I opened the speakers up to check the internals, see attached;; now if I have to connect the amps direct to the speakers, there is a pair of my amps that has the Vellemen DC protection kit installed so I guess this should be the one to connect to the tweeter directly, and use the other pair of amps for the woofers, I am still a little uncertain of how to wire the woofers after the cross over pcb removed and replace with new cup, please advise.
Thanks,
Chris
Thanks,
Chris
Attachments
An MTM is a two way, the woofer are paralleled, of course you can use three amps and that's what I did but there is only one crossover point and that makes it a biamp and not a triamp.
I think we can safely assume that everybody knows that an active system has no passive xover components between the anp and the driver anywhere in sight.
I can definitely confirm what many have said, including NP:
4th order Xovers sound like crap. That is not stopping many from promoting their use including some pretty popular designer in both active and passive xovers.
Makes ME question what people out there are REALLY hearing and what they consider "good sound".
I think we can safely assume that everybody knows that an active system has no passive xover components between the anp and the driver anywhere in sight.
I can definitely confirm what many have said, including NP:
4th order Xovers sound like crap. That is not stopping many from promoting their use including some pretty popular designer in both active and passive xovers.
Makes ME question what people out there are REALLY hearing and what they consider "good sound".
Crossovers being phase coherent and having complementary amplitude response don't sound in any special way, being the order 2, 3 or 4
To test it simply try a simple double blind test [please don't cheat, if you know what you are testing then test is no longer blind and has no value, you must be blind]
Option 1 is to pass the source signal through a state-variable 4th order Linkwitz-Riley crossover with unity gain, then sum the outputs with unity gain and fed this signal to the amplifier [one filter for each channel is required for stereo]
Option 2 is to fed the amplifier directly from the source
If the crossover circuit works properly, I doubt you would be able to distinguish between the presende and the absence of the crossover in a double blind test [as it works as a simple phase shifter after summing its outputs and it has been demonstrated long time ago that there is no way we can hear these 'absolute phase shifts' when they are in the order of a wavelenght or a few wavelenghts]
Of course, a badly designed or aligned 4th crossover [there are almost infinite options for this and infinite mistakes are possible] will create a sharp dip or peak in the frequency response at the crossover region [or even combinations of both things!, like peak-dip-peak or dip-peak-dip] making anything soud disgusting, but this is the fault of the designer, not of 4th order crossovers bi itself
For any people without skill or measuring equipment to design, test and adjust properly 4th crossovers, 2nd order crossovers are advisable because they are easier to use and far more forgiving on design faults
Also, 2nd order tends to be enough for Hi-Fi drivers because they usually have wide frequency response and very progressive rolloffs [also 2nd order gives wide crossover region and thus wide cancellation zone out of axis and less reflected sound provided you optimize the system for this with proper electronic-phase-shifting or driver-placement delays]
4th crossovers are only advisable when you need a sharp cutoff, for example when using horns and compression drivers or when trying to get a driver to play a frequency range as wide as possible without showing possible driver abnormalities out of band
Just use the minimum order and the minimum number of drivers capable of matching the performance you want
To test it simply try a simple double blind test [please don't cheat, if you know what you are testing then test is no longer blind and has no value, you must be blind]
Option 1 is to pass the source signal through a state-variable 4th order Linkwitz-Riley crossover with unity gain, then sum the outputs with unity gain and fed this signal to the amplifier [one filter for each channel is required for stereo]
Option 2 is to fed the amplifier directly from the source
If the crossover circuit works properly, I doubt you would be able to distinguish between the presende and the absence of the crossover in a double blind test [as it works as a simple phase shifter after summing its outputs and it has been demonstrated long time ago that there is no way we can hear these 'absolute phase shifts' when they are in the order of a wavelenght or a few wavelenghts]
Of course, a badly designed or aligned 4th crossover [there are almost infinite options for this and infinite mistakes are possible] will create a sharp dip or peak in the frequency response at the crossover region [or even combinations of both things!, like peak-dip-peak or dip-peak-dip] making anything soud disgusting, but this is the fault of the designer, not of 4th order crossovers bi itself
For any people without skill or measuring equipment to design, test and adjust properly 4th crossovers, 2nd order crossovers are advisable because they are easier to use and far more forgiving on design faults
Also, 2nd order tends to be enough for Hi-Fi drivers because they usually have wide frequency response and very progressive rolloffs [also 2nd order gives wide crossover region and thus wide cancellation zone out of axis and less reflected sound provided you optimize the system for this with proper electronic-phase-shifting or driver-placement delays]
