GM,
I haven't kept up with Radian lately. I know that they have changed the compression drivers since I last played with them. The older models needed more than a simple xo to work. They had some definite hash in the upper ranges that required a notch filter and they also needed a sloping curve in the xo due to their dropping response at higher frequencies. I hope they are better than they were in the past, they have had plenty of time to correct some of this.
I haven't kept up with Radian lately. I know that they have changed the compression drivers since I last played with them. The older models needed more than a simple xo to work. They had some definite hash in the upper ranges that required a notch filter and they also needed a sloping curve in the xo due to their dropping response at higher frequencies. I hope they are better than they were in the past, they have had plenty of time to correct some of this.
He's referring to a vintage RCA 15" 'FR' driver that has a separate little tweeter built in.
GM
Attachments
FWIW, the two I auditioned several years ago were digitally EQ'd flat at the LP via a laptop, but they still sounded distinctly PA, which my tinnitus doesn't tolerate.
GM
Gm,
I'm going to take a guess and assume that they were probably using one of their 2" exit compression drivers. Those designs would have way to much on axis response and little dispersion giving that recognizable horn honk. It is so simple to understand I just can't fathom why designers keep doing that. I love the picture of the RCA design, I saved the image, it gave me some real ideas for a future device. Thanks for that one.
Steven
I'm going to take a guess and assume that they were probably using one of their 2" exit compression drivers. Those designs would have way to much on axis response and little dispersion giving that recognizable horn honk. It is so simple to understand I just can't fathom why designers keep doing that. I love the picture of the RCA design, I saved the image, it gave me some real ideas for a future device. Thanks for that one.
Steven
Why aren't coaxial speakers more popular?
The subject is interesting...though, as in all things related to "statistics", the POV can be twisted: "why musical good taste is not more common?";
"why common sense is not more prevalent?"; "why people like more rumbling noises than natural tones?"
Popular means mass predilection, and we are not here to fulfil our mass-derived expectations...lets go build proper coaxial based speakers instead...
The subject is interesting...though, as in all things related to "statistics", the POV can be twisted: "why musical good taste is not more common?";
"why common sense is not more prevalent?"; "why people like more rumbling noises than natural tones?"
Popular means mass predilection, and we are not here to fulfil our mass-derived expectations...lets go build proper coaxial based speakers instead...
The trouble with many, perhaps most, coaxials is the response of the woofer is pretty messy in the crossover region. Digital EQ is not a cure-all, since there are very large response changes with changes in emission angle (spikes in the polar pattern that are frequency dependent).
This comes down to the big horn right smack dab in the middle of the cone. The cone doesn't like this; I'm not even sure the "phase plugs" that are popular these days are an improvement. In other words, hard reflecting surfaces in the middle of the cone does odd things to the time response, which also leads to directional spikes in the polar pattern.
I agree 100% that a driver like the famous RCA LC-1A needs to go back into production. The patents have long expired - heck, RCA itself is just a trade-name owned by Thomson of France. Another fun driver to bring back would be the Altec Biflex, with an elegant decoupling ring, which provides an inner and outer cone.
This comes down to the big horn right smack dab in the middle of the cone. The cone doesn't like this; I'm not even sure the "phase plugs" that are popular these days are an improvement. In other words, hard reflecting surfaces in the middle of the cone does odd things to the time response, which also leads to directional spikes in the polar pattern.
I agree 100% that a driver like the famous RCA LC-1A needs to go back into production. The patents have long expired - heck, RCA itself is just a trade-name owned by Thomson of France. Another fun driver to bring back would be the Altec Biflex, with an elegant decoupling ring, which provides an inner and outer cone.
Last edited:
My first speaker build was a Lafayette 12" coax, I had four of them, two each to a cabinet
Driven by 20 watts per ch from a Lafayette receiver, wow!
Can you really hear offset from separated components?
After building a pair from DynAudio and using hefty components, the sound-stage was melted, it was gone, wide!
But then we get into the components driving the drivers?
Would I use coax? Yes, definitely, I would! but what are you listening to and where?
How good is the source, the microphone used to record, the in-between?
Depends on what you listen to, string quartet? Or rock band?
Funny thing, I was looking at drivers and noticed coax, really expensive!
KEF, they are nice!
Driven by 20 watts per ch from a Lafayette receiver, wow!
Can you really hear offset from separated components?
After building a pair from DynAudio and using hefty components, the sound-stage was melted, it was gone, wide!
