am still not happy but it is better than the original circuit components. I hope this information will avoid you spending unnecessarily, but if you wan't to satisfy your own curiosity, I understand.
I am now going to split the x-over elements (HF/Mid/Bass) into seperate cicuits each driven by its own amplifier channel (Tri-Amping) and move the coils further appart and better oriented to avoid interaction, which in theory could contribute to the distortion, but I am not convinced on that one. If all else fails I will ditch the x-over all together and use the 6 channels of amplification used for Tri-Amping and go with an active x-over feeding the amps.
If you do hit on a magic combination, please let us all know, I just love the overall presentation of these speakers, more than anything I have heard at sensible money, so removing these niggling little faults would make them my keepers.
Thanks for the info Qwin.
The reason I am doing this is mostly educational purposes. I was especially surprised the Elcaps did indeed sound better. My experience with recaps on speakers with elcaps is that when I replace them with Alcaps there is always a big improvement in sound. Usually they did not sound good beforehand. But this confirms what you are saying regarding the ESR and resonance issue.
I plan to follow what you have advised on which will be redoing the treble filter and using an Alcap on the midrange filter and some better polypropylene caps for the rest.
My crossovers are currently separated on to 3 boards by the way, with much more distance between the coils. In recalling my memory from hearing an all original pair, it reduces the coloration and distortion. It's a good improvement. I won't be tri amping (although that is a possibility) however.
I think my next step will be to find a proper board to put the crossover on and mount it in the cabinet. Mine are in the same place you decided on, behind the ABR.
I'll certainly post my results when I reach the point of putting the Alcaps in. Thanks again Qwin.
Sent from my iPhone using Tapatalk
The reason I am doing this is mostly educational purposes. I was especially surprised the Elcaps did indeed sound better. My experience with recaps on speakers with elcaps is that when I replace them with Alcaps there is always a big improvement in sound. Usually they did not sound good beforehand. But this confirms what you are saying regarding the ESR and resonance issue.
I plan to follow what you have advised on which will be redoing the treble filter and using an Alcap on the midrange filter and some better polypropylene caps for the rest.
My crossovers are currently separated on to 3 boards by the way, with much more distance between the coils. In recalling my memory from hearing an all original pair, it reduces the coloration and distortion. It's a good improvement. I won't be tri amping (although that is a possibility) however.
I think my next step will be to find a proper board to put the crossover on and mount it in the cabinet. Mine are in the same place you decided on, behind the ABR.
I'll certainly post my results when I reach the point of putting the Alcaps in. Thanks again Qwin.
Sent from my iPhone using Tapatalk
Also, I've been reading this thread with some interest..
Celestion Ditton 44 - HF 2000 tweeter replacement? - Page 3 - UK Vintage Radio Repair and Restoration Discussion Forum
It seems that he used a 10ohm resistor in parallel with the tweeter, and then added a 1.1ohm resistor in series. The idea was to achieve the same resistance of the HF2000. Although it's a Ditton 44, he seems to be very pleased with that..
Any thoughts? I think once the midrange filter is set back to Alcaps, the next step is to finish off the treble filter because that has to be the last thing to get them sounding proper. I'm tempted to try the above.
Celestion Ditton 44 - HF 2000 tweeter replacement? - Page 3 - UK Vintage Radio Repair and Restoration Discussion Forum
It seems that he used a 10ohm resistor in parallel with the tweeter, and then added a 1.1ohm resistor in series. The idea was to achieve the same resistance of the HF2000. Although it's a Ditton 44, he seems to be very pleased with that..
Any thoughts? I think once the midrange filter is set back to Alcaps, the next step is to finish off the treble filter because that has to be the last thing to get them sounding proper. I'm tempted to try the above.
Jeffrey88 - The Seas tweeter he uses (19TFF1) is the same one that I have used.
I did a frequency sweep on both mine using DATS and the lowest impedence on the plot was 7 Ohm which just happened to be at the 5kHz crossover point. He was working on 7.5 Ohm but I'm not sure how he got to that figure, in any case it's not that far out. Trouble is I can not see where he arrives at the S= 1.1 Ohm, P=10 Ohm for his L-Pad. The on line calculator I use won't get any where near that combination no matter what dB reduction value you enter.
