I think that there probably isn't room for the ribbon too deep, as the ribbon tweeters are pretty deep themselves. I was however suggesting this as a DIY experiment. It would require one to remove the dust cap in the center of the driver. Anyone who has used one of the kits to replace a driver's rotted suspension can tell you that this is pretty simple. Then glue a piece of pipe, PVC aluminum, cardboard, or whatever comes in an exterior 4" diameter onto the solid steel polepiece which is just inside where the dustcap was. Add some damping inside and then glue the tweeter into the end of the tube. A clever way should be found to try different lengths of tube, but I suspect that getting the tweeter as close in as possible is probably best. The tube could be then wrapped in felt. A 3" tube and tweeter in a 4" voicecoil would probably give a lot more room for experimentation, but even if you remove the tweeter flange I can't remember if it will be small enough- will check This scheme seems a bit more doable now that we are aware of 12" drivers that can go up to 4k or more...
I checked: Darn! 92dB / 2.83V / 1M not efficient enough! Well.... the Phy drivers use piezo tweeters....
OK the Aurum Cantus 2si ribbon has a diameter of 4-5/16" (110mm) but the flange could be removed and easily machined to less than 4" without damage as you can see from the link below.It's 96 db
http://www.e-speakers.com/catalog/aurum_cantus_g2si_4284956.htm
Hmm I wonder if the pressure would damage it? I have a 2si 4.5" laterally away from a 15" woofer and no problems so far...
To make things more simple, why can't a good efficiency dome go in a 90deg constant directivity waveguide to match the efficiency of the woofer. My limited research and experience with waveguides tells me that a dome with a CD waveguide in front would be easier to get right than a CD WG for any other driver. You want the wave front edges perpendicular to the waveguide sides. Typically a lot of trouble or size goes into getting the necessary spherical wavefront shape entering the waveguide. With a dome the wave starts out the proper shape, so no throat is needed.
The only compression driver I ever opened up had a dome membrane in it producing the sound, and what a cluster f#$& the sound had to go through to get to your ears. Radiating from a dome straight into say a 90deg WG seems a whole lot simpler to me, and if you can get the woofer up to 4-5khz, the WG doesn't need to be very big. Radiating 94db/w/m into half space has to get close to your efficiency goal radiating into a 90deg CD WG. With the CD WG you should even run into far less problems related to the woofer operation interfering with the tweeter operation, IM etc.
The whole thing seems so simple to me that there must be some basic flaw, unless the manufacturers are just out to sell you the high priced compression drivers. Why else would they insist on strapping a 106db/w/m compression driver to a 99db/w/m woofer. Let's make our own 99db/w/m tweeter.
Am I that far off the mark?
The only compression driver I ever opened up had a dome membrane in it producing the sound, and what a cluster f#$& the sound had to go through to get to your ears. Radiating from a dome straight into say a 90deg WG seems a whole lot simpler to me, and if you can get the woofer up to 4-5khz, the WG doesn't need to be very big. Radiating 94db/w/m into half space has to get close to your efficiency goal radiating into a 90deg CD WG. With the CD WG you should even run into far less problems related to the woofer operation interfering with the tweeter operation, IM etc.
The whole thing seems so simple to me that there must be some basic flaw, unless the manufacturers are just out to sell you the high priced compression drivers. Why else would they insist on strapping a 106db/w/m compression driver to a 99db/w/m woofer. Let's make our own 99db/w/m tweeter.
Am I that far off the mark?
My small contribution...
I would like to contribute but I have very little speaker knowledge but a reasonable ear for music (musician). Most of this stuff is over my head.
I've noticed the TT Alnico is edge wound, no information on the ceramic TT VC construction (round wound?).
All the coaxes mentioned here (unless I've missed some) are not edge wound. Some of the pro sound non coaxes are edge wound from the same vendors with coaxes.
Not sure if the VC construction has any relevance, but I must say the 12" TT alnico have amazing mids.
Regards
Frank.
I would like to contribute but I have very little speaker knowledge but a reasonable ear for music (musician). Most of this stuff is over my head.
I've noticed the TT Alnico is edge wound, no information on the ceramic TT VC construction (round wound?).
All the coaxes mentioned here (unless I've missed some) are not edge wound. Some of the pro sound non coaxes are edge wound from the same vendors with coaxes.
Not sure if the VC construction has any relevance, but I must say the 12" TT alnico have amazing mids.
Regards
Frank.
