OK - one measured 4.4ohms the other 6.6ohms. Looks like this accounts for part of the problem. One of them was repaired, and that is the one measuring 6.6ohms.
Good thing I had the spare drivers lying around.
They are marked T.2619 on the baskets.
The part numbers are correct for 66 very late production with large dust cap, which is what you thought you had.
So it looks like these drivers are the problem.
I would concentrate on the other pair of drivers and the x-over which you seem happy with.
I think its worth mentioning, that in my view Polyprop replacements for electrilytics never give quite the same results. Sure they are more transparent and linear in performance, but if the engineers spent months developing the circuit using the parts they did you can bet that it is right for the job. The use of ESR compensating resistors helps but to me the values being used are only effecting the frequency range around the x-over point. This is where any changes will be most noticeable and the minor differences in what folks are using is probably down to fine tunning in terms of integrating their individual drivers and of personel taste. The problem is that it is a fixed value resistor effecting just a small window. The Electrolytics are effected by ambient temperature, current and frequency to a fairly large degree. The amount of variation in ESR over the range they are handeling will be large. This is why a Polyprop will never work like the electrolytic it has replaced and not allways in a good way.
I read a couple of threads earlier in the week on this very same topic, actually many hundreds of threads but these two grabbed my attention as they were from KEF owners working on speakers from the same period. Lots of similarities, two and three way designs using an ABR and electrolytics in the x-over. One of these chaps had replaced his caps one at a time and noticed the more Polyprops he added the worse they sounded. He ended up removing them, he was reasonably knowledgeable and done quite a lot of DIY projects on his amp etc. He was expecting great things from the new better quality caps but he just didn't like the way they changed the character of his speakers. There was pleanty that they did well but he just couldn't live with all that they changed. You see this time after time on recap projects. Initially the owner is full of praise about the clarrity, transparency, less muddy bass, but how many times do you see them come back and get into endless tweaking of parts and values because after living with them for a while they are not happy with the overall balance of the sound they are producing. The circuit was designed to work with these inferior and very differently behaved parts. If you want to get the best from the polyprops you may need to change the circuit design and taylor what is happening in terms of frequency response and impedence matching, a much more involved process.
The results you get with Polyprop replacements will very much depend on the circuit design and some swap outs will work out better than others. But if you like the sound of your speakers and don't want to "upgrade" the sound and spend lots of cash on bean can size components, consider just renewing the electrolytics to protect the drivers and sit back and enjoy the sound you appreciated in the first place.
This is just my point of view and there are bound to be those that disagree very strongly, but they can not argue with the fact that Llytics and Polyprops are different animals and behave very differently in a circuit regardless of quality.
Rant over
So it looks like these drivers are the problem.
I would concentrate on the other pair of drivers and the x-over which you seem happy with.
I think its worth mentioning, that in my view Polyprop replacements for electrilytics never give quite the same results. Sure they are more transparent and linear in performance, but if the engineers spent months developing the circuit using the parts they did you can bet that it is right for the job. The use of ESR compensating resistors helps but to me the values being used are only effecting the frequency range around the x-over point. This is where any changes will be most noticeable and the minor differences in what folks are using is probably down to fine tunning in terms of integrating their individual drivers and of personel taste. The problem is that it is a fixed value resistor effecting just a small window. The Electrolytics are effected by ambient temperature, current and frequency to a fairly large degree. The amount of variation in ESR over the range they are handeling will be large. This is why a Polyprop will never work like the electrolytic it has replaced and not allways in a good way.
I read a couple of threads earlier in the week on this very same topic, actually many hundreds of threads but these two grabbed my attention as they were from KEF owners working on speakers from the same period. Lots of similarities, two and three way designs using an ABR and electrolytics in the x-over. One of these chaps had replaced his caps one at a time and noticed the more Polyprops he added the worse they sounded. He ended up removing them, he was reasonably knowledgeable and done quite a lot of DIY projects on his amp etc. He was expecting great things from the new better quality caps but he just didn't like the way they changed the character of his speakers. There was pleanty that they did well but he just couldn't live with all that they changed. You see this time after time on recap projects. Initially the owner is full of praise about the clarrity, transparency, less muddy bass, but how many times do you see them come back and get into endless tweaking of parts and values because after living with them for a while they are not happy with the overall balance of the sound they are producing. The circuit was designed to work with these inferior and very differently behaved parts. If you want to get the best from the polyprops you may need to change the circuit design and taylor what is happening in terms of frequency response and impedence matching, a much more involved process.
The results you get with Polyprop replacements will very much depend on the circuit design and some swap outs will work out better than others. But if you like the sound of your speakers and don't want to "upgrade" the sound and spend lots of cash on bean can size components, consider just renewing the electrolytics to protect the drivers and sit back and enjoy the sound you appreciated in the first place.
This is just my point of view and there are bound to be those that disagree very strongly, but they can not argue with the fact that Llytics and Polyprops are different animals and behave very differently in a circuit regardless of quality.