But then we get into the components driving the drivers?
Would I use coax? Yes, definitely, I would! but what are you listening to and where?
How good is the source, the microphone used to record, the in-between?
Depends on what you listen to, string quartet? Or rock band?
Funny thing, I was looking at drivers and noticed coax, really expensive!
KEF, they are nice!
How much would it cost to get these manufactured in China?
If they are indeed that good, then you might get 30 people crazy enough to put their money where their mouth is and get the ball rolling and have them for sale in the diyaudio store
KEF makes some pretty good coaxials. Their Blade is certainly an interesting speaker to say the least. The Q900 mid they use could have very interesting results with a more aggressive crossover. The R900 mid looks sweet.
And of course, there's TAD and Pioneer EX. They do have slightly worse on-axis frequency response, but are still some of the finest speakers out there.
I also think what AJ in FLA has been doing with coaxials is worthy of attention.
And of course, there's TAD and Pioneer EX. They do have slightly worse on-axis frequency response, but are still some of the finest speakers out there.
I also think what AJ in FLA has been doing with coaxials is worthy of attention.
Last edited:
Without directing my attention to any one design I will give you my take on the three basic design constructions and my opinion of the pluses and minuses of each.
The first design to consider here are the cone loudspeaker, any size and compression driver combination. In this design the mid/bass cone assembly is basically the same as any other direct radiator and has all the pluses and minuses of that traditional design. Nothing special, a well known and understood design. The obvious difference is that the center dust cap is removed from the design. A compression driver is mounted to the back plate of the magnet structure and the pole piece rather than solid as is normal is a hollow cylinder. This hollow pole piece is traditionally shaped as the first section of a normal compression driver whether that is a straight conic section or an exponential expansion which is much less common. This section is connected to the cone driver by one of two methods. The cone can be used as the waveguide, Tannoy and others, or a small short horn is mounted in front of the throat section in front of the cone. Here are the limitations of each design. Both of these designs will have a rather long throat section that has a limited expansion rate, this causes beaming in the upper frequencies of either of the two implementations. The longer this transition the greater the beaming. One solution has been to foreshorten this area as much as possible and use a very fast expansion rate allowing for greater dispersion. Another attempt to correct for this is the use of larger format drivers, typically 2" exit drivers to increase mid-range output, but this comes at the cost of both less upper frequency response and a greater tendency to beam on axis. The second implementation of these devices is the short horn lens coupled to the throat section in front of the cone, Altec and UREI are two of these designs. These designs have two short comings. The horns are very small, the lower cutoff frequency is limited to the 1/4 wave mouth size of the waveguide and this is typically fairly high in the frequency band. Meaning that the cone driver has to go much higher in frequency before cutoff and this leads to both cone breakup and poor polar response at the crossover frequency. Another problem is the destructive interference to the radiation pattern of the cone driver. Neither of these two designs has a perfect response curve but they are highly efficient and can sound extremely good but you have to be aware of the limitations of each of the two basic designs. Also remember that though the acoustic center is concentric the time alignment of these designs is not optimum without an active time alignment circuit in an active crossover. The distance is to great to correct this with a passive alignment circuit.
The Second typical coaxial design is that of a dome tweeter or some version of that in place of the above compression driver at the apex of where the dust cap would normally be on the cone driver. This implementation has the advantage of both coaxial and time alignment of the two devices.The RCA design earlier shown is a very early design of this design taken to a high level of integration. Many times these designs are simply a dome tweeter with a separate magnetic system mounted on the pole piece of the driver where the dust cap would reside. The time alignment will be slightly askew as the magnetic system will offset the tweeter forward in time alignment. This design can have a fairly good sound depending on the quality of the two devices but will have a slight interference pattern in the upper mid-range response of the cone due to the assembly in front of the cone driver. Typically the efficiency match of the dome tweeter is on the lower side due to the minimization of the motor assembly allowable in front of the bass cone.
The third implementation is a tweeter or mid range and tweeter device mounted on a plate in front of the cone loudspeaker. This design has many drawbacks to its design. The time alignment of these devices are no better than a dome tweeter separately mounted on a baffle, though the on axis implementation is superior. Mounting the device in front of the cone loudspeaker will fundamentally interfere with both the radiation pattern of the cone portion of the driver. There will be a shading of the cone driver, cavity resonances, reflective energy redirected back at the cone and many times resonance output from the actual mounting plate emanating in the audio band. This is the typical design of a car loudspeaker where this design is used due to space constraints and ease of installation. This is the most destructive implementation of the three designs, with poor polar response, lack of time alignment and cavity resonances. Carefully evaluate your needs and desires as in any loudspeaker selection. do not believe the marketing hype that discounts the design flaws of each implementation. Listen carefully to any of these designs before settling on any particular brand or design. Let your ears tell you the story, not a disingenuous specification sheet.