Dave, 93mm doesn't surprise me. It could be done. You are Mr. Phaseplug 
so I'd think you'd be tempted. A nice turned wood nose tht tapers from 100mm to 93mm,that fits into a 4"tube which attaches to the polepiece. The tube could just have a piece of steel inserted on the inside end which I would presume would snap right onto the magnatized polepiece. Carefully!!! like the guys that use a socket wrench socket as a phaseplug.
At the moment I cross to the 2si at about 4500 hz and I like it!
At 96db it probably is about the right efficiency to work with 97 Db drivers. People say that padding down a ribbon hurts the sound. I don't have to pad mine down because it matches the mid efficiency so well. That is really elegant IMHO.
I think most of the rear of the ribbon is magnet. It's just that additional bump which I think is the transformer, but I'm just guessing.
I don't know if any hemp 12" drivers have 4" coils though.
Quite likely the Super Boy 15" with cast basket does, but ceramic magnet and how high does it go?
Might have to use the eighteen sound..or some other for an experiment. I don't have measuring equipment.
http://www.6moons.com/audioreviews/bastanis/prometheus_2.html
John,
Here are the new Bastanis tweeters. As you mentioned, they are the insides of a compression driver made to work with a custom horn(below link) I do believe that he leaves the backplate off to get dipole action! The 6 moons site shows this. They look like they could be used as dome tweeters, but without compression, their efficiency might be as low as other dome tweeters.I don't know Also compression driver magnets are pretty huge. They wouldn't fit in a phase plug... All commercial dome tweeters that I have found top out at 93 dB which really isn't enough for what Lynn wants. I think he'll accept 97 dB but I think he dreams of 99 dB
http://www.bastanis.com/catalog/index.php?language=en
http://www.bastanis.com/us/atlas.htm
so I'd think you'd be tempted. A nice turned wood nose tht tapers from 100mm to 93mm,that fits into a 4"tube which attaches to the polepiece. The tube could just have a piece of steel inserted on the inside end which I would presume would snap right onto the magnatized polepiece. Carefully!!! like the guys that use a socket wrench socket as a phaseplug. At the moment I cross to the 2si at about 4500 hz and I like it!
At 96db it probably is about the right efficiency to work with 97 Db drivers. People say that padding down a ribbon hurts the sound. I don't have to pad mine down because it matches the mid efficiency so well. That is really elegant IMHO.
I think most of the rear of the ribbon is magnet. It's just that additional bump which I think is the transformer, but I'm just guessing.
I don't know if any hemp 12" drivers have 4" coils though.
Quite likely the Super Boy 15" with cast basket does, but ceramic magnet and how high does it go?
Might have to use the eighteen sound..or some other for an experiment. I don't have measuring equipment.
http://www.6moons.com/audioreviews/bastanis/prometheus_2.html
John,
Here are the new Bastanis tweeters. As you mentioned, they are the insides of a compression driver made to work with a custom horn(below link) I do believe that he leaves the backplate off to get dipole action! The 6 moons site shows this. They look like they could be used as dome tweeters, but without compression, their efficiency might be as low as other dome tweeters.I don't know Also compression driver magnets are pretty huge. They wouldn't fit in a phase plug... All commercial dome tweeters that I have found top out at 93 dB which really isn't enough for what Lynn wants. I think he'll accept 97 dB but I think he dreams of 99 dB
http://www.bastanis.com/catalog/index.php?language=en
http://www.bastanis.com/us/atlas.htm
A Tale of Two Versions
Keep the discussion going, guys, this is getting interesting. Curious to see what the master of phase-plugs has to say about customized cylindrical ribbons - a coax with a difference, dare I say the first of its kind.
Here are my current favorite candidates from all the discussions so far, with probable crossovers mapped out:
1) The 18Sound 12NDA520 with BMS 4540ND compression driver, possibly with the 18Sound XT1086 80 x 60 elliptical horn.
Both drivers are quite flat as pro gear goes, with not much equalization required. A 1.5 kHz 12 dB/oct lowpass applied to the 12NDA520 will produce an acoustic crossover in the vicinity of 2 ~ 2.4 kHz. The 4540 needs a simple shelving filter, a 2 ~ 2.2 kHz highpass of at least 12 dB/octave, and an autoformer attenuator with 1 dB steps. I suspect this is the simplest possible crossover that will yield flat response and acceptably low distortion.
It does not fulfill prosound criteria of constant directivity, but I'm more interested in low coloration and low IM distortion from the HF unit. Every half-octave I can raise the HF crossover results in substantial reductions of HF IM distortion, which rises very fast with decreasing frequency with nearly every type of tweeter (with the possible exception of cone tweeters).
