DVV.
I just grabbed a 10" driver I made and measured the cone diameter as about 8" (200mm) and I would say if it was properly designed should be able to go easily as high as 1.7khz without breakup modes. I'd guess as I don't have a 4" cone on hand that would be about a 2 1/2" (65mm) diameter cone and that should go up to about 5.5Khz no problems. I use my 6 1/2" driver with about a 4 3/4" ( 120mm) cone up to 2.6Khz with no problems and it actually go much higher than that. I think your crossover points are very conservative and have no problem with that. Of course you have to look at the actual FR of each device to make these decisions but in general your choices seem reasonable.
As far as crossovers go I am of the mind as Eperado stated earlier that higher order filters are preferred if the out of band response of each device is not very smooth and the two devices do not mate well. On the other hand I think that first order networks just allow to much overlap in most instances. Second order filters just bother me with the phase reversal from the network and you really should invert the input of the upper frequency device. One issue to think about is the real slope of the devices at the crossover points, many talk about using a lower order filter plus the slope of the device but you are probably working in an area where you are far from that slope, the FR would be rather flat at the points you have chosen and therefor you have no acoustical slope adding to the electrical slope and then it is all up to the electrical slope. In that instance I would use a 4th order Linkwitz-Riley crossover if we are talking passive networks. Networks get complex as we start looking at phase angle and time alignment so what I am saying is just generalities. Another point to keep in mind is to use the actual ac impedance of the device at the crossover point, not the rated dc impedance of the device, the overlap of the two devices will shift if you do that and nothing will be correct.
I actually was a fan of the Son-Audax drivers before they got bought by Harman, I don't know what happened to those drivers after that point. I have had some of the so called aerogel drivers, a misnomer use for advertisement only. they aren't really aerogels as far as I am concerned. I also have tested some other Danish drivers with foam cell cones and wasn't real impressed with them, somewhat of a polypropylene cone kind of sound, a little listless in sound production, just seems to kill the dynamics to me.
I just grabbed a 10" driver I made and measured the cone diameter as about 8" (200mm) and I would say if it was properly designed should be able to go easily as high as 1.7khz without breakup modes. I'd guess as I don't have a 4" cone on hand that would be about a 2 1/2" (65mm) diameter cone and that should go up to about 5.5Khz no problems. I use my 6 1/2" driver with about a 4 3/4" ( 120mm) cone up to 2.6Khz with no problems and it actually go much higher than that. I think your crossover points are very conservative and have no problem with that. Of course you have to look at the actual FR of each device to make these decisions but in general your choices seem reasonable.
As far as crossovers go I am of the mind as Eperado stated earlier that higher order filters are preferred if the out of band response of each device is not very smooth and the two devices do not mate well. On the other hand I think that first order networks just allow to much overlap in most instances. Second order filters just bother me with the phase reversal from the network and you really should invert the input of the upper frequency device. One issue to think about is the real slope of the devices at the crossover points, many talk about using a lower order filter plus the slope of the device but you are probably working in an area where you are far from that slope, the FR would be rather flat at the points you have chosen and therefor you have no acoustical slope adding to the electrical slope and then it is all up to the electrical slope. In that instance I would use a 4th order Linkwitz-Riley crossover if we are talking passive networks. Networks get complex as we start looking at phase angle and time alignment so what I am saying is just generalities. Another point to keep in mind is to use the actual ac impedance of the device at the crossover point, not the rated dc impedance of the device, the overlap of the two devices will shift if you do that and nothing will be correct.
I actually was a fan of the Son-Audax drivers before they got bought by Harman, I don't know what happened to those drivers after that point. I have had some of the so called aerogel drivers, a misnomer use for advertisement only. they aren't really aerogels as far as I am concerned. I also have tested some other Danish drivers with foam cell cones and wasn't real impressed with them, somewhat of a polypropylene cone kind of sound, a little listless in sound production, just seems to kill the dynamics to me.
Merci, Chrostophe, I wanted Kindhorneman's opinion because he seems very knowledgable about speakers, certainly ore than myself.
FYI, the bass is cut off at 800 Hz in my speakers. That particular point was chosen because the mid is linear from about 600 Hz or so, making the transition smooth in terms of driver capability.
FYI, the bass is cut off at 800 Hz in my speakers. That particular point was chosen because the mid is linear from about 600 Hz or so, making the transition smooth in terms of driver capability.
No, that's not how it works. The issue is transverse or flex vibrations of materials, which is not related to the speed of sound as a compression wave through the material. Rather dispersion is a matter of flex, mass and loss - related to the shape and the material and how it is retained.
Kindhorman, thanks for the input.
In my case, the bass drivers's net diameter is actually about 220 mm, and the midrange's about 85-90 mm. Of course, this also depends and where you measure it from, because if I measured the bass driver's metal edges, it wouldn't be 256 mm (10"), but more like 280 mm. 220 mm from suspenstion inner edge to the other inner edge. Don't be surprised by that, Celestion also had a model they actually sold as with an 11" bass dirver (outer edge to outer edge).
