Hi Tinitus,
Thanks for the reply
Yes I agree that any change in cap types is only going to be minor compared to designing the circuit properly in the first place Even using bipolar electros (in a correct design) should sound better than using the most exotic caps in a poor design. I think your point about contruction of the caps is also very valid. I've read about microphonics of caps and also about how sqeezing caps will show changes in capacitance... Since most crossovers are inside the speaker cabinets, they are exposed to all sorts of resonances and pressure changes so I can see how caps that are more immune to these effects will be beneficial
Lucas,
If you add impedance compensation you will almost certainly change the sound of the speakers. The impedance compensation will basically present a more uniform impedance to the filter meaning it will work the same (more or less) at all frequencies. If there was no comensation the original network will get less effective as the natural driver impedance rises... This may or may not be desirable depending on how the overall original design was arrived at. Also You will probably get an overall cut in the output level of the woofer, which will likely then necessitate changes in other areas of the crossover to comensate.
Tony.
Thanks for the reply
Yes I agree that any change in cap types is only going to be minor compared to designing the circuit properly in the first place Even using bipolar electros (in a correct design) should sound better than using the most exotic caps in a poor design. I think your point about contruction of the caps is also very valid. I've read about microphonics of caps and also about how sqeezing caps will show changes in capacitance... Since most crossovers are inside the speaker cabinets, they are exposed to all sorts of resonances and pressure changes so I can see how caps that are more immune to these effects will be beneficial Lucas,
If you add impedance compensation you will almost certainly change the sound of the speakers. The impedance compensation will basically present a more uniform impedance to the filter meaning it will work the same (more or less) at all frequencies. If there was no comensation the original network will get less effective as the natural driver impedance rises... This may or may not be desirable depending on how the overall original design was arrived at. Also You will probably get an overall cut in the output level of the woofer, which will likely then necessitate changes in other areas of the crossover to comensate.
Tony.
Well Tony, here's the schematic of the original design, for crossovers points at 5 & 5K.
What impedance compensation is there? I am far from expert - in fact, I'm a novice, but I see no resistors. Maybe the driver's impedances were made to match? Correct me if I'm wrong, but that's almost certainly the case I would think as all of the drivers were actually made by Celestion themselves...
Lucas
What impedance compensation is there? I am far from expert - in fact, I'm a novice, but I see no resistors. Maybe the driver's impedances were made to match? Correct me if I'm wrong, but that's almost certainly the case I would think as all of the drivers were actually made by Celestion themselves...
Lucas
Last edited:
Thanks Tony
Lucas, its really impossible to suggest anything without hearing your speaker
But I could suggest to try and remove the 72uf closest to woofer
If that doesnt sound good, try a smaller value, with resistor to ground
Thats an impedance correction
Mayb also try a resistor to ground on the first 72uf, but its value may need changed too
I expect the first woofer inductor L1 to be bigger than L4?
I could also suggest to mount a resistor in series with the paralel 6uf on midrange(maybe try other values from 4.7uf to 10uf, cheap ones)
This isnt easy, and could become a new crossover nightmare
It depend on how much effort you want to put in this, how many months you like to spend with your speaker
btw, earlier I gave you some very simple ways to determine whether a component should preferably be smaller or bigger
But I wouldnt rely on a 35year old design to be correct
Lucas, its really impossible to suggest anything without hearing your speaker
But I could suggest to try and remove the 72uf closest to woofer
If that doesnt sound good, try a smaller value, with resistor to ground
Thats an impedance correction
Mayb also try a resistor to ground on the first 72uf, but its value may need changed too
I expect the first woofer inductor L1 to be bigger than L4?
I could also suggest to mount a resistor in series with the paralel 6uf on midrange(maybe try other values from 4.7uf to 10uf, cheap ones)
This isnt easy, and could become a new crossover nightmare
It depend on how much effort you want to put in this, how many months you like to spend with your speaker
btw, earlier I gave you some very simple ways to determine whether a component should preferably be smaller or bigger
But I wouldnt rely on a 35year old design to be correct
Last edited:
Hi Lucas, yes I believe you are correct in saying there is no impedance compensation, it just looks like a standard low pass filter on the woofer to me.. That is why I said if you add it, it will almost certainly change the sound of the speakers I thought you were mostly interested in keeping the crossover as close to original as possible but upgrading the aging capacitors... I haven't seen some of the more recent posts, so sorry if I missed some important info...