For me, the choice between directivity control vs IM distortion from the tweeter is easy and obvious - I go for low IM distortion (especially in the 1 to 5 kHz region) every time. This goes back to my days of struggling with the difficult and cantankerous KEF T27 tweeter, which needed little provocation to sound dreadful.
2) The BMS 12C362 coaxial, with the BMS 12C262 and 18Sound 12CX800 as alternates.
The crossovers required here are more complex. The bass driver requires a 2.2 kHz notch filter (for all three drivers, surprisingly), and a possible lowpass filter somewhere in the same region. The tweeter requires a somewhat stronger shelf filter than the one mentioned above, and the highpass filter probably needs to come in a bit early at 2.2 ~ 2.5 kHz, to offset the steeply rising response.
The complexity of the curves indicates the crossover will need substantially more development time - both to get the desired response and tuning for subjectively flat response with pink-noise stimulus. The more complex acoustical structure of the coax driver has implications for subjective coloration, and the required crossover tuning to "dial out" the coloration over a reasonably wide listening angle.
What I am most wary of are directional colorations, which cannot be removed without spoiling the response in the very important +/- 20 degree frontal arc. When I encounter these, the driver has to rejected and the project re-started with another driver. The reason I bring up the difficult issue of directional colorations - the bane of all speaker designers - is that drivers with acoustically complex radiating surfaces are much more prone to these problems. If it looks complex acoustically, it is, and the required EQ issues become more complex as well.
The two examples shown above represent examples of simple and complex crossovers. Gently tilting the response by bringing in a crossover "early", or shelf filtering, is simple, and generally stays within the character of the driver - it doesn't violate its nature, so to speak. This falls in the general domain of system integration, or harmonizing the drivers.
Once you enter the wilder territory of notch filters combined with rolloffs at nearby frequencies, though, results can be more unpredictable. The driver can be EQ'ed perfectly flat, but may sound lifeless and alienated from the basic "personality" that gives it the charm in the first place. This has happened to me before with other speakers I've designed, so it's something I'm cautious about. I don't like having to choose between sparkly sound with heavy colorations and boring, dull sound with minor colorations. Choices like that indicate something is wrong with the basic concept of the speaker system.
Keep the discussion going, guys, this is getting interesting. Curious to see what the master of phase-plugs has to say about customized cylindrical ribbons - a coax with a difference, dare I say the first of its kind.
Here are my current favorite candidates from all the discussions so far, with probable crossovers mapped out:
1) The 18Sound 12NDA520 with BMS 4540ND compression driver, possibly with the 18Sound XT1086 80 x 60 elliptical horn.
Both drivers are quite flat as pro gear goes, with not much equalization required. A 1.5 kHz 12 dB/oct lowpass applied to the 12NDA520 will produce an acoustic crossover in the vicinity of 2 ~ 2.4 kHz. The 4540 needs a simple shelving filter, a 2 ~ 2.2 kHz highpass of at least 12 dB/octave, and an autoformer attenuator with 1 dB steps. I suspect this is the simplest possible crossover that will yield flat response and acceptably low distortion.
It does not fulfill prosound criteria of constant directivity, but I'm more interested in low coloration and low IM distortion from the HF unit. Every half-octave I can raise the HF crossover results in substantial reductions of HF IM distortion, which rises very fast with decreasing frequency with nearly every type of tweeter (with the possible exception of cone tweeters).
For me, the choice between directivity control vs IM distortion from the tweeter is easy and obvious - I go for low IM distortion (especially in the 1 to 5 kHz region) every time. This goes back to my days of struggling with the difficult and cantankerous KEF T27 tweeter, which needed little provocation to sound dreadful.
2) The BMS 12C362 coaxial, with the BMS 12C262 and 18Sound 12CX800 as alternates.
The crossovers required here are more complex. The bass driver requires a 2.2 kHz notch filter (for all three drivers, surprisingly), and a possible lowpass filter somewhere in the same region. The tweeter requires a somewhat stronger shelf filter than the one mentioned above, and the highpass filter probably needs to come in a bit early at 2.2 ~ 2.5 kHz, to offset the steeply rising response.
The complexity of the curves indicates the crossover will need substantially more development time - both to get the desired response and tuning for subjectively flat response with pink-noise stimulus. The more complex acoustical structure of the coax driver has implications for subjective coloration, and the required crossover tuning to "dial out" the coloration over a reasonably wide listening angle.