To adopt the Seas Tweeter I changed the filter values, you need to do this as the driver is a different impedance, 7 not 4 Ohm.
To do this I changed the inductor from 0.14mH to 0.17mH.
The large cap from 6uf to 9.1uf.
The small cap from 4uf to 3.3uf. This calculated as 3.0uf but after trying both I thought the 3.3uf sounded better. Calculators only get you in the right ball park.
I have -0.6dB attenuation L-Pad of S=0.5 Ohm, P=100 Ohm based on 7.0 Ohm impedance at crossover point.
Note - for your Mid range you can change the big series cap to PP, I used an Ansar Supersound 30uf, but needed an L-Pad before the driver as withought ESR or Resistors attached, it is too forward. I used 0.62dB of attenuation, which also worked out at S=0.5 Ohm, P=100 Ohm for the measured 7.5 Ohm impedance drivers. The L-Pad does not excite the resonance as much as adding ESR resistors to the big cap. Note - I have both the MF and MD mid range drivers and prefere the earlier MF. I think it is more subtle and realistic with acoustic instruments and the MF500 is bigger and bolder, more suited to rock etc. If using the MD version you may have to use more attenuation as you get an excess of upper mid/lower treble in the mix, even with the matching 24uf cap.
These are all compromises, but there do not appear to be any audible holes or peaks in the spectrum, I listen to recordings of concert grand piano for this and pieces that run up and down the scales. It sounds right with good timming and seperation of notes.
I did a frequency sweep on both mine using DATS and the lowest impedence on the plot was 7 Ohm which just happened to be at the 5kHz crossover point. He was working on 7.5 Ohm but I'm not sure how he got to that figure, in any case it's not that far out. Trouble is I can not see where he arrives at the S= 1.1 Ohm, P=10 Ohm for his L-Pad. The on line calculator I use won't get any where near that combination no matter what dB reduction value you enter.
To adopt the Seas Tweeter I changed the filter values, you need to do this as the driver is a different impedance, 7 not 4 Ohm.
To do this I changed the inductor from 0.14mH to 0.17mH.
The large cap from 6uf to 9.1uf.
The small cap from 4uf to 3.3uf. This calculated as 3.0uf but after trying both I thought the 3.3uf sounded better. Calculators only get you in the right ball park.
I have -0.6dB attenuation L-Pad of S=0.5 Ohm, P=100 Ohm based on 7.0 Ohm impedance at crossover point.
Note - for your Mid range you can change the big series cap to PP, I used an Ansar Supersound 30uf, but needed an L-Pad before the driver as withought ESR or Resistors attached, it is too forward. I used 0.62dB of attenuation, which also worked out at S=0.5 Ohm, P=100 Ohm for the measured 7.5 Ohm impedance drivers. The L-Pad does not excite the resonance as much as adding ESR resistors to the big cap. Note - I have both the MF and MD mid range drivers and prefere the earlier MF. I think it is more subtle and realistic with acoustic instruments and the MF500 is bigger and bolder, more suited to rock etc. If using the MD version you may have to use more attenuation as you get an excess of upper mid/lower treble in the mix, even with the matching 24uf cap.
These are all compromises, but there do not appear to be any audible holes or peaks in the spectrum, I listen to recordings of concert grand piano for this and pieces that run up and down the scales. It sounds right with good timming and seperation of notes.
Last edited:
Hey Qwin,
I have to admit I'm a bit confused by your suggestion to use a polypropylene in the midrange filter. I thought through our discussion it was best advised to stick with an electrolytic with similar ESR properties to the Elcaps. It seems to me that adding in resistors and L-pads is simply complicating the crossover and potentially the balance of the speaker unnecessarily - I get that PPs are better capacitors (technically speaking) but considering the original crossover is designed with electrolytics, would it not be smarter to simply go with that?
