Your digital crossover is changing phase, the effect in phase of an additional capacitor is subtracted from it. The combined phase will be the same whether you use the capacitor or not.
It remains to ask, can you start with a given response and have your processor to turn it into another wanted response?
It remains to ask, can you start with a given response and have your processor to turn it into another wanted response?
To explain a little better : my speakers are for home use not disco so they rarely if ever will be used at very high volume , and used for normal music reproduction . They are fairly big to get good low frequency reproduction and will most likely be assisted by a subwoofer in the future.
The high supply rail voltages of the bass amps being used for the midrange and tweeters in unison with the active filters should assure the tweeter and , to a lesser extend , the midrange are NEVER driven to clipping so no problem there. The midrange and tweeter frequencies are not "riding" on the bass signal voltages so their signals should always be clean. No clipping = no undue stress on their voice-coils. Both tweeters and midrange are dome type , of 1" and 2" respectively.
Under these conditions what is left ? Effectively protecting the tweeters and possibly the midrange and their voice-coils from DC in case of failures.
to answer AllenB : no processor , just active linkwitz-rileys 24dB/octave with exremely good ops and carefully selected matching capacitors and resistors
The high supply rail voltages of the bass amps being used for the midrange and tweeters in unison with the active filters should assure the tweeter and , to a lesser extend , the midrange are NEVER driven to clipping so no problem there. The midrange and tweeter frequencies are not "riding" on the bass signal voltages so their signals should always be clean. No clipping = no undue stress on their voice-coils. Both tweeters and midrange are dome type , of 1" and 2" respectively.
Under these conditions what is left ? Effectively protecting the tweeters and possibly the midrange and their voice-coils from DC in case of failures.
to answer AllenB : no processor , just active linkwitz-rileys 24dB/octave with exremely good ops and carefully selected matching capacitors and resistors
Last edited:
Ok, different situation. Your thoughts that phase will only be worse are uncertain. It may actually be better, that is not clear.
Hello AllanB
I am after perfection not starting air-planes so yes , different situation . My best protective circuitry concerning overload are my neighbours .
Both the tweeters and midrange are vintage´s finest from the seventies and irreplaceable hence my wish to protect them at all costs. Their sound is cristal clear with no hint at any sharpness.
I am after perfection not starting air-planes so yes , different situation . My best protective circuitry concerning overload are my neighbours .
Both the tweeters and midrange are vintage´s finest from the seventies and irreplaceable hence my wish to protect them at all costs. Their sound is cristal clear with no hint at any sharpness.
Last edited:
So you say lowest frequency is 3000Hz Tweeter has 4 Ohms
Then the cap would be something like 13.5uf
12uf will be fine.. get some WIMA (cost 2 Euros) or similar so cutoff or lowest frequency will be somewhat higher similar to 3.32khz if you use a larger cap, then frequency will come down.
And frequency will be 2.65 Khz with 15uf
Then the cap would be something like 13.5uf
12uf will be fine.. get some WIMA (cost 2 Euros) or similar so cutoff or lowest frequency will be somewhat higher similar to 3.32khz if you use a larger cap, then frequency will come down.
And frequency will be 2.65 Khz with 15uf
Last edited:
You're in the right place. I'd suggest you should measure the phase difference between your woofer and tweeter, with filters in place, then use the added capacitor strategically to match them up..
there is obviously a changing phase as frequency changes but with the filters being Linkwitz-Riley there is no difference in phase between tweeter/midrange and midrange/bass at any given frequency , so the output of the filters give a perfectly flat response with the speakers down 6 dB at the two cross-over-frequencies and therefore adding up to flat response. Possible difference in sensitivity can simply be adjusted with potentiometers at filter outputs. The fact that all outputs of a 24 dB Linkwitz-Riley filter are perfectly in phase is one of the beauties of said filters.
This is why a capacitor in series with the tweeter needs to be big enough not to introduce any change in phase at all frequencies reproduced by the tweeter , say above 1000Hz .At this frequency the signal reaching the tweeter is negligible so a change in phase is as well..
My problem here is I do not know how to calculate this correcly for a tweeter impedance of 8 ohms
This is why a capacitor in series with the tweeter needs to be big enough not to introduce any change in phase at all frequencies reproduced by the tweeter , say above 1000Hz .At this frequency the signal reaching the tweeter is negligible so a change in phase is as well..
My problem here is I do not know how to calculate this correcly for a tweeter impedance of 8 ohms
Last edited:
We love mathematical truths 😉 but acoustics ain't electronics and drivers, enclosures and location of acoustical centers do mess big time with your ideal phase prerequisite.