What I am most wary of are directional colorations, which cannot be removed without spoiling the response in the very important +/- 20 degree frontal arc. When I encounter these, the driver has to rejected and the project re-started with another driver. The reason I bring up the difficult issue of directional colorations - the bane of all speaker designers - is that drivers with acoustically complex radiating surfaces are much more prone to these problems. If it looks complex acoustically, it is, and the required EQ issues become more complex as well.
The two examples shown above represent examples of simple and complex crossovers. Gently tilting the response by bringing in a crossover "early", or shelf filtering, is simple, and generally stays within the character of the driver - it doesn't violate its nature, so to speak. This falls in the general domain of system integration, or harmonizing the drivers.
Once you enter the wilder territory of notch filters combined with rolloffs at nearby frequencies, though, results can be more unpredictable. The driver can be EQ'ed perfectly flat, but may sound lifeless and alienated from the basic "personality" that gives it the charm in the first place. This has happened to me before with other speakers I've designed, so it's something I'm cautious about. I don't like having to choose between sparkly sound with heavy colorations and boring, dull sound with minor colorations. Choices like that indicate something is wrong with the basic concept of the speaker system.
Eton likes the cylindrical phase plug idea. They seem to use it as a heatsink:
http://www.madisound.com/cgi-bin/index.cgi?cart_id=8368061.11057&pid=282
Similar model w/o it to compare:
http://www.madisound.com/cgi-bin/index.cgi?cart_id=8368061.11057&pid=281
http://www.madisound.com/cgi-bin/index.cgi?cart_id=8368061.11057&pid=282
Similar model w/o it to compare:
http://www.madisound.com/cgi-bin/index.cgi?cart_id=8368061.11057&pid=281
Variac said:John,
Here are the new Bastanis tweeters. As you mentioned, they are the insides of a compression driver made to work with a custom horn(below link) I do believe that he leaves the backplate off to get dipole action! The 6 moons site shows this. They look like they could be used as dome tweeters, but without compression, their efficiency might be as low as other dome tweeters.I don't know Also compression driver magnets are pretty huge. They wouldn't fit in a phase plug... All commercial dome tweeters that I have found top out at 93 dB which really isn't enough for what Lynn wants. I think he'll accept 97 dB but I think he dreams of 99 dB
Variac,
At the risk of being repetitive, it seems pretty logical that if you take a dome tweeker that is 93db into half space and put it in a 90deg constant directivity waveguide, you are going to gain efficiency because the same energy is focused into 1/8th space even with no horn loading. Doesn't that also make a number of other things easier especially since polar response is a better match at the XO point? I'm sure there are plenty of tweeters whose magnet structure will fit in a 3" diameter space and end up much easier to get right than using a ribbon, which has vertical dispersion problems anyway.
Late breaking news:
Interestingly, Lynn has found us an interesting candidate for a horn driver-in-a-phase-plug right here:
http://www.bmspro.info/photos/bmspro_info/bms-4540nd.pdf
It is his choice for a horn drive that seems to have good response and reputation. What is interesting is that it's only 3" in diameter and very short also. Would fit well in a phase plug. It's so short that it could be placed to be in phase with the woofer. Just another possibility. Yesterday I saw a photo on this forum somewhere of a Japanese speaker which is exactly what I have been discussing. It is a cone driver on a stand with a small horn driver embedded in a cylindrical phaseplug. Now I can't find it . Anyone see it?
John,
Of course the first thing that I thought of was a dome, back a year ago when this occurred to me. Round polepiece/phase plug ergo round dome tweeter. 93 db domes are actually rather rare and the few there are are barely 93. If you look at them, they are already partially horn loaded so I'm not sure they will get up to the 97 or 98 db range that I think Lynn is looking for, even with wave guide. I spent a lot of time looking for one that works. I found an older fabric Audax that was quite efficient but not over 93. The Morel has the partial horn already. You may be right though. Please post likely candidates here. I suspect there is a lot of good info on this topic here:
http://www.zaphaudio.com/hornconversion.html
Interestingly, Lynn has found us an interesting candidate for a horn driver-in-a-phase-plug right here:
http://www.bmspro.info/photos/bmspro_info/bms-4540nd.pdf
It is his choice for a horn drive that seems to have good response and reputation. What is interesting is that it's only 3" in diameter and very short also. Would fit well in a phase plug. It's so short that it could be placed to be in phase with the woofer. Just another possibility. Yesterday I saw a photo on this forum somewhere of a Japanese speaker which is exactly what I have been discussing. It is a cone driver on a stand with a small horn driver embedded in a cylindrical phaseplug. Now I can't find it . Anyone see it?