Rant over
Never ignore > 100 V polyester . Although not relevant to this conversion polyester sometimes out ranks polypropylene ( memory effect ) . My brother laughs at polypropylene . His company uses it a lot , it is used for extreme cheapness .
Where polyester might work better is if using a Quad 303 as I do . When using the capacitor both as crossover and DC blocking ( 33.5 V ) an opportunity for a no money upgrade exists . That is to bi or tri wire and remove a mild demerit of the Quad ( Sugden A48 , and many more , the 2000 uF output cap in series with everything ) . For what it is worth the Quad might be the finest amplifier ever made . Mostly as it's low distortion also has ideal maths for it's distortion harmonics ( a super tube amp with 40 dB better distortion ) . This simple upgrade might be decisive . Being class AB sometimes means it looses detail . Even that is so close to not being so ( Output triples and good bias behaviour ) . If a Quad sounds less than wonderful look elsewhere or repair it . The 405 was not as good although a nice amp . It phase inverts so always have black to speaker red . No one ever says that , why ?
I note that EPCOS 100 and 250 V polyesters are very cheap . I suspect it is the end of the line for them . The 400 V types also . Don't use 63 V types . The build requirements for space dominate the calculations when 63V . The high voltage types as a byproduct have a better tan theta . The 63 V types for want of a better word sound bloated . You can use mains suppression caps . These sound rather good as they follow the same logical . A suppression cap might be 630 VDC and 250VAC ( that is about testing rather than RMS , it implies 445 VAC which would be right , 250 VAC is the UL rating ) . 1uF costs about 50 pence and most likely will be polypropylene . The cheap ones sold by Rapid Electronics ( e.g 10-2508 ) being fine . Having taken so many of these apart I have great doubts about Audio Grade caps . They all look the same to me .
If an electrolytic is replaced be prepared for not liking the sound . Sometimes the designer did have good ears and balanced off the vices . The special electrolytic's used in speakers are exceptionally good at low voltages . In fact they give benchmark figures up to 0.4 V ( - 150 dB distortion ) . I use them a lot in low level circuits where they can considerably outperform all other types . They are dirt cheap which makes people think them no good . After 0.4 V they become ordinary, high voltage types seem the best as before . When a tweeter 0.4 V is about what it might be . Never polarize an electrolytic if you can help it ( respect it's preferred direction as there is always some bias in an active circuit ) . Above 0.4 volts they start to behave as rectifiers . It is surprising how often in low level circuits 0.4 V is fine . And to think , put them in any way you like . I usually find a circuit with 100 % DC feedback sounds best . I make things for people and find the caps replaced . Never mind .
Where polyester might work better is if using a Quad 303 as I do . When using the capacitor both as crossover and DC blocking ( 33.5 V ) an opportunity for a no money upgrade exists . That is to bi or tri wire and remove a mild demerit of the Quad ( Sugden A48 , and many more , the 2000 uF output cap in series with everything ) . For what it is worth the Quad might be the finest amplifier ever made . Mostly as it's low distortion also has ideal maths for it's distortion harmonics ( a super tube amp with 40 dB better distortion ) . This simple upgrade might be decisive . Being class AB sometimes means it looses detail . Even that is so close to not being so ( Output triples and good bias behaviour ) . If a Quad sounds less than wonderful look elsewhere or repair it . The 405 was not as good although a nice amp . It phase inverts so always have black to speaker red . No one ever says that , why ?
I note that EPCOS 100 and 250 V polyesters are very cheap . I suspect it is the end of the line for them . The 400 V types also . Don't use 63 V types . The build requirements for space dominate the calculations when 63V . The high voltage types as a byproduct have a better tan theta . The 63 V types for want of a better word sound bloated . You can use mains suppression caps . These sound rather good as they follow the same logical . A suppression cap might be 630 VDC and 250VAC ( that is about testing rather than RMS , it implies 445 VAC which would be right , 250 VAC is the UL rating ) . 1uF costs about 50 pence and most likely will be polypropylene . The cheap ones sold by Rapid Electronics ( e.g 10-2508 ) being fine . Having taken so many of these apart I have great doubts about Audio Grade caps . They all look the same to me .
If an electrolytic is replaced be prepared for not liking the sound . Sometimes the designer did have good ears and balanced off the vices . The special electrolytic's used in speakers are exceptionally good at low voltages . In fact they give benchmark figures up to 0.4 V ( - 150 dB distortion ) . I use them a lot in low level circuits where they can considerably outperform all other types . They are dirt cheap which makes people think them no good . After 0.4 V they become ordinary, high voltage types seem the best as before . When a tweeter 0.4 V is about what it might be . Never polarize an electrolytic if you can help it ( respect it's preferred direction as there is always some bias in an active circuit ) . Above 0.4 volts they start to behave as rectifiers . It is surprising how often in low level circuits 0.4 V is fine . And to think , put them in any way you like . I usually find a circuit with 100 % DC feedback sounds best . I make things for people and find the caps replaced . Never mind .