Tony.
I thought you were mostly interested in keeping the crossover as close to original as possible but upgrading the aging capacitors... I haven't seen some of the more recent posts, so sorry if I missed some important info... Tony.
I get that.
Why would I want to "correct" an impedance of a speaker that was designed by a notoriously excellent speaker driver manufacturer? What makes you think it's wrong, or doesn't sound good? The inductors are almost if not exactly the same size.
OK, perhaps you're right....but why?
Sounds like a real crossover nightmare! I wish I could change the stupid name of this thread....it wasn't even the crossover at fault in the original problem....I'll continue to swap them out, see how it sounds and take it from there...
Why? Has crossover technology or knowledge changed so much since 1973? Celestion were making some pretty amazing stuff back then and these still sound better in many respects than most expensive modern two-way speakers I've heard.
After 35 years it shure woun`t hurt to replace it
But be aware; when/if the amp seem to loose control of the woofers membans it`s most likely because the electrolytic cap aren`t able to handle the power. (-if you`re speakerwires aren`t too tiny, that`s another woofer-killer)
If you could afford to go up to a set of cheap SCR-caps would make a remarkable uppgrade.
Only place I could find the SCR`s was here:
http://www.slc35.com/-c-438_440_526.html?page=2&sort=2a
Yes, it really has
There were many things they didnt realise at the time
Even the very simple zobels/impedance correction wasnt used, or just beginning to enter the scene
Strange to think of how little they really knew
But sure, if you are happy with the sound as is, then dont change anything, or only very little, which was suggested before
But to put a great deal of effort in obtaining the exact values used in original design doesnt makes much sense to me
I would use closest standard value
I did tell you how to determine whether to choose the bigger or smaller standard value
I avoid multiple paralelled caps, if possible, and the same goes fore resistors
But thats just me
Maybe you just have to do what you believe is right, or you may always have a nagging doubt about it, thinking "what if", "maybe I should have", and so on
Im not here to mess with your speaker, or the good people helping you
But there are a lot of questions asked, and lots of assumptions made
Tinitus, if I were to go down that road, I'd be getting advice to completely design new crossovers to entirely replace the existing ones. How would a hi-fidelity 3-way modern loudspeaker crossover look? The ones I've seen seem to have many fewer caps and inductors, and a few resistors in there. They look much smaller to me.
Just to "recap," so to speak:
This started with -
"Hey Lucas, nice speakers. They'll sound even better if you swap out those old electrolytic caps for new ones."
Then -
"Yeah, don't use more electros for speakers. Polyprops are much better if you can afford it."
Then -
"Yeah but you'll have to compensate for the ESR with resistors."
And all along, I have tinitus, the nagging ringing noise in my head, that seems to be telling me that it's going to muck everything up, because you're not here to hear my speakers and so can't advise me! I can either not replace my caps and put up with the effects of 35 year old electros, or spend my life devoted to learning crossover technology, testing and fiddling about. Surely there's a middle way - replacing my caps and keeping the design as is (because even in 1973 Celestion knew more about speaker design than I ever will), and making a few subtle changes like compensating for ESR.
After that, we'll see.
Just to "recap," so to speak:
This started with -
"Hey Lucas, nice speakers. They'll sound even better if you swap out those old electrolytic caps for new ones."
Then -
"Yeah, don't use more electros for speakers. Polyprops are much better if you can afford it."
Then -
"Yeah but you'll have to compensate for the ESR with resistors."
And all along, I have tinitus, the nagging ringing noise in my head, that seems to be telling me that it's going to muck everything up, because you're not here to hear my speakers and so can't advise me! I can either not replace my caps and put up with the effects of 35 year old electros, or spend my life devoted to learning crossover technology, testing and fiddling about. Surely there's a middle way - replacing my caps and keeping the design as is (because even in 1973 Celestion knew more about speaker design than I ever will), and making a few subtle changes like compensating for ESR.
After that, we'll see.