4th crossovers are only advisable when you need a sharp cutoff, for example when using horns and compression drivers or when trying to get a driver to play a frequency range as wide as possible without showing possible driver abnormalities out of band
Just use the minimum order and the minimum number of drivers capable of matching the performance you want
chris ma :
Try to figure out if both drivers are fed by the same filter circuit [effectively connected in paralell] or by independent filter circuits
Also, in case there are independen circuits, try to find out if they are identical or use different component values and/or layouts
If there are two midwoofers, even being identical, but the enclosure is designed to place it vertically, then both drivers may not be receiving the same signal, for example, to compensate baffle step or to equalize the system in a particular way, these kind of tricks are common
Passive filters sometimes have responses more complex than expected and thus they can't be replaced with generic active filters, custom active filters based in the design of the passive ones have to be used [using the same equalizing tricks to get the same response] except when you don't like the original response of the system at all
Try to figure out if both drivers are fed by the same filter circuit [effectively connected in paralell] or by independent filter circuits
Also, in case there are independen circuits, try to find out if they are identical or use different component values and/or layouts
If there are two midwoofers, even being identical, but the enclosure is designed to place it vertically, then both drivers may not be receiving the same signal, for example, to compensate baffle step or to equalize the system in a particular way, these kind of tricks are common
Passive filters sometimes have responses more complex than expected and thus they can't be replaced with generic active filters, custom active filters based in the design of the passive ones have to be used [using the same equalizing tricks to get the same response] except when you don't like the original response of the system at all
To summarize in a very general way a couple of the longer posts:
Loudspeaker designers use a number of techniques to compensate for phase differences in the XO region. When replaceing a passive XO with an active XO may one may discover that removing the passive circuit exposes the effect of some of the compensation techniques. These may be difficult to deal with.
My personal generalization:
Well designed, top quality conventional loudspeakers don't get much benefit from an active XO and may even present special problems. The best candidates for active XO are DIY loudspeakers, loudspeakers designed with an active XO option in mind, decent budget speakers or any loudspeaker where you know the drivers have flat, overlapping response curves and are all mounted on the same plane.
Loudspeaker designers use a number of techniques to compensate for phase differences in the XO region. When replaceing a passive XO with an active XO may one may discover that removing the passive circuit exposes the effect of some of the compensation techniques. These may be difficult to deal with.
My personal generalization:
Well designed, top quality conventional loudspeakers don't get much benefit from an active XO and may even present special problems. The best candidates for active XO are DIY loudspeakers, loudspeakers designed with an active XO option in mind, decent budget speakers or any loudspeaker where you know the drivers have flat, overlapping response curves and are all mounted on the same plane.
tiroth said:
The Vellemen speaker protection circuit is very effective and fast, it also has a turn on delay about 5-7 seconds before it connects to the output as well. I am for sure I will use this pair of amps for the tweeters.
It is the woofers connections that I am not sure, according to Eva it seems that the stock speaker is a 3 way and now if I connect the woofers to one amp, in effect I will change a 3 way to 2 way....I am just a little bit puzzled..
Thanks,
Chris
chris ma:
Could you post detailed schematics and/or pictures of those passive filters?
[Component values are useful too if available]
If I'm not wrong, CX-3400 has phase shift/delay capability and you may have to play somewhat with it to tune up the system
Also, feel free to play with gain, crossover freq and relative polarities until you find a combination giving optimum sum of tweeter and midwoofers
Follow this rule : The better the sum you are getting , the bigger the cancellation you will get when changing tweeter polarity [this is position-dependant so you sould perform all tests from the usual listening position and with the usual enclosure placement]
A graphic equalizer of 15 bands or more may be useful to compensate baffle step and low fequency peaks and dips present in every room due to its acoustics
sam9 :
I agree that passive crossovers with very clever design are already almost optimun and there is not much benefit from active filtering except maybe a reduction in THD due to the lack of linearity of iron-core inductors and voice coil inductance [near zero crossing and near saturation as can be seen in the oscilloscope], and the removal of cross couplig between air core inductors
But phase compensation is by far easier and cheaper to implement with active circuits and there are lots of passive systems without optimum filtering so with some active filtering knowledge you can easily improve any passive system with low to medium complexity crossovers or tailor it to your needs [delays and hot-spot placement are very hard to get without active components]
PD: In my opinion, any expensive loudspeaker should come with its own otpimized active filter as a part of the product
Could you post detailed schematics and/or pictures of those passive filters?