Qwin,
I definitely agree with you about the differences in the capacitors and their impact and effect on the sound in designs where a specific part was used (such as 'lytics). In almost all of my speaker restorations I have used electrolytics, and in fact most of my speaker restorations have been British speakers that used Elcaps or Alcaps and I replace those with Alcaps to get the absolute closest thing to the original. Such as my IMF RSPM Mk IVs - I used Alcaps and the sound is wonderful.
Back to the polyprops. I was concerned about the sound quality of my Celestion 66s when I through in my crossovers that I put together using the inductors from old 66 boards and Solen polypropylene capacitors for the mid and treble circuit and Alcap for the bass section. I know that even ESR cannot fully compensate for the difference. However, as I sit here and type this I must say the 66s really do sound wonderful. I took your advice and dropped in another 6uF Solen to bring up the midrange filter to 30uF. After some swapping back and forth between the other one with 24uF I decided that it sounds more balanced to my ear.
Unfortunately through this testing I discovered one of the MF500s has developed some resonance issues at very specific frequencies. A shame, but overall it is working well. However it will need to be addressed at some point. I think I will rather find some spare MF500s and do experiments attempting to fix this one before I bother pulling this one apart, as I fear if I mess it up I will be stuck without any midrange at all in one speaker and that won't do.
Another thing I wanted to mention was charge coupling - ever heard of this technique? You apply a 9volt battery to the crossover, add in resistors and double the capacitance - it results in electrolytics performing much better. I have had one pair of speakers with this modification done and I compared them directly against the same model that had just recently been recapped (using Alcaps to replace Elcaps) and couldn't believe the difference. It's a cheap but very effective way to improve the speaker's performance and may alleviate any concern about swapping out different parts.
I definitely agree with you about the differences in the capacitors and their impact and effect on the sound in designs where a specific part was used (such as 'lytics). In almost all of my speaker restorations I have used electrolytics, and in fact most of my speaker restorations have been British speakers that used Elcaps or Alcaps and I replace those with Alcaps to get the absolute closest thing to the original. Such as my IMF RSPM Mk IVs - I used Alcaps and the sound is wonderful.
Back to the polyprops. I was concerned about the sound quality of my Celestion 66s when I through in my crossovers that I put together using the inductors from old 66 boards and Solen polypropylene capacitors for the mid and treble circuit and Alcap for the bass section. I know that even ESR cannot fully compensate for the difference. However, as I sit here and type this I must say the 66s really do sound wonderful. I took your advice and dropped in another 6uF Solen to bring up the midrange filter to 30uF. After some swapping back and forth between the other one with 24uF I decided that it sounds more balanced to my ear.
Unfortunately through this testing I discovered one of the MF500s has developed some resonance issues at very specific frequencies. A shame, but overall it is working well. However it will need to be addressed at some point. I think I will rather find some spare MF500s and do experiments attempting to fix this one before I bother pulling this one apart, as I fear if I mess it up I will be stuck without any midrange at all in one speaker and that won't do.
Another thing I wanted to mention was charge coupling - ever heard of this technique? You apply a 9volt battery to the crossover, add in resistors and double the capacitance - it results in electrolytics performing much better. I have had one pair of speakers with this modification done and I compared them directly against the same model that had just recently been recapped (using Alcaps to replace Elcaps) and couldn't believe the difference. It's a cheap but very effective way to improve the speaker's performance and may alleviate any concern about swapping out different parts.
If I can find it I will try to show how badly an electrolytic cap performs once charged , I have some data somewhere . Naturally when in a PSU or as the output device there is no choice . Folk-Law has it that charging the caps to half stated voltage is ideal . Inside the cap is a lattice of resistance and inductance . As 0.4V is reached a rectifier effect kicks in .This makes the open sound of the sub 0.4V become dull and pinched ( sorry to ascribe it a sound , best to say something to give a hint ) .
I suspect what the person thinks is that two standard electrolytic's will be cheaper or easier to get . To polarize them might help them carry more current .
I use a large number of non polars in my work . None of that is loudspeakers . I use them because they are very good . A cheap Panasonic or Nikkia non polar kept below 0.4V will out perform an expensive Black Gate cap .
One trick I have seen is a non polar and poly cap in parallel . About 90 % non polar and 10 % film type .
One cap you could try although I have doubts would be Rifa Evox paper X2 grade suppression caps . I suspect they won't be far away from the old paper caps JBL used . If paper it almost becomes possible to make them . I dare say if bulk is not a problem kitchen foil would be OK . I would use cling film also to compare as dielectric ( cling film and toilet paper ? ) . Mostly we see 10 V rms so flash-over should not be a problem . A cheap capacitance meter and reference cap will be needed and clone caps you have ( open one to see how it is done ) . The " bad " thing about being bulky might sound better . That is sound exactly like the electrolytic , yet maintain that sound at higher volume ( > 0.4V ) . The bad thing being loss of > 20 kHz perhaps . Reality say more like 50 kHz .