Edit: same applies for levels by the way. Any driver acts as a natural bandpass filter. With it's phase and level behavior. You have to take this into account when designing crossovers. Like Linkwitz already did back in the seventies IIRC.
Edit: same applies for levels by the way. Any driver acts as a natural bandpass filter. With it's phase and level behavior. You have to take this into account when designing crossovers. Like Linkwitz already did back in the seventies IIRC.
Last edited:
One builds said filter to get a flat response I´d say.
I know perfection does not exist but one can do ones best to get close , which is what I am trying
Try conecting one speaker out of phase see what the phase change does.I simply am trying to avoid mistakes wherever possible. And was hoping to get the info and help here.
I know perfection does not exist but one can do ones best to get close , which is what I am trying
Try conecting one speaker out of phase see what the phase change does.I simply am trying to avoid mistakes wherever possible. And was hoping to get the info and help here.
I am perfectly aware of all this but why should it keep me from doing the best I can wherever possible?
You will have help from peoples here this is sure.
But most of the time it doesn't take the form of a straigth answer as it usually rise interesting points ( and food for thoughts). 😉
I think that what are Allen, Markbakk (and others) trying to tell you is that what does matter is the way the whole system react and not nescesseraly the perfection of a filter implemented.
Iow, the way your drivers are going to interact in their final location will not nescessarely be the one your implemented filter suggest.
Center to center distance, eventual offset of emmissing point,native response of driver, width and shape of baffle,etc,etc,... will change the behavior you could expect from your choosen filter type.
What does matter is the final acoustic slope not the way it is implemented electronicaly ( to a degree).
So let's say you want to implement your capacitor (for protection concern) you could choose to make its value effective at 3khz ( so a 6db/octave hp) and then complement the filtering through a 18db cell implemented electronicaly.
In practice you'll have your 24db/octave slope and if choosen wisely your electronic cell will make it a LR.
But here again nothing tells you you won't have surprise from the interaction with your mid.
So all in all either you can take measurements and tweak your filters to have the final type you have choosen either you can simulate the behavior of your loudspeaker through Vituix or Lspcad,... and adapt your electronic circuit to your goal.
But most of the time it doesn't take the form of a straigth answer as it usually rise interesting points ( and food for thoughts). 😉
I think that what are Allen, Markbakk (and others) trying to tell you is that what does matter is the way the whole system react and not nescesseraly the perfection of a filter implemented.
Iow, the way your drivers are going to interact in their final location will not nescessarely be the one your implemented filter suggest.
Center to center distance, eventual offset of emmissing point,native response of driver, width and shape of baffle,etc,etc,... will change the behavior you could expect from your choosen filter type.
What does matter is the final acoustic slope not the way it is implemented electronicaly ( to a degree).
So let's say you want to implement your capacitor (for protection concern) you could choose to make its value effective at 3khz ( so a 6db/octave hp) and then complement the filtering through a 18db cell implemented electronicaly.
In practice you'll have your 24db/octave slope and if choosen wisely your electronic cell will make it a LR.
But here again nothing tells you you won't have surprise from the interaction with your mid.
So all in all either you can take measurements and tweak your filters to have the final type you have choosen either you can simulate the behavior of your loudspeaker through Vituix or Lspcad,... and adapt your electronic circuit to your goal.
Last edited:
Thx krivium!
Let me put it otherwise: how do you acquire that the acoustic output of your speaker drivers behave like perfectly tuned LR function filters? Not by feeding them electric LR functions for sure...
Let me put it otherwise: how do you acquire that the acoustic output of your speaker drivers behave like perfectly tuned LR function filters? Not by feeding them electric LR functions for sure...
We all know that a room with bad acoustics can ruin the sound of even the best system worldwide , like say a square livingroom with tiled walls and so on ...Haha!
So please lets forego that part of the drama and try to concentrate on building the best possible active 3-way speaker cabinets . Flat response , no or only minimal colouring ot the reproduced sound and power of 100 to 130 watts RMS , more than enough to listen to good music at home without ending up deaf.
So far I have the electronics as perfect as I can make them towards the above goals.
I have a perfectly flat response right to the outputs of the three amps.
I have a reasonable reserve for the midrange and a very good one for the highs so they should never be driven into clipping sparing their coils and my ears from the nefarious effects this produces , and totally independent of the bass , and possible occasional clipping in the woofers .
At the same time I avoid any interaction between the different speakers which would be present using a passive crossover , avoiding the distortions and powerloss inherent to the latter. Which can , by the way , ruin an otherwise good system quite easily. not in vain good crossovers are very expensive and must be tailored to the various chassis they serve. Not only is my aproach not much more expensive ( my work doesnt count in this ) but it is independent of the individual drivers response-curves as well and therefore minimizes any diference s there. All thats necesary is to adjust the amps so bass/midrange and midrange/tweeter reproduce exactly the same volume at their corresponding Xover to be back in business with the best possible flat frequency-response throughout.