John,
Of course the first thing that I thought of was a dome, back a year ago when this occurred to me. Round polepiece/phase plug ergo round dome tweeter. 93 db domes are actually rather rare and the few there are are barely 93. If you look at them, they are already partially horn loaded so I'm not sure they will get up to the 97 or 98 db range that I think Lynn is looking for, even with wave guide. I spent a lot of time looking for one that works. I found an older fabric Audax that was quite efficient but not over 93. The Morel has the partial horn already. You may be right though. Please post likely candidates here. I suspect there is a lot of good info on this topic here:
http://www.zaphaudio.com/hornconversion.html
Re: A Tale of Two Versions
I don't see why not... someone is going to need to check the efficiency... eSpeakers & CSS list 96 dB, the ones i have are rated 94.5.
You might well want to remove the transformer to an external location.
Ruler flat impedance & pretty flat FR, lovely waterfall (i wish more measuring kit allowed setting the time axis in periods).
The but is that to get near the limit of their range you need at least a 4th order XO.
dave
Lynn Olson said:
I don't see why not... someone is going to need to check the efficiency... eSpeakers & CSS list 96 dB, the ones i have are rated 94.5.
You might well want to remove the transformer to an external location.
Ruler flat impedance & pretty flat FR, lovely waterfall (i wish more measuring kit allowed setting the time axis in periods).
The but is that to get near the limit of their range you need at least a 4th order XO.
dave
Variac said:
That little BMS driver is cute! A little too big to fit inside the voice coil of the 18Sound 12NDA520, but almost. (72mm vs 64mm)
The idea, of course, is not to attenuate the compressin driver but to run it from a lower gain, lower power amp. But if you do run it with a passive filter the attenuation helps smooth out the impendance peaks a great deal. That's a good thing.
And being at 106dB/w/m means the driver never has to work very hard, at least compaired to a dome or cone.
I don't wanna rain on anyone's parade - but as far as I know, and I'm not the world's expert on horns (as just about everyone on Audio Asylum informed me) - Little Horns = High Cutoff Frequency, pretty much regardless of profile.
So a little bitty horn that's say, maybe 3 inches long and 3 inches wide, is probably a horn with a 3 kHz cutoff. That doesn't help with a driver that goes away at 3 kHz, although maybe it could meet a Tone Tubby at the edge of its plateau before the TT drops like a stone.
If you're bold enough and use ALK's crossovers, yes, you could use elliptical filters with near-brickwall slopes. That's living too close to the edge for me - I prefer gentle 12 dB/oct filters when I can get away with it, maybe 18 dB/oct when pushed.
So if we're using a big 12 or 15-inch driver in a more "classical" manner, and want to use the horn at least an octave away from the point where it falls off the cliff, the horn gets bigger - maybe 5 ~ 8 inches deep, maybe larger, depending on how much distance we want between the horn cutoff and the crossover, and where the big-boy driver is going to cross over.
Maybe, just maybe, the 18Sound XD125 could have a streamlined housing designed for it - it's pretty small and has a minimum recommended crossover of 2.5 kHz. It all comes down to how close the HF crossover is to horn cutoff, I guess.
So a little bitty horn that's say, maybe 3 inches long and 3 inches wide, is probably a horn with a 3 kHz cutoff. That doesn't help with a driver that goes away at 3 kHz, although maybe it could meet a Tone Tubby at the edge of its plateau before the TT drops like a stone.
If you're bold enough and use ALK's crossovers, yes, you could use elliptical filters with near-brickwall slopes. That's living too close to the edge for me - I prefer gentle 12 dB/oct filters when I can get away with it, maybe 18 dB/oct when pushed.
So if we're using a big 12 or 15-inch driver in a more "classical" manner, and want to use the horn at least an octave away from the point where it falls off the cliff, the horn gets bigger - maybe 5 ~ 8 inches deep, maybe larger, depending on how much distance we want between the horn cutoff and the crossover, and where the big-boy driver is going to cross over.
Maybe, just maybe, the 18Sound XD125 could have a streamlined housing designed for it - it's pretty small and has a minimum recommended crossover of 2.5 kHz. It all comes down to how close the HF crossover is to horn cutoff, I guess.
Yeah, I thought that you were going to to try to get well up to about 4k with the 12" But if as Dave says, a three inch diameter horn will handle 2k then a 3k crossover should be fine with the 3" horn.
has there been a curve shown for the TT?
Curses! I didn't notice the small V.C. That probably helps it with it's high response.
OK, anyone know the TT voice coil diameter? They don't seem to give out much info... I think we might be in a loop here. The coax horn in a phase plug needs to be crossed high but few 12" drivers that cover high frequencies have large voice coils.