The concept of a coaxial time aligned single source driver is one of the holy grails in loudspeaker design, but there have been no perfect implementations of all of these design parameters. Chose wisely, each design has issues that must be accepted, just as any distributed loudspeaker design has its own design concerns and limitations.
The first design to consider here are the cone loudspeaker, any size and compression driver combination. In this design the mid/bass cone assembly is basically the same as any other direct radiator and has all the pluses and minuses of that traditional design. Nothing special, a well known and understood design. The obvious difference is that the center dust cap is removed from the design. A compression driver is mounted to the back plate of the magnet structure and the pole piece rather than solid as is normal is a hollow cylinder. This hollow pole piece is traditionally shaped as the first section of a normal compression driver whether that is a straight conic section or an exponential expansion which is much less common. This section is connected to the cone driver by one of two methods. The cone can be used as the waveguide, Tannoy and others, or a small short horn is mounted in front of the throat section in front of the cone. Here are the limitations of each design. Both of these designs will have a rather long throat section that has a limited expansion rate, this causes beaming in the upper frequencies of either of the two implementations. The longer this transition the greater the beaming. One solution has been to foreshorten this area as much as possible and use a very fast expansion rate allowing for greater dispersion. Another attempt to correct for this is the use of larger format drivers, typically 2" exit drivers to increase mid-range output, but this comes at the cost of both less upper frequency response and a greater tendency to beam on axis. The second implementation of these devices is the short horn lens coupled to the throat section in front of the cone, Altec and UREI are two of these designs. These designs have two short comings. The horns are very small, the lower cutoff frequency is limited to the 1/4 wave mouth size of the waveguide and this is typically fairly high in the frequency band. Meaning that the cone driver has to go much higher in frequency before cutoff and this leads to both cone breakup and poor polar response at the crossover frequency. Another problem is the destructive interference to the radiation pattern of the cone driver. Neither of these two designs has a perfect response curve but they are highly efficient and can sound extremely good but you have to be aware of the limitations of each of the two basic designs. Also remember that though the acoustic center is concentric the time alignment of these designs is not optimum without an active time alignment circuit in an active crossover. The distance is to great to correct this with a passive alignment circuit.
The Second typical coaxial design is that of a dome tweeter or some version of that in place of the above compression driver at the apex of where the dust cap would normally be on the cone driver. This implementation has the advantage of both coaxial and time alignment of the two devices.The RCA design earlier shown is a very early design of this design taken to a high level of integration. Many times these designs are simply a dome tweeter with a separate magnetic system mounted on the pole piece of the driver where the dust cap would reside. The time alignment will be slightly askew as the magnetic system will offset the tweeter forward in time alignment. This design can have a fairly good sound depending on the quality of the two devices but will have a slight interference pattern in the upper mid-range response of the cone due to the assembly in front of the cone driver. Typically the efficiency match of the dome tweeter is on the lower side due to the minimization of the motor assembly allowable in front of the bass cone.
The third implementation is a tweeter or mid range and tweeter device mounted on a plate in front of the cone loudspeaker. This design has many drawbacks to its design. The time alignment of these devices are no better than a dome tweeter separately mounted on a baffle, though the on axis implementation is superior. Mounting the device in front of the cone loudspeaker will fundamentally interfere with both the radiation pattern of the cone portion of the driver. There will be a shading of the cone driver, cavity resonances, reflective energy redirected back at the cone and many times resonance output from the actual mounting plate emanating in the audio band. This is the typical design of a car loudspeaker where this design is used due to space constraints and ease of installation. This is the most destructive implementation of the three designs, with poor polar response, lack of time alignment and cavity resonances. Carefully evaluate your needs and desires as in any loudspeaker selection. do not believe the marketing hype that discounts the design flaws of each implementation. Listen carefully to any of these designs before settling on any particular brand or design. Let your ears tell you the story, not a disingenuous specification sheet.
The concept of a coaxial time aligned single source driver is one of the holy grails in loudspeaker design, but there have been no perfect implementations of all of these design parameters. Chose wisely, each design has issues that must be accepted, just as any distributed loudspeaker design has its own design concerns and limitations.