As for the 10ohm + 1.1ohm resistor, at his estimation of 7.5ohms (which is what the spec sheet indicates on the Seas website) at 5kHz, adding 10ohms in parallel reduces the resistance down to 4.2ohms, then it adds 1.1ohms to that for an estimated 5.3ohms (why he wouldn't just employ more resistance in parallel, I'm not sure).
I've determined that in my treble filter, with the cap values altered and the 30ohms resistance, the tweeter is seeing a (roughly) 5kHz crossover point as per calculations. Some fine tuning I am sure would improve it but it should be close. Besides removing the resistors, I can't think of a reason to change the treble filter - would it really sound better with different values on each part?
My next step will be to build much tidier crossovers, then replace the Elcap with an Alcap. After this, save for some final tweaking, they should be sounding very good. Right now they are sounding great, I am pretty happy with them but still detect a veil in the midrange.
I have to admit I'm a bit confused by your suggestion to use a polypropylene in the midrange filter. I thought through our discussion it was best advised to stick with an electrolytic with similar ESR properties to the Elcaps. It seems to me that adding in resistors and L-pads is simply complicating the crossover and potentially the balance of the speaker unnecessarily - I get that PPs are better capacitors (technically speaking) but considering the original crossover is designed with electrolytics, would it not be smarter to simply go with that?
As for the 10ohm + 1.1ohm resistor, at his estimation of 7.5ohms (which is what the spec sheet indicates on the Seas website) at 5kHz, adding 10ohms in parallel reduces the resistance down to 4.2ohms, then it adds 1.1ohms to that for an estimated 5.3ohms (why he wouldn't just employ more resistance in parallel, I'm not sure).
I've determined that in my treble filter, with the cap values altered and the 30ohms resistance, the tweeter is seeing a (roughly) 5kHz crossover point as per calculations. Some fine tuning I am sure would improve it but it should be close. Besides removing the resistors, I can't think of a reason to change the treble filter - would it really sound better with different values on each part?
My next step will be to build much tidier crossovers, then replace the Elcap with an Alcap. After this, save for some final tweaking, they should be sounding very good. Right now they are sounding great, I am pretty happy with them but still detect a veil in the midrange.
Hi Jeffrey
After many back and forth comparisons I think that it is safe to use a PP for the series cap in the mid range filter, but not for the parallel 4uf one.
If you use it without any ESR resistors it sounds much more transparent than an Alcap or Mundorf Lytic. The resonance is no worse with the bare PP than with the lytics in this location, remember this particular lytic is LL which has less ESR. It is however rather forward as I pointed out and adding the attenuation still leaves the mid range sounding more transparent and with much more attack on tom toms etc. than with a lytic. The reason I had shyed away from it previous was the forwardness, this is why I tried an L-Pad, which seems to work.
The values on Seas web site are nominal for production targets. I was told by a designer to measure the drivers and mine are 7.05 and 7.2 Ohm at 5kHz so this is what I used.
The HF filter is designed for a 4 Ohm driver, different values are required for a 7/7.5 Ohm. Try using an online tool and see how it varies when you change the driver impedance. The cap values and inductor values change if you want to keep the same crossover frequency. I briefely tried the Seas with the stock values while waiting for new caps to arrive, it sounded reasonable but improved quite a lot when I changed them and the inductor.
After many back and forth comparisons I think that it is safe to use a PP for the series cap in the mid range filter, but not for the parallel 4uf one.
If you use it without any ESR resistors it sounds much more transparent than an Alcap or Mundorf Lytic. The resonance is no worse with the bare PP than with the lytics in this location, remember this particular lytic is LL which has less ESR. It is however rather forward as I pointed out and adding the attenuation still leaves the mid range sounding more transparent and with much more attack on tom toms etc. than with a lytic. The reason I had shyed away from it previous was the forwardness, this is why I tried an L-Pad, which seems to work.
The values on Seas web site are nominal for production targets. I was told by a designer to measure the drivers and mine are 7.05 and 7.2 Ohm at 5kHz so this is what I used.