I suspect what the person thinks is that two standard electrolytic's will be cheaper or easier to get . To polarize them might help them carry more current .
I use a large number of non polars in my work . None of that is loudspeakers . I use them because they are very good . A cheap Panasonic or Nikkia non polar kept below 0.4V will out perform an expensive Black Gate cap .
One trick I have seen is a non polar and poly cap in parallel . About 90 % non polar and 10 % film type .
One cap you could try although I have doubts would be Rifa Evox paper X2 grade suppression caps . I suspect they won't be far away from the old paper caps JBL used . If paper it almost becomes possible to make them . I dare say if bulk is not a problem kitchen foil would be OK . I would use cling film also to compare as dielectric ( cling film and toilet paper ? ) . Mostly we see 10 V rms so flash-over should not be a problem . A cheap capacitance meter and reference cap will be needed and clone caps you have ( open one to see how it is done ) . The " bad " thing about being bulky might sound better . That is sound exactly like the electrolytic , yet maintain that sound at higher volume ( > 0.4V ) . The bad thing being loss of > 20 kHz perhaps . Reality say more like 50 kHz .
Thought I should say a bit more . With a non polar cap what we are getting is a similar thing to class AB amplifiers . That is we maintain class A as long as we can . This technique has been discredited by the " measurements are better than ears " brigade . Another thread I am on one guy makes an old fashioned AB and has pushed it to the right level to work . Here we have the same . The more critical levels are served by the cap at it's best .
In my work I use tons of measurements . Mostly I ignore them beyond a certain point . After the listening phase I measure again . Usually the design even in measured terms has advanced . The measurements from then are to say the amp performs to book spec . Now to the crunch . Any deviation will not sound the same . Why it sounds good is seldom totally in the measurements except to say it is or it is not working to book .
If interested . The ideal production test would be a final null test against the prototype . That would be a musical signal or another fed in to both that devices that will produce 0V output if in some form of anti-phase . Tools designed for testing should never do the listening . To stop daft designs they are invaluable . Many things I design have their daft moments .
The polarized caps being better might be simply be that they are better devices . Paper caps should win all the arguments the more I think about it . The difference in construction between paper and non polar is zero . The difference is the electrolyte . This for want of a better way of saying it mimics a battery . In a non polar it has to polarize every time the voltage changes direction/polarity . Inside the cap both foils are shiny . Take apart a polar electrolytic to see that one side of the foils is mildly matt . That capacitor must not be reversed . Up to 0.4V you can reverse polar caps with reasonable success . For example in an amplifier input long tail pair it probably will be OK ( feedback cap ) . As soon as current is involved you can't . Thus a battery is required . In theory 9V is just about enough . Ideally it should be more . The 0.4V is the tipping point . After that it is more about what gives the better protection of the caps . I dare say 63 V caps with 30 V polarization would be best . 63 V electrolytic caps have far better tan theta ( 0.2 reduces to 0.12 perhaps ) . I dare say 200 V types set to 100 V would be ideal if taking that route . To keep a cap safe at high voltage also makes it sound better , it is a byproduct . All of these things cost very little to try so might pass the winter months very well . If you take the crossover outside of the speaker to test it might sound slightly different when returned to the box . Most likely inductors close to big magnets cause that .
In my work I use tons of measurements . Mostly I ignore them beyond a certain point . After the listening phase I measure again . Usually the design even in measured terms has advanced . The measurements from then are to say the amp performs to book spec . Now to the crunch . Any deviation will not sound the same . Why it sounds good is seldom totally in the measurements except to say it is or it is not working to book .
If interested . The ideal production test would be a final null test against the prototype . That would be a musical signal or another fed in to both that devices that will produce 0V output if in some form of anti-phase . Tools designed for testing should never do the listening . To stop daft designs they are invaluable . Many things I design have their daft moments .
The polarized caps being better might be simply be that they are better devices . Paper caps should win all the arguments the more I think about it . The difference in construction between paper and non polar is zero . The difference is the electrolyte . This for want of a better way of saying it mimics a battery . In a non polar it has to polarize every time the voltage changes direction/polarity . Inside the cap both foils are shiny . Take apart a polar electrolytic to see that one side of the foils is mildly matt . That capacitor must not be reversed . Up to 0.4V you can reverse polar caps with reasonable success . For example in an amplifier input long tail pair it probably will be OK ( feedback cap ) . As soon as current is involved you can't . Thus a battery is required . In theory 9V is just about enough . Ideally it should be more . The 0.4V is the tipping point . After that it is more about what gives the better protection of the caps . I dare say 63 V caps with 30 V polarization would be best . 63 V electrolytic caps have far better tan theta ( 0.2 reduces to 0.12 perhaps ) . I dare say 200 V types set to 100 V would be ideal if taking that route . To keep a cap safe at high voltage also makes it sound better , it is a byproduct . All of these things cost very little to try so might pass the winter months very well . If you take the crossover outside of the speaker to test it might sound slightly different when returned to the box . Most likely inductors close to big magnets cause that .