Last edited:
It might not look much different from now, only minor changes
As said before, if they sound good why change anything
But sure, you could just change the big bipolar caps on woofers
But those big film caps are expencive, though the ones I linked to are affordable
Here your main problem is the need fore TWO caps each channel, 4 i total, so it gets costly nontheless
Partly the reason why I suggested to try with just one 75uf cap
Anyway, IF they do sound good as is, then even the existing bipolars might be ok, in which case theres no reason to change them fore even cheaper ones
But only you can be the judge of that
But maybe just take one step at a time
Only do whats really needed, and not get too exited to go hunting fore even more
When this stressing have gone and things have calmed down a bit you may want to try other small changes, nice and easy
One step at a time, or it will be confused and you wont know what works and what doesnt
At least dont let the saving on shipping cost determine your choises
Last edited:
Parallel resistors
Indeed one should - if the particular application is one where a lot of power will be dissipated.
Higher power ratings in Welwyn W** series are:-
W23 - 10watt@25*C <---> 9watt@70*C
W24 - 14watt@25*C <---> 12watt@70*C
These are longer, thus more inductive, though of course inductance is reduced when resistors are connected in parallel.
Inductance is not a problem in some applications,
and electrolytic caps usually have larger inductance than wirewound resistors.
In the particular Celestion 44 bass application I specified the 1ohm <---> 3.9ohm WP4S resistors for
the current is sufficiently low there for the voltage drop to be low across those low values of resistance, and thus not a lot of power to dissipate.
If I was using 4.7ohms and higher I would use higher powered resistors there.
For Zobels -{impedance correction}- in woofer circuits I would usually use at least a 10watt resistor for Hi-Fi, and greater in Public Address systems loudspeakers.
I am wary of parallel resistors because if one of them goes slightly faulty during operation there is no obvious warning until it fails further
and the other resistor is carrying most of the current and thus the voltage has changed significantly then.
If only one resistor is used, then if it starts to fail, usually there is an audible effect sooner than when a parallel pair.
I use pairs only when necessary, and always in equal, or close to equal, resistances so as to close to equally distribute the dissipation.
Indeed one should - if the particular application is one where a lot of power will be dissipated.
Higher power ratings in Welwyn W** series are:-
W23 - 10watt@25*C <---> 9watt@70*C
W24 - 14watt@25*C <---> 12watt@70*C
These are longer, thus more inductive, though of course inductance is reduced when resistors are connected in parallel.
Inductance is not a problem in some applications,
and electrolytic caps usually have larger inductance than wirewound resistors.
In the particular Celestion 44 bass application I specified the 1ohm <---> 3.9ohm WP4S resistors for
the current is sufficiently low there for the voltage drop to be low across those low values of resistance, and thus not a lot of power to dissipate.
If I was using 4.7ohms and higher I would use higher powered resistors there.
For Zobels -{impedance correction}- in woofer circuits I would usually use at least a 10watt resistor for Hi-Fi, and greater in Public Address systems loudspeakers.
I am wary of parallel resistors because if one of them goes slightly faulty during operation there is no obvious warning until it fails further
and the other resistor is carrying most of the current and thus the voltage has changed significantly then.
If only one resistor is used, then if it starts to fail, usually there is an audible effect sooner than when a parallel pair.
I use pairs only when necessary, and always in equal, or close to equal, resistances so as to close to equally distribute the dissipation.
For the Celestion 44 design
Hello Tony,
Celestion didn't bother with additional impedance correction circuits often then,
and few manufacturers do unless it is absolutely necessary, and does not cause the cross-over to be more complex than necessary.
When high-slope filters are needed, and thus a lot of parallel connected components, there are audible colourations caused by caps resonating with inductors, including with the inductance of the driver's voice-coil.
Such types of multi-path circuit cross-overs were used by B&W and KEF in several of their loudspeakers during the 1970s and 1980s, and most of them sounded worse when used with their moving-coil drivers than the simpler circuits used by Celestion with theirs.
Too much Science was incompetantly applied, and not sufficient listening to the results, but trying to impress tech-head buyers by showing photos of impressive looking cross-over circuitry !
{there are a few classic KEF and B&W models than do not sound bad, but not all their old models.}
The woofers used in Celestion's 44 and 66 models have significant rise in output as the frequency response increases, thus the large parallel caps,
in conjunction with the large series inductors, to pull down the output rise which starts from a little above 200Hz with those woofers.
Those woofers were developed from Public Address and Electric Guitar 12" speakers.
They were not like other developed for bass only applicable 12" drivers.
Celestion's main business was manufacturing drivers for Public Address and Electric Instruments' amplifiers then, and it still is to large degree.