The HF filter is designed for a 4 Ohm driver, different values are required for a 7/7.5 Ohm. Try using an online tool and see how it varies when you change the driver impedance. The cap values and inductor values change if you want to keep the same crossover frequency. I briefely tried the Seas with the stock values while waiting for new caps to arrive, it sounded reasonable but improved quite a lot when I changed them and the inductor.
OK, I see.
Well, after my tests using the Solen polypropylene capacitor in place of the Elcap 24uF, I think I am going to stick with electrolytic. I am at a point where I am tired of messing around with these and just want to sit down and enjoy the music. If I feel up to it I might try the L-pad idea later on but, it sounds like more experimenting to me and I'm not up for it. I also don't want to keep spending money on these (as it is, to date I've spent around $1100 in total on them and that is far more than they are worth).
Anyway, when I get around to cleaning up the crossovers and installing the new electrolytics I'll post some pictures and impressions.
Thanks for your help Qwin.
Jeffrey
OK - a little more experimenting.
I decided to pop in the 4uF Elcap as well on the midrange filter, just to see. Previously I had a 4.4uF mylar film capacitor there.
In comparing the two, it's hard to say one sounds better than the other, but I do detect a bit more of a balance with the speaker with the Elcap. It seems like there's a slight suck-out on the speaker with the mylar film cap still in place. Honestly though, the difference is fairly subtle, it took quite a bit of back and forth to really hear any amount of difference between the two. I think a fresh electrolytic will make a good improvement, as I am liking the way the old Elcaps sound over what I had before, and recalling what I heard when I replaced Elcaps, I am certain this will be the ticket.
I also designed a layout for a large, single board to put all the components on that maintains at least 4" of distance between all the coils. I'll post this up when I have the time, I have yet to figure out what size of board will fit in there. The opening is just over 11" wide, I'm thinking an 11"x12-12.5" board will just fit. I will have to do some experimenting to see. Qwin, have you sorted out what size board you are going to use?
I decided to pop in the 4uF Elcap as well on the midrange filter, just to see. Previously I had a 4.4uF mylar film capacitor there.
In comparing the two, it's hard to say one sounds better than the other, but I do detect a bit more of a balance with the speaker with the Elcap. It seems like there's a slight suck-out on the speaker with the mylar film cap still in place. Honestly though, the difference is fairly subtle, it took quite a bit of back and forth to really hear any amount of difference between the two. I think a fresh electrolytic will make a good improvement, as I am liking the way the old Elcaps sound over what I had before, and recalling what I heard when I replaced Elcaps, I am certain this will be the ticket.
I also designed a layout for a large, single board to put all the components on that maintains at least 4" of distance between all the coils. I'll post this up when I have the time, I have yet to figure out what size of board will fit in there. The opening is just over 11" wide, I'm thinking an 11"x12-12.5" board will just fit. I will have to do some experimenting to see. Qwin, have you sorted out what size board you are going to use?
Jeffrey – this is what I am currently using but as an external lash up. I’ve started making and have finished the boards, but need to fit the parts.
The High/Mid/Low filters have been separated and will connect to the three sets of binding posts already fitted. Coils have been re-spaced/oriented to avoid interaction.
Note: This is the biggest board that will go through the ABR aperture and is 310 x 260mm or 12 3/16" x 10 1/4".
You have to angle the board as it goes in, going from vertical (speaker laid on its back) and rotating to horizontal in its final location. This means the board has to be smaller than you might think. I have cut a piece of 3mm ply and tried it, it only just fits.
I am using the Seas Tweeter with the polarity reversed, like the original.
The High/Mid/Low filters have been separated and will connect to the three sets of binding posts already fitted. Coils have been re-spaced/oriented to avoid interaction.
Note: This is the biggest board that will go through the ABR aperture and is 310 x 260mm or 12 3/16" x 10 1/4".
You have to angle the board as it goes in, going from vertical (speaker laid on its back) and rotating to horizontal in its final location. This means the board has to be smaller than you might think. I have cut a piece of 3mm ply and tried it, it only just fits.
I am using the Seas Tweeter with the polarity reversed, like the original.