Jeffrey88 - Not come accross Charge Coupling before, interesting.
Sorry to here about your MF500, try and find a replacement and do like you described.
Nigel Pearson - Thanks for the input on cap construction/workings.
I will be keeping my mods simple at first, just replacing like for like and keeping the x-overs internal. I want to keep the speakers as original as possible. Its not easy to do external x-overs on these sealed boxes even while just experimenting. I put the cables through the vent port on bass reflex models, but with the ABR and sealed system it would mean making a hole/holes big enough for the three sets of driver cables, so not ideal.
Sorry to here about your MF500, try and find a replacement and do like you described.
Nigel Pearson - Thanks for the input on cap construction/workings.
I will be keeping my mods simple at first, just replacing like for like and keeping the x-overs internal. I want to keep the speakers as original as possible. Its not easy to do external x-overs on these sealed boxes even while just experimenting. I put the cables through the vent port on bass reflex models, but with the ABR and sealed system it would mean making a hole/holes big enough for the three sets of driver cables, so not ideal.
Hey guys,
It's been a while!
First of all my MF500 woes have been solved. I took my MF500 to a local repair guy and he fixed the issue. He said it had to do with delamination of the dome from the magnet. He replace the lamination material, and voila, problem solved!
So, now I am hearing them fully for the first time. I must admit, they sound good but I can't help but feel like something is still missing. My suspicion is with the tweeter.. on my receiver, I find bumping up the treble by +2dB really improves the sound a lot. The midrange just sounds too strong otherwise. Perhaps it's my young ears, but this to me sounds much, much better!
What I am asking is, what modification do I need to do to the crossover to achieve this so I can run the filter on the receiver flat?
For what it's worth, the crossovers are totally rebuilt. I am using Seas 19TFF1 tweeters with the crossover amply modified - I am using 30ohms resistance to allow the tweeter to realize the same resistance as the HF2000 for the crossover. The capacitance has been modified to fit the new tweeter.
Any ideas? I feel like I am on the verge of really satisfying sound. But it's just not quite there yet. If you need more info please let me know.
It's been a while!
First of all my MF500 woes have been solved. I took my MF500 to a local repair guy and he fixed the issue. He said it had to do with delamination of the dome from the magnet. He replace the lamination material, and voila, problem solved!
So, now I am hearing them fully for the first time. I must admit, they sound good but I can't help but feel like something is still missing. My suspicion is with the tweeter.. on my receiver, I find bumping up the treble by +2dB really improves the sound a lot. The midrange just sounds too strong otherwise. Perhaps it's my young ears, but this to me sounds much, much better!
What I am asking is, what modification do I need to do to the crossover to achieve this so I can run the filter on the receiver flat?
For what it's worth, the crossovers are totally rebuilt. I am using Seas 19TFF1 tweeters with the crossover amply modified - I am using 30ohms resistance to allow the tweeter to realize the same resistance as the HF2000 for the crossover. The capacitance has been modified to fit the new tweeter.
Any ideas? I feel like I am on the verge of really satisfying sound. But it's just not quite there yet. If you need more info please let me know.
Hi Jeffrey
I have also gone for the Seas 19TFF1.
Took a slightly different approach and replaced the 0.14mH inductor with a 0.17Mh and used 9uf and 3uf caps in series. This is what most online calculators show for a 5kHz crossover point when using a 7 Ohm driver. Note that the Seas driver is 7 Ohm inductance at 5kHz, I did a frequency sweep using Dayton Audio's DATS system to establish this. Using these values you don't need any resistors and the Tweeter integrates well.
I have also split the x-over circuit into its three elements and gone Tri-Wired. This helps lower resonance.
When disconecting the tweeter you can hear how much of the top end is actually supplied by the mid driver. I think it is your mid range circuit that needs sorting.
Have you used Polyprops in the mid range and worse still used ESR compensating resistors? If so this is where your problem lies. The mid range band pass circuit is highly resonant as discussed earlier in the thread with graphs to show it. Using Polyprops makes the resonance much more pronounced and adding ever increasing value of ESR resistors reduces the resonance but kills the drive and dynamics and dulls the lower treble. I have tried ALL the variations on this site and I mean ALL and none of it works in the mid range. Going back to the original Elcaps actually improved the sound over the many makes and values of Polyprops and dozens of resistor combinations I had tried.
I have my x-overs external and all caps were pre burned 60hrs before use. I had terminal blocks on the board and could change caps and resistors while music was playing. I have spent hundreds of hours listening and tweaking the mid range and eventually gone back to electrolytics here, which sound a massive amount better.
Note I tried Sonicap, ClarityCap ESA, Solen, and Ansar, caps in all positions.