Please let me get back to the capacitor intended to protec the tweeters but without modifying the response of the above.
All I need to figure out how big it has to be not to cause any shift in phase above 1000 Hz for a load of 8Ohm
So please lets forego that part of the drama and try to concentrate on building the best possible active 3-way speaker cabinets . Flat response , no or only minimal colouring ot the reproduced sound and power of 100 to 130 watts RMS , more than enough to listen to good music at home without ending up deaf.
So far I have the electronics as perfect as I can make them towards the above goals.
I have a perfectly flat response right to the outputs of the three amps.
I have a reasonable reserve for the midrange and a very good one for the highs so they should never be driven into clipping sparing their coils and my ears from the nefarious effects this produces , and totally independent of the bass , and possible occasional clipping in the woofers .
At the same time I avoid any interaction between the different speakers which would be present using a passive crossover , avoiding the distortions and powerloss inherent to the latter. Which can , by the way , ruin an otherwise good system quite easily. not in vain good crossovers are very expensive and must be tailored to the various chassis they serve. Not only is my aproach not much more expensive ( my work doesnt count in this ) but it is independent of the individual drivers response-curves as well and therefore minimizes any diference s there. All thats necesary is to adjust the amps so bass/midrange and midrange/tweeter reproduce exactly the same volume at their corresponding Xover to be back in business with the best possible flat frequency-response throughout.
Please let me get back to the capacitor intended to protec the tweeters but without modifying the response of the above.
All I need to figure out how big it has to be not to cause any shift in phase above 1000 Hz for a load of 8Ohm
Last edited:
A bulb is a resistor in series with the tweeter , and a variable one at that . While it does protect the tweeter it is a limiter and compressor as well and probably the right thing for a PA-System but not for what I aim at . Then I might as well use a crowbar and a fuse , something I would only consider as a last resort if all else fails.
Either I don’t understand what you wrote or you are making big thinking errors about electrodynamic transducers and how sound sources work. I for now think the latter is the case.
I think the first is the case and maybe my comments are not detailed enough. Sorry
Originally Posted by MalleMike
I have a perfectly flat response right to the outputs of the three amps.
I have three amps connected to a three-way Linkwitz-Riley filter with 24 dB xovers. Summing their outputs the frequency-response it totally flat from 20Hz to 20 KHz.
Three potentiometers at the filters outputs allow to adjust for drivers of different sensitivity to equalize the drivers acoustic output at the xover frequencies. No complicated redesigning os passive crossover needed . Of course this does not compensate for the response curves of the drivers themselves . That is a different matter. But it allows me to use drivers of different sensitivity by simply adjusting one or more of the amps. Say a woofer with 90dB , a midrange with 96 dB and a tweeter with 108 dB.
Originally Posted by MalleMike
I have a perfectly flat response right to the outputs of the three amps.
I have three amps connected to a three-way Linkwitz-Riley filter with 24 dB xovers. Summing their outputs the frequency-response it totally flat from 20Hz to 20 KHz.
Three potentiometers at the filters outputs allow to adjust for drivers of different sensitivity to equalize the drivers acoustic output at the xover frequencies. No complicated redesigning os passive crossover needed . Of course this does not compensate for the response curves of the drivers themselves . That is a different matter. But it allows me to use drivers of different sensitivity by simply adjusting one or more of the amps. Say a woofer with 90dB , a midrange with 96 dB and a tweeter with 108 dB.
Last edited:
Let me say a few things concerning passive cross-overs. First they reduce the power which actually reaches the speakers. Second they introduce their own distortions. Third they allow the speakers to influence each other in unforseeable ways. And forth they totally mess up any damping on the part of the amp . .The only coils that do not go into saturation are airwound but these are big and and have considerable resistances so cause powerloss, and are expensive . The magnetic fields of different coils can and do influence each other
For what all this implies and costs I prefer active crossovers at line level and Bi- or tri-Amping. If anyone has never listened to a good active three way please do .I believe it will convince you. Building good amps is not as difficult nor as expensive as 10 to 20 years ago
For what all this implies and costs I prefer active crossovers at line level and Bi- or tri-Amping. If anyone has never listened to a good active three way please do .I believe it will convince you. Building good amps is not as difficult nor as expensive as 10 to 20 years ago
Last edited:
all good points, but why such long confusing thread when all you had to do is to put a cap on the tweeter (or a mid if you want to protect that too), which as you say, you do not know how to calculate, while google is full of links to such things
2-Way Crossover Calculator / Designer
2-Way Crossover Calculator / Designer