[Component values are useful too if available]
If I'm not wrong, CX-3400 has phase shift/delay capability and you may have to play somewhat with it to tune up the system
Also, feel free to play with gain, crossover freq and relative polarities until you find a combination giving optimum sum of tweeter and midwoofers
Follow this rule : The better the sum you are getting , the bigger the cancellation you will get when changing tweeter polarity [this is position-dependant so you sould perform all tests from the usual listening position and with the usual enclosure placement]
A graphic equalizer of 15 bands or more may be useful to compensate baffle step and low fequency peaks and dips present in every room due to its acoustics
sam9 :
I agree that passive crossovers with very clever design are already almost optimun and there is not much benefit from active filtering except maybe a reduction in THD due to the lack of linearity of iron-core inductors and voice coil inductance [near zero crossing and near saturation as can be seen in the oscilloscope], and the removal of cross couplig between air core inductors
But phase compensation is by far easier and cheaper to implement with active circuits and there are lots of passive systems without optimum filtering so with some active filtering knowledge you can easily improve any passive system with low to medium complexity crossovers or tailor it to your needs [delays and hot-spot placement are very hard to get without active components]
PD: In my opinion, any expensive loudspeaker should come with its own otpimized active filter as a part of the product
The woofers are run as a 2.5 way system so to properly multiamp you will need 6 amps or leave the bottom choke in your drawing in series with the 2nd (bottom probably) driver. 1st order XO on these. And the series Rs have to be there to change the system Q, i can't see any other reason for them.
The tweeter is a little trickier with the R shunting the series cap. And the R in series with the choke that shunts the T is giving some EQ too. Someone better versed than i in passive XOs will have to comment on what is happening.
dave
The tweeter is a little trickier with the R shunting the series cap. And the R in series with the choke that shunts the T is giving some EQ too. Someone better versed than i in passive XOs will have to comment on what is happening.
dave
So it seems the Mirage is not easily convert to active bi-amps. OK I guess I have to bite the bullet and order the Jordan 92s, but before I do that what tweeter I can use ? so I may as well order it the same time from Creative Sound in one go this afternoon?
I still want to bi-amp,
which way would be better with the J92s
a tweeter with J92
or
a 12inch woofer with J92
Sealed box or open buffle is easier for my limited wood working skill and tools
Chris
I still want to bi-amp,
which way would be better with the J92s
a tweeter with J92
or
a 12inch woofer with J92
Sealed box or open buffle is easier for my limited wood working skill and tools
Chris
If you can do only 2-way with the JX92, add a woofer. If really wants to be the heart of a mostly FR system with a W on the bottom (<125Hz) and a T on the top (>7-10k) ... i've often recommended the Aurum Cantus ribbons as superTs, and Bob just recently made it impossible for me not to buy a pr.
If you just want to do a 2-way, the JX150 or JX125 with a JX53 makes more sense (1st order XO between 200 & 4k) Check out Rune's project at my site for a JX150/JX53 design)
dave
Note: the JX125 TTL with full details won't fit within upload guidlines. (design inspired by Woody, done with MJKs tables and amazingly close to a KEF B200 line i designed in 1979)
If you just want to do a 2-way, the JX150 or JX125 with a JX53 makes more sense (1st order XO between 200 & 4k) Check out Rune's project at my site for a JX150/JX53 design)
dave
Note: the JX125 TTL with full details won't fit within upload guidlines. (design inspired by Woody, done with MJKs tables and amazingly close to a KEF B200 line i designed in 1979)
Attachments
The filter of the tweeter appears to work as 12dB/oct from 1800Hz to about 700Hz and 6dB/oct from 700Hz to 20Hz, then no filtering at all [but 52dB attenuation]
It also appears to be compensating for raised impedance around 1800Hz due to tweeter resonance since I had to put about 15 ohms as tweeter impedance and 1mH as inductor value to get -6dB and 90º at about 1,8kHz
R1 provides about 1dB of attenuation [providing tweeter impedance is about 7 ohms] and R3 is responsible for turning the filter from 12dB/oct into 6dB/oct
When going active you no longer need to compensate for raised impedance near Fs of the driver, and if you use a 4th order crossover, compensating for tweeter rolloff shouln't be as critical since crossover region isn't as wide
Woofer filter appears to be first order for both drivers, you can think of the coupled inductor as an alternative to two independent inductors of different value, but with lower cost, space saving and making profit of mutual coupling instead of having to avoid it placing one inductor 20-40cm away from the other
This coupled inductor provides an early rolloff for one of the drivers and crosses the other driver with the tweeter at a higher frequency, the purpose of this practice is to compensate for baffle step making the lower driver start to work when the higher driver loses the acoustical load provided by the baffle [it happens at a frequency lower than 1800Hz so just attenuating the tweeter doesn't work, it would leave the range from baffle step frequency to 1800Hz boosted]
To compensate for baffle step you would need to add some equalization at the LF ouput of the active crossover, things as simple as a R in series and then an RC to ground may work [this is called a shelving filter, its response looks like this : ¯¯¯\____