Replacing the 4uf parallel cap with a regular Alcap 100v item and the 30uf with Alcap 25 + 5 uf 50v Low loss works much better. Much more open sound with greater dynamics. A friend explained that its not just the ESR that is different with Polyprops but also the disapation which in the electrolytics acts as a damping factor on the natural resonance of this generic bandpass filter.
I know what I am suggesting goes against much of what has been done on this thread but I am telling you what I have actually tried and got results with.
Unfortunately Alcap suplies, especially in the LL values are almost exhausted, I have found the Mundorf 70v "Plain" electrolytics which are LL work well (15 + 15uf). They have a slightly different sound on their own and need a 1.2 to 1.5 Ohm resistor in series to get a sound like the originals. The original 30uf was made up of a 24uf LL and a 6uf normal electrolytic. The LL have lower ESR so getting the balance of the originals is tricky. The resonance with the Mundorfs is minimal and way better than the Poly props and has none of the shut in dull nature. They are fast/dynamic and give a sound more like the originals.
After trying the full Polyprop/ESR set up and then this combination of PP on the tweeter and electrolytics on the Mid driver, there is just no comparison and the later sounds so much more like the original voicing.
I found the best combination for the Tweeter was a Sonicap gen1 9uf and and a ClarityCap ESA for the 3uf. Using a Sonicap for the 3uf is a bit bright and has less texture and timbre on piano notes, the mixture of the two worked best for me. In terms of changing the coil to 0.17mH it is just not worth trying to **** the circuit to use the originals as these coils are cheep as chips. I used a 0.45 Ohm version I had in my parts bin.
I could fill pages with what I have tried but my advice to you is this - the electrolytics for the mid range are cheap, so try what I have suggested and make your own mind up, I think it will lift your lower treble. Change your tweeter cap/coil values for the Seas driver. I tried the values suggested on here and it sounded terrible. Even fitting the Seas driver to the original celestion values for the HF2000 worked as good.
Somthing to consider.
I have also gone for the Seas 19TFF1.
Took a slightly different approach and replaced the 0.14mH inductor with a 0.17Mh and used 9uf and 3uf caps in series. This is what most online calculators show for a 5kHz crossover point when using a 7 Ohm driver. Note that the Seas driver is 7 Ohm inductance at 5kHz, I did a frequency sweep using Dayton Audio's DATS system to establish this. Using these values you don't need any resistors and the Tweeter integrates well.
I have also split the x-over circuit into its three elements and gone Tri-Wired. This helps lower resonance.
When disconecting the tweeter you can hear how much of the top end is actually supplied by the mid driver. I think it is your mid range circuit that needs sorting.
Have you used Polyprops in the mid range and worse still used ESR compensating resistors? If so this is where your problem lies. The mid range band pass circuit is highly resonant as discussed earlier in the thread with graphs to show it. Using Polyprops makes the resonance much more pronounced and adding ever increasing value of ESR resistors reduces the resonance but kills the drive and dynamics and dulls the lower treble. I have tried ALL the variations on this site and I mean ALL and none of it works in the mid range. Going back to the original Elcaps actually improved the sound over the many makes and values of Polyprops and dozens of resistor combinations I had tried.
I have my x-overs external and all caps were pre burned 60hrs before use. I had terminal blocks on the board and could change caps and resistors while music was playing. I have spent hundreds of hours listening and tweaking the mid range and eventually gone back to electrolytics here, which sound a massive amount better.
Note I tried Sonicap, ClarityCap ESA, Solen, and Ansar, caps in all positions.
Replacing the 4uf parallel cap with a regular Alcap 100v item and the 30uf with Alcap 25 + 5 uf 50v Low loss works much better. Much more open sound with greater dynamics. A friend explained that its not just the ESR that is different with Polyprops but also the disapation which in the electrolytics acts as a damping factor on the natural resonance of this generic bandpass filter.
I know what I am suggesting goes against much of what has been done on this thread but I am telling you what I have actually tried and got results with.
Unfortunately Alcap suplies, especially in the LL values are almost exhausted, I have found the Mundorf 70v "Plain" electrolytics which are LL work well (15 + 15uf). They have a slightly different sound on their own and need a 1.2 to 1.5 Ohm resistor in series to get a sound like the originals. The original 30uf was made up of a 24uf LL and a 6uf normal electrolytic. The LL have lower ESR so getting the balance of the originals is tricky. The resonance with the Mundorfs is minimal and way better than the Poly props and has none of the shut in dull nature. They are fast/dynamic and give a sound more like the originals.
After trying the full Polyprop/ESR set up and then this combination of PP on the tweeter and electrolytics on the Mid driver, there is just no comparison and the later sounds so much more like the original voicing.
I found the best combination for the Tweeter was a Sonicap gen1 9uf and and a ClarityCap ESA for the 3uf. Using a Sonicap for the 3uf is a bit bright and has less texture and timbre on piano notes, the mixture of the two worked best for me. In terms of changing the coil to 0.17mH it is just not worth trying to **** the circuit to use the originals as these coils are cheep as chips. I used a 0.45 Ohm version I had in my parts bin.