Tempests inna Teapot.
Tannoy's had evolved, by late 60's into arguably the
'best' configuration if not implementation of the Coax driver: Huge LF cone with a Compression driver HF feeding an expo horn using the LF cone as part of it
It worked! Yes.. some wee anomalies inherent. There is a crossover spike at ~2hz with the LF ! cone. Due in large part to the early design Crossover implementation.. Correctable actually
But these are Speakers ...of which NONE are of as yet; perfect.
But well setup there are few, If any drivers that can produce a more pleasing performance.
There are no imitators.
Buy a pair and listen.. talk is wayyy to facile
Presumably the cost of Manufacture and likely Patents? Not to mention the serious R& D involved in developing the design.
Speaker 'makers' typically source a few V cheap drivers out of China.
Slap on some cheap Chinese crossover bits.. dick with it for a short while and then sell them en masse for Home Theatre buyers. Who are Happy as Long as the Explosions sound dramatic.
Face it.. Audio is on Life Support. As a guess: average Audio Weenie Age is ~60.
Which in itself is amusing, given that normal physiology dictates a substantial hearing loss in Human males of said age.
Tannoy's had evolved, by late 60's into arguably the
'best' configuration if not implementation of the Coax driver: Huge LF cone with a Compression driver HF feeding an expo horn using the LF cone as part of it
It worked! Yes.. some wee anomalies inherent. There is a crossover spike at ~2hz with the LF ! cone. Due in large part to the early design Crossover implementation.. Correctable actually
But these are Speakers ...of which NONE are of as yet; perfect.
But well setup there are few, If any drivers that can produce a more pleasing performance.
There are no imitators.
Buy a pair and listen.. talk is wayyy to facile
Presumably the cost of Manufacture and likely Patents? Not to mention the serious R& D involved in developing the design.
Speaker 'makers' typically source a few V cheap drivers out of China.
Slap on some cheap Chinese crossover bits.. dick with it for a short while and then sell them en masse for Home Theatre buyers. Who are Happy as Long as the Explosions sound dramatic.
Face it.. Audio is on Life Support. As a guess: average Audio Weenie Age is ~60.
Which in itself is amusing, given that normal physiology dictates a substantial hearing loss in Human males of said age.
Last edited:
Bare,
Hey I almost resemble that last remark, but I have a few years to go and my hearing is still there into the upper limits....
I agree that currently the Tannoy is most likely the best of the compression driver arraignments out there. All the Altecs are very old by now and would most likely need a recone, paper only lasts so long. The small horn always was a limitation in the mid-band crossover region. I assume you meant to say 2Khz for the problem with the Tannoy and not 2hz, that would be nice! The tulip design transition to the cone is much better than an open throat, that never works on the top end, just spits it out on axis and dies anywhere else. I am a fan of the idea, just wish more would follow that implementation to get the dispersion correct. Still need an active crossover to get the time alignment correct and they never made a very nice passive crossover for them.
Steven
Hey I almost resemble that last remark, but I have a few years to go and my hearing is still there into the upper limits....
I agree that currently the Tannoy is most likely the best of the compression driver arraignments out there. All the Altecs are very old by now and would most likely need a recone, paper only lasts so long. The small horn always was a limitation in the mid-band crossover region. I assume you meant to say 2Khz for the problem with the Tannoy and not 2hz, that would be nice! The tulip design transition to the cone is much better than an open throat, that never works on the top end, just spits it out on axis and dies anywhere else. I am a fan of the idea, just wish more would follow that implementation to get the dispersion correct. Still need an active crossover to get the time alignment correct and they never made a very nice passive crossover for them.
Steven
"…just spits it out on axis and dies anywhere else."
Except that even the old pepperpots have a very uniform 90deg dispersion.
As I told you before elsewhere they behave exactly like any other constant directivity horn.
The timing difference with ferrite magnets is about 6 microsecs, too little to bother with ie the cure is worse than the disease.
The AlNiCos are in phase and as such the timing matters very little indeed. Again the cure would be worse…
Except that even the old pepperpots have a very uniform 90deg dispersion.
As I told you before elsewhere they behave exactly like any other constant directivity horn.
The timing difference with ferrite magnets is about 6 microsecs, too little to bother with ie the cure is worse than the disease.
The AlNiCos are in phase and as such the timing matters very little indeed. Again the cure would be worse…
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.