I recommend getting some program capable of simulating simple electronic circuits and showing bode plots in a simple and quick way, it helps a lot with experimentation [I use electronics workbench for this purpose and PSpice for more detailed simulations]
It also appears to be compensating for raised impedance around 1800Hz due to tweeter resonance since I had to put about 15 ohms as tweeter impedance and 1mH as inductor value to get -6dB and 90º at about 1,8kHz
R1 provides about 1dB of attenuation [providing tweeter impedance is about 7 ohms] and R3 is responsible for turning the filter from 12dB/oct into 6dB/oct
When going active you no longer need to compensate for raised impedance near Fs of the driver, and if you use a 4th order crossover, compensating for tweeter rolloff shouln't be as critical since crossover region isn't as wide
Woofer filter appears to be first order for both drivers, you can think of the coupled inductor as an alternative to two independent inductors of different value, but with lower cost, space saving and making profit of mutual coupling instead of having to avoid it placing one inductor 20-40cm away from the other
This coupled inductor provides an early rolloff for one of the drivers and crosses the other driver with the tweeter at a higher frequency, the purpose of this practice is to compensate for baffle step making the lower driver start to work when the higher driver loses the acoustical load provided by the baffle [it happens at a frequency lower than 1800Hz so just attenuating the tweeter doesn't work, it would leave the range from baffle step frequency to 1800Hz boosted]
To compensate for baffle step you would need to add some equalization at the LF ouput of the active crossover, things as simple as a R in series and then an RC to ground may work [this is called a shelving filter, its response looks like this : ¯¯¯\____
I recommend getting some program capable of simulating simple electronic circuits and showing bode plots in a simple and quick way, it helps a lot with experimentation [I use electronics workbench for this purpose and PSpice for more detailed simulations]
Thanks Eva for the explaination of the crossover network of my Mirage M490is..
I have just ordered a pair of each of the following
AurumCantus G2Si
Jordan JX53
Jordan JX150
FWIW is to bi-amp these guys with the CX3400 active crossover
Can anyone provide me a set of temp crossover network for the above drivers for break in purposes..or any suggestions as to how to break them in before I actually turn them into a pair of speakers...for now
Thanks in advance
Chris
I have just ordered a pair of each of the following
AurumCantus G2Si
Jordan JX53
Jordan JX150
FWIW is to bi-amp these guys with the CX3400 active crossover
Can anyone provide me a set of temp crossover network for the above drivers for break in purposes..or any suggestions as to how to break them in before I actually turn them into a pair of speakers...for now
Thanks in advance
Chris
Hi all.
To the initial thread poster:
Bi-amping is not at all easy to do. It can be hard to adjust and the active crossovers and power amps do have to be good.
However, if you can do all this you will be TOTALLY BLOWN AWAY by the sound quality.
I have been listening using STAX SR-5 headphones since 1977, so I have good ears.
After a number of designs starting in 1979 for the active filters and speaker enclosures, finally my home made system settled in 1986 at a 7 channel (55 W rms/ch) four band quad amped setup, using a (home made) JANIS W2 clone for the subwoofer. The sound is both spectacular and accurate (very accurate). The sub is a real window and door rattler. I manage about 115 dB in the loungeroom at full power with more distortion in my ears than the speakers. I'm not kidding.
I would not even consider a conventional speaker system,
regardless of cost. The very best are very good, but then if you bi-amped them, well, you know what I am getting at.
I am considering designing an integrated six channel power amp and active filter system for DIY bi-ampers. Doing it really well is not as easy as you might think at first... so I'll have a design with PCBs available evntually (+ six months).
My current audio project is a precision intelligent parametric equalizer, but thats going in another thread.
Best of luck with your projects guys!
Don Wallace
To the initial thread poster:
Bi-amping is not at all easy to do. It can be hard to adjust and the active crossovers and power amps do have to be good.
However, if you can do all this you will be TOTALLY BLOWN AWAY by the sound quality.
I have been listening using STAX SR-5 headphones since 1977, so I have good ears.
After a number of designs starting in 1979 for the active filters and speaker enclosures, finally my home made system settled in 1986 at a 7 channel (55 W rms/ch) four band quad amped setup, using a (home made) JANIS W2 clone for the subwoofer. The sound is both spectacular and accurate (very accurate). The sub is a real window and door rattler. I manage about 115 dB in the loungeroom at full power with more distortion in my ears than the speakers. I'm not kidding.
I would not even consider a conventional speaker system,
regardless of cost. The very best are very good, but then if you bi-amped them, well, you know what I am getting at.
I am considering designing an integrated six channel power amp and active filter system for DIY bi-ampers. Doing it really well is not as easy as you might think at first... so I'll have a design with PCBs available evntually (+ six months).
My current audio project is a precision intelligent parametric equalizer, but thats going in another thread.
Best of luck with your projects guys!
Don Wallace
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.