I could fill pages with what I have tried but my advice to you is this - the electrolytics for the mid range are cheap, so try what I have suggested and make your own mind up, I think it will lift your lower treble. Change your tweeter cap/coil values for the Seas driver. I tried the values suggested on here and it sounded terrible. Even fitting the Seas driver to the original celestion values for the HF2000 worked as good.
Somthing to consider.
Last edited:
Further to the previous post - Note that I should have said the Seas tweeter was 7 Ohm impedance not inductance.
I have been doing tests this morning on the 4 Ohm parallel cap in the mid range filter. I find that with the Alcap electrolytics, the 50v version works better than the 100v one. Don't ask me why but it does, slightly less resonant and a bit cleaner sounding in general. tried them in and out three times each and its a fairly obvious difference. Both sets of caps have many hours on them, were fitted the same way round in each case (not that that should make a difference) and were purchased pre graded by the supplier to 2% tolerance, rather than 10% as is standard.
I have been doing tests this morning on the 4 Ohm parallel cap in the mid range filter. I find that with the Alcap electrolytics, the 50v version works better than the 100v one. Don't ask me why but it does, slightly less resonant and a bit cleaner sounding in general. tried them in and out three times each and its a fairly obvious difference. Both sets of caps have many hours on them, were fitted the same way round in each case (not that that should make a difference) and were purchased pre graded by the supplier to 2% tolerance, rather than 10% as is standard.
Last edited:
Hey Qwin,
First of all thank you for your time and response. It's very much appreciated.
Secondly I must commend you on your efforts, that is quite a lot of experimenting!
I'm not entirely surprised by your findings. I've been thinking the issue lies with the caps. As you suspected, I am using Solen polypropylene capacitors on the midrange and treble filters and also have resistors in place to compensate for ESR. I am using a 0.14mH coil (the original) on the tweeter. The capacitance was adjusted, I can't remember exactly to what but the values were exactly what was suggested to me earlier in another thread (by Alan).
I've sent an email to Falcon to inquire about the Alcaps. I've used them quite a bit, in restoring speakers, so I am going to get those if I can.
For the new inductor what gauge did you end up going with?
Again, I sincerely appreciate your input. I have a feeling you are "spot on" in your evaluation of my issue.
Jeffrey
Sent from my iPhone using Tapatalk
First of all thank you for your time and response. It's very much appreciated.
Secondly I must commend you on your efforts, that is quite a lot of experimenting!
I'm not entirely surprised by your findings. I've been thinking the issue lies with the caps. As you suspected, I am using Solen polypropylene capacitors on the midrange and treble filters and also have resistors in place to compensate for ESR. I am using a 0.14mH coil (the original) on the tweeter. The capacitance was adjusted, I can't remember exactly to what but the values were exactly what was suggested to me earlier in another thread (by Alan).
I've sent an email to Falcon to inquire about the Alcaps. I've used them quite a bit, in restoring speakers, so I am going to get those if I can.
For the new inductor what gauge did you end up going with?
Again, I sincerely appreciate your input. I have a feeling you are "spot on" in your evaluation of my issue.
Jeffrey
Sent from my iPhone using Tapatalk
Hi Jeffrey
The coils were some that I had just lying around, they were from an old set of B&O speakers and about 0.2mH. I wound some coils off to get them down to 0.17mH were they ended up at 0.45 Ohm (RCD). The wire is about 0.7mm diameter when I measure it with a vernier caliper guage.
I am still tweaking the resistor values etc. I have added 0.6dB attenuation on the tweeter this afternoon as an L-Pad circuit (0.5 Ohm series/100 Ohm parallel), which sounds good. Need to do some more listening to decide if that's a keeper.
If you use 0.17mH inductor you will need 9uf and 3uf caps. Solen are not the best for HF circuits in my opinion, a bit grainy and over bright. You don't need ESR resistors for the tweeter caps and any L-Pad attenuation is done on the driver end of the circuit.
I also tried swapping the HF mixture of caps which has made a small improvement, 9uf Sonicap now ClarityCap, 3uf ClarityCap now 3.3uf Sonicap. (Reversed the mix between the two). I only had a 3.3uf to hand but it computes at 3.03uf on the online calculator I used. Never the less it sounds good, the calculators are only a starting point anyway, you allways end up adjusting by ear.
Note: Falcon Accoustics are running out of many values (thats who I use for Alcaps). That's why I looked at the Mundorf electrolytics which are current production and not NOS (HiFi Collective).
Ken
The coils were some that I had just lying around, they were from an old set of B&O speakers and about 0.2mH. I wound some coils off to get them down to 0.17mH were they ended up at 0.45 Ohm (RCD). The wire is about 0.7mm diameter when I measure it with a vernier caliper guage.
I am still tweaking the resistor values etc. I have added 0.6dB attenuation on the tweeter this afternoon as an L-Pad circuit (0.5 Ohm series/100 Ohm parallel), which sounds good. Need to do some more listening to decide if that's a keeper.