Last edited:
Very nice Qwin, that looks great! I may have to borrow your template when I put mine onto actual PCBs.
Well, I just placed my order for the Alcaps I'm going to need. I'm not sure if they will arrive before Christmas but when they do I'll be sure to report back. Besides tweaking the treble filter I think I'm pretty well done at this point.
Sent from my iPhone using Tapatalk
Well, I just placed my order for the Alcaps I'm going to need. I'm not sure if they will arrive before Christmas but when they do I'll be sure to report back. Besides tweaking the treble filter I think I'm pretty well done at this point.
Sent from my iPhone using Tapatalk
Hi Jeffrey
I can always send you a PDF of the hole centres I used as a drilling template. I printed it out and stuck it to the board with spray adhesive then drilled through, removed paper by wetting it with white spirit. It is A3 size though so you need a large printer, or as I did, print it on two pages and join them together.
I was advised by someone who makes/sells loudspeakers to stick exactly to what Celestion had in the circuit. The 30uf was originally made up of a 24uf LL Lytic (CCL) and a parallel Elcap 6uf Lytic (not LL).
I hope you ordered the right Alcaps.
The problem is even the Alcaps are now in short supply, Falcon Acoustics no longer have 24uf LL and I got the last of the 25uf LL they had. I tried an Alcap 25uf LL + Alcap 5uf (not LL) and it sounds reasonable, but nothing sounds as good as the originals that Celestion fitted. I’m thinking of leaving the originals in, but folks recon these are well past their best, the age they are, 1973 printed on mine and they could be passing driver damaging DC. I’m looking for a long term solution and fitting Alcaps would mean renewing them in 5 to 10 years. By that time there won’t be any left, you just can’t win.
Mundorf do LL lytics but 22uf is the closest they do to 26, but they don't do an 8uf (not LL). Aaahhhhhh.
Like you I am fed up of the whole thing, I am seriously thinking of ditching the passive crossover all together and going active. I could have built an active x-over for less than I've spent on parts for the passive.
I'm looking at a second hand 6 channel amp and may pull the trigger. This would give each driver its own indipendant channel. I could use op-amp based analogue modules I've spotted to feed filtered signals to each amp channel, or could go for the more expensive but infinately tweakable MiniDSP digital route.
I think I've allready made up my mind and the passive is just a sticking plaster to tide me over till I go active.
I can always send you a PDF of the hole centres I used as a drilling template. I printed it out and stuck it to the board with spray adhesive then drilled through, removed paper by wetting it with white spirit. It is A3 size though so you need a large printer, or as I did, print it on two pages and join them together.
I was advised by someone who makes/sells loudspeakers to stick exactly to what Celestion had in the circuit. The 30uf was originally made up of a 24uf LL Lytic (CCL) and a parallel Elcap 6uf Lytic (not LL).
I hope you ordered the right Alcaps.
The problem is even the Alcaps are now in short supply, Falcon Acoustics no longer have 24uf LL and I got the last of the 25uf LL they had. I tried an Alcap 25uf LL + Alcap 5uf (not LL) and it sounds reasonable, but nothing sounds as good as the originals that Celestion fitted. I’m thinking of leaving the originals in, but folks recon these are well past their best, the age they are, 1973 printed on mine and they could be passing driver damaging DC. I’m looking for a long term solution and fitting Alcaps would mean renewing them in 5 to 10 years. By that time there won’t be any left, you just can’t win.
Mundorf do LL lytics but 22uf is the closest they do to 26, but they don't do an 8uf (not LL). Aaahhhhhh.
Like you I am fed up of the whole thing, I am seriously thinking of ditching the passive crossover all together and going active. I could have built an active x-over for less than I've spent on parts for the passive.
I'm looking at a second hand 6 channel amp and may pull the trigger. This would give each driver its own indipendant channel. I could use op-amp based analogue modules I've spotted to feed filtered signals to each amp channel, or could go for the more expensive but infinately tweakable MiniDSP digital route.
I think I've allready made up my mind and the passive is just a sticking plaster to tide me over till I go active.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.