But there are significant differences between the 12" in the 44 and those used in Marshall's guitar and bass cabinets,
and the Hi-Fi versions would not live long if driven hard by Rock Musicians' amplifiers -{so don't try it !}.
One could redesign the bass filters in the x-over and use a Zobel in parallel with the woofer,
but then one would have to change the values of the inductors and of the cap to ground in between the inductors.
Would the final result be any better ?
It would be different, but as a lot of listeners like the basic sound of these old 44s and 66s -{and see the ridiculously high prices that some are selling for}- then why change the design ?
What I am attempting to do here is to restore the Celestions to as close as possible to their original liked sound, but with using modern longer life components -{when the budget allows}.
If one wants to Improve on the original design then one will have to measure ALL the drivers and redesign the x-over.
There is some Zobel type Impedance Correction present in the bass filter, and that is the ESR in the 72uF electro-caps.
{Celestion's design engineers knew about ESR.}
That obviously worked well enough when those caps were new, and similar type new electro caps will have similar ESR now,
-{but not all new bipolar electros do now, thus why I specified the particular ones in this Thread, and NOT the Elnas and Nichicons which are intended for lower impedance applications}.
If one wants long life, then install poly caps, but with the necessary series-connected resistances to keep the relative impedances correct in this crossover.
*******************
I'll get to your inductor winding question next time Lucas.
Hello Tony,
Celestion didn't bother with additional impedance correction circuits often then,
and few manufacturers do unless it is absolutely necessary, and does not cause the cross-over to be more complex than necessary.
When high-slope filters are needed, and thus a lot of parallel connected components, there are audible colourations caused by caps resonating with inductors, including with the inductance of the driver's voice-coil.
Such types of multi-path circuit cross-overs were used by B&W and KEF in several of their loudspeakers during the 1970s and 1980s, and most of them sounded worse when used with their moving-coil drivers than the simpler circuits used by Celestion with theirs.
Too much Science was incompetantly applied, and not sufficient listening to the results, but trying to impress tech-head buyers by showing photos of impressive looking cross-over circuitry !
{there are a few classic KEF and B&W models than do not sound bad, but not all their old models.}
The woofers used in Celestion's 44 and 66 models have significant rise in output as the frequency response increases, thus the large parallel caps,
in conjunction with the large series inductors, to pull down the output rise which starts from a little above 200Hz with those woofers.
Those woofers were developed from Public Address and Electric Guitar 12" speakers.
They were not like other developed for bass only applicable 12" drivers.
Celestion's main business was manufacturing drivers for Public Address and Electric Instruments' amplifiers then, and it still is to large degree.
But there are significant differences between the 12" in the 44 and those used in Marshall's guitar and bass cabinets,
and the Hi-Fi versions would not live long if driven hard by Rock Musicians' amplifiers -{so don't try it !}.
One could redesign the bass filters in the x-over and use a Zobel in parallel with the woofer,
but then one would have to change the values of the inductors and of the cap to ground in between the inductors.
Would the final result be any better ?
It would be different, but as a lot of listeners like the basic sound of these old 44s and 66s -{and see the ridiculously high prices that some are selling for}- then why change the design ?
What I am attempting to do here is to restore the Celestions to as close as possible to their original liked sound, but with using modern longer life components -{when the budget allows}.
If one wants to Improve on the original design then one will have to measure ALL the drivers and redesign the x-over.
There is some Zobel type Impedance Correction present in the bass filter, and that is the ESR in the 72uF electro-caps.
{Celestion's design engineers knew about ESR.}
That obviously worked well enough when those caps were new, and similar type new electro caps will have similar ESR now,
-{but not all new bipolar electros do now, thus why I specified the particular ones in this Thread, and NOT the Elnas and Nichicons which are intended for lower impedance applications}.
If one wants long life, then install poly caps, but with the necessary series-connected resistances to keep the relative impedances correct in this crossover.
*******************
I'll get to your inductor winding question next time Lucas.