If you use 0.17mH inductor you will need 9uf and 3uf caps. Solen are not the best for HF circuits in my opinion, a bit grainy and over bright. You don't need ESR resistors for the tweeter caps and any L-Pad attenuation is done on the driver end of the circuit.
I also tried swapping the HF mixture of caps which has made a small improvement, 9uf Sonicap now ClarityCap, 3uf ClarityCap now 3.3uf Sonicap. (Reversed the mix between the two). I only had a 3.3uf to hand but it computes at 3.03uf on the online calculator I used. Never the less it sounds good, the calculators are only a starting point anyway, you allways end up adjusting by ear.
Note: Falcon Accoustics are running out of many values (thats who I use for Alcaps). That's why I looked at the Mundorf electrolytics which are current production and not NOS (HiFi Collective).
Ken
Last edited:
Something was said about a resistor in part of the crossover. One thing I have always thought is the BBC auto-transformer is a good idea. It keeps the damping as high as possible and accepts that it might not be bass units alone that like to see the amplifier? This might be why active produces a very different sound? Often a worse sound as drive units are made to work in that design.
The other nice possibility of the transformer is to arive at > 16 R. Many amplifiers look much nicer into 16R above 10 kHz. If 4R up to 1 kHz that can be fine. Then 8R from 1kHz to 7 kHz ? Better still if not that dramatic.
The other nice possibility of the transformer is to arive at > 16 R. Many amplifiers look much nicer into 16R above 10 kHz. If 4R up to 1 kHz that can be fine. Then 8R from 1kHz to 7 kHz ? Better still if not that dramatic.
Resonance isn't such a problem with 2 way systems.
As soon as you add the third driver you need a bandpass filter that adjusts either end of the drivers range.
Everything I have read about this and I am no expert, suggests that the generic band pass filter is resonant. Basicaly if you put a coil and cap in series it will resonate, everything I have tried would suggest this is the case. Without adding additional knotch filters etc the resonance is allways going to be a problem and those kind of mods require measuring equipment and the knowledge to interperate the results.
Electrolytic caps with their variable ESR and damping factor tend to hide the resonance. I can't honestly say I was aware of it till I put Plolyprop caps and fixed ESR simulating resistors in the midrange filter. Then it was like a poke in the eye, Uughhh what is that distortion, kind of moment. I have been trying to get it to an acceptable level ever since. Thing is, now that I am focused on the fault and how it effects certain tracks and voices I am more sensitive to the issue and even pick up on it when putting the original Celestion circuit parts back into use. It is far less noticeable but still there to a degree in the original, hence my return to electrolytics. Elcap electrolytics are no longer available and Alcap supplies are dwindelling and it doesn't help that different brands of the same value have quite different ESR/Dissapation properties.
If I can't get a satisfactory compromise, I might try going active with some low power/cost T-Amps just to see how it compares.
As soon as you add the third driver you need a bandpass filter that adjusts either end of the drivers range.
Everything I have read about this and I am no expert, suggests that the generic band pass filter is resonant. Basicaly if you put a coil and cap in series it will resonate, everything I have tried would suggest this is the case. Without adding additional knotch filters etc the resonance is allways going to be a problem and those kind of mods require measuring equipment and the knowledge to interperate the results.
Electrolytic caps with their variable ESR and damping factor tend to hide the resonance. I can't honestly say I was aware of it till I put Plolyprop caps and fixed ESR simulating resistors in the midrange filter. Then it was like a poke in the eye, Uughhh what is that distortion, kind of moment. I have been trying to get it to an acceptable level ever since. Thing is, now that I am focused on the fault and how it effects certain tracks and voices I am more sensitive to the issue and even pick up on it when putting the original Celestion circuit parts back into use. It is far less noticeable but still there to a degree in the original, hence my return to electrolytics. Elcap electrolytics are no longer available and Alcap supplies are dwindelling and it doesn't help that different brands of the same value have quite different ESR/Dissapation properties.
If I can't get a satisfactory compromise, I might try going active with some low power/cost T-Amps just to see how it compares.
I am curious if any one tried to design a new crossover based on using new Polypropylene caps instead of playing with ESR compensation? Since nobody know the exact value of the ESR with the original Elcap. Would it be easier or harder to get the aimed crossover point if re-designing a crossover based on new components?
After reccap my 66 with the ESR simulation, I still feel that the mid and high could be improved. I have been reading about different type of speaker design. I came across the open baffle design and it get me very interested in trying and OB in the future. That will belong to another topic.Back to the 66, I then also read about the Audio Note speaker which they use no damping material. This inspired me to do some experiment with the 66. I took out the foam at the top behind the tweeter and mid range while keeping the rest of the foam. I am not sure if it is my imagination or I do actually hearing improvement. It does seems to me the high is more extended and there is just a little more air so it sound more live like. So I suggest other 66 owner to try it and report if they hear any differences.
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.
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.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.