Last edited:
I use 5watt wirewound cement fore tweeter series attenuation, no problems, XT25 sounds just sweet
Also 5watt ditto on midrange zobel, no problems
5watt does seem to sound better used in particular these two positions
20watt wirewound cement everywhere else
Also, 20watt seem to sound better in all these other low frequency "high power" positions
None of them ever gets hot or even slightly varm, so no problems at all
I have no idea about durability of film resistors
As said, I will never accept multiple paralelled components
They mostly sounds worse than choosing the closest standard value, if not too far off
The trick is to choose the best closest standard value, which I admit takes some experimenting
Thats all
Also 5watt ditto on midrange zobel, no problems
5watt does seem to sound better used in particular these two positions
20watt wirewound cement everywhere else
Also, 20watt seem to sound better in all these other low frequency "high power" positions
None of them ever gets hot or even slightly varm, so no problems at all
I have no idea about durability of film resistors
As said, I will never accept multiple paralelled components
They mostly sounds worse than choosing the closest standard value, if not too far off
The trick is to choose the best closest standard value, which I admit takes some experimenting
Thats all
Last edited:
Using two equally sized caps value in paralel with inductors
Been there, done that, and its a never again, and noone else does, if they have brains
Pardon me if wrong, but are there not TWO equally sized value 72uf in paralel on woofer ?
Its pretty daft
If it at least dont ruin response too much, it certainly will screw up phase
I dont understand how they left it like that
I bet almost anything else works better
But at the time I believe it was one of those slick ideas some believed in
Man, there have even been very skilled guys who made 2ways 12db with equally sized mirror imaged reversed values on woofer and tweeter, or whatever we should call it
Last edited:
Tinitus, enough already! It's not helpful. It's not fun. It's pretty discouraging.
Have you ever heard Celestion Ditton 66 or 44 speakers? They are VERY good. You can bet all you like how awful they might sound, but the fact is, Alan and I own some, and we like them a hell of a lot, as does everybody that's heard mine. The bass is in fact highly renowned on both models. I'm trying to get the best out of mine, and Alan is helping out with his FIRST HAND EXPERIENCE of rejuvenating this particular combo of crossover and drive units, and it's the kind of advice that makes sense to me.
Have you ever heard Celestion Ditton 66 or 44 speakers? They are VERY good. You can bet all you like how awful they might sound, but the fact is, Alan and I own some, and we like them a hell of a lot, as does everybody that's heard mine. The bass is in fact highly renowned on both models. I'm trying to get the best out of mine, and Alan is helping out with his FIRST HAND EXPERIENCE of rejuvenating this particular combo of crossover and drive units, and it's the kind of advice that makes sense to me.
Hi Alan, I've only played around with impedance compensation on my current project, and didn't care for it that much. Sorry if my post came across as supporting the idea of putting in impedance compensation, that was not the intent.. The intent was simply to state that adding it would more than likely change the sound, and possibly also necessitate making other crossover changes. Something I didn't think Lucas really wanted to do.
Tony.
Thats a completely total misunderstanding
I said, "I bet almost anything else works better", which was regarding using double 72uf caps
I thought that was quite clear
I have never commented on the sound of your speakers as such, or said anything about sounding awful
But if you want my silence
fine, no problemTake care
Last edited:
Tinitus, I din't mean to be rude, but maybe I was, so I'm sorry for that.
I find that Alan has personal experience of these units and has done what I am trying to do, so he can be less theoretical in his advise about what to expect, and I find that reassuring.
You place more doubts and questions in the air, which may be based on your extensive experience but these questions can't be answered, as you have said yourself, without listening, and all of that just leads to a negative feeling that there's no point to it all and it will make a mess of my speakers. That's not what I need to hear, sorry! Maybe you're right. Maybe they're a terrible design and I should really change them a lot, like removing caps completely and changing values, but you're contradicting your earlier advice not to change much, so it gets very confusing and not helpful in the end.
Again, sorry if I was rude. I know you were trying to help, and I may look back later and see that you were right about some things.
Lucas
I find that Alan has personal experience of these units and has done what I am trying to do, so he can be less theoretical in his advise about what to expect, and I find that reassuring.
You place more doubts and questions in the air, which may be based on your extensive experience but these questions can't be answered, as you have said yourself, without listening, and all of that just leads to a negative feeling that there's no point to it all and it will make a mess of my speakers. That's not what I need to hear, sorry! Maybe you're right. Maybe they're a terrible design and I should really change them a lot, like removing caps completely and changing values, but you're contradicting your earlier advice not to change much, so it gets very confusing and not helpful in the end.
Again, sorry if I was rude. I know you were trying to help, and I may look back later and see that you were right about some things.
Lucas
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.