My XO uses 18 dB/oct cut off points, only the midrange driver starts with a 12 dB/oct slope, but finishes wih 18 dB/oct. This was the last addition to the speakers befoew we called it a day because this definitely sounded better than 3rd order.
At the time (2002), all three were what were thoought of as Son Audax' bes quality drivers. The mid especially was highly thought of and appeared in several very high end speakers at the time. Just by listening to it, and that's the crticial midrange, it's as clean and clear as any I have ever heard.
In my case, the bass drivers's net diameter is actually about 220 mm, and the midrange's about 85-90 mm. Of course, this also depends and where you measure it from, because if I measured the bass driver's metal edges, it wouldn't be 256 mm (10"), but more like 280 mm. 220 mm from suspenstion inner edge to the other inner edge. Don't be surprised by that, Celestion also had a model they actually sold as with an 11" bass dirver (outer edge to outer edge).
My XO uses 18 dB/oct cut off points, only the midrange driver starts with a 12 dB/oct slope, but finishes wih 18 dB/oct. This was the last addition to the speakers befoew we called it a day because this definitely sounded better than 3rd order.
At the time (2002), all three were what were thoought of as Son Audax' bes quality drivers. The mid especially was highly thought of and appeared in several very high end speakers at the time. Just by listening to it, and that's the crticial midrange, it's as clean and clear as any I have ever heard.
Attachments
See my reply to vacuphile above - the issue is not about the velocity of a compression wave in (say) cone material. Rather it's about the velocity of flex waves, a function of cone shape and geometry, and material properties such as elasticity, mass and loss. And indeed about methods to manage flex, such as cone material weave, lamination etc.
The cone does not move as a body, rather it flexes in 3D. In response to an impulse motion of the voice coil, a waveshape propagates within the cone outwards. This wave is typically dispersive, ie velocity depends on frequency, so its shape changes as it radiates, and the outer part of the cone moves later than the inner.
This happens far far slower than velocity of sound in the cone material. Just as the velocity of a transverse wave on a taught rope is unrelated to the speed of sound in cotton !
I may still have a sample of that mid-range in my collection somewhere. I never used any of those bass drivers but have played with some other 10" drivers they had for low bass only. I actually really liked one of the Audax dome tweeters that was cheap but really worked well, the TWO25 driver but I think that was marketed as an Audax and I'm not sure how the Son-Audax and Audax differ if they do at all. That is actually one of the best measuring and sounding soft dome tweeters in my eyes, the elliptical shape of the soft dome seemed to work well and it was just s simple well designed unit. It has some of the cleanest top octave output of many dome tweeters I tested at the time. I still have a few of those and replacement diaphragms, they just sound right for a dome tweeter.
As a side note, Kindhorman, the way you descrived the sound of aerogel or other such like bass drivers is eaxctly why I hate Dynaudio, beacuse its bass cannot fail to sound plasticky to me. The rest id fine, but the bass is killing me.
As for dynamics, let me put it this way - if I ever suspect anything in my system constraining the dynamics, it's out of the house in a jiffy. Remember, in my wild and thoroughly misspent youth, I used to band a drum set, so I know damn well how dynamic it can sound when pushed hard. If evenr I should forget, my old AR94s will remind me, even if they don't have the extension my regular speakers do. As classic bass as they come, dual 8" woofers, lower one has a larger magnet than the top one, and is gradually cut off by an inductor around 150 Hz. That was the world's first 2.5 way speaker, I am told.
As for dynamics, let me put it this way - if I ever suspect anything in my system constraining the dynamics, it's out of the house in a jiffy. Remember, in my wild and thoroughly misspent youth, I used to band a drum set, so I know damn well how dynamic it can sound when pushed hard. If evenr I should forget, my old AR94s will remind me, even if they don't have the extension my regular speakers do. As classic bass as they come, dual 8" woofers, lower one has a larger magnet than the top one, and is gradually cut off by an inductor around 150 Hz. That was the world's first 2.5 way speaker, I am told.
TWO25A16 - that's my tweeter. Not exactly cheap, deaker price was around US$160 at the time, but clean and clear is its middle name, and was really worth the asking price. Super light titanium. You have to work really hard to make it sound bad. Exceptionally well defined. which is why I wanted it.
Son Audax was the full name of the company, but it was often abbreviated to just Audax. Its' the same outfit. Harman shut them down in 2006. This made quite a few people very unhappy.
Son Audax was the full name of the company, but it was often abbreviated to just Audax. Its' the same outfit. Harman shut them down in 2006. This made quite a few people very unhappy.
Are you sure it is group-delay compensated?? Or just way for flat freq response?
http://www.mathworks.com/help/signa...lay-introduced-by-a-filter-with-feedback.html
http://www.mathworks.com/help/dsp/examples/iir-filter-design-given-a-prescribed-group-delay.html
THx-RNMarsh
Last edited:
Luckythedog, you surely must have references to this novel explanation. I don't know what flex waves are, my books on speakers don't mention it, I have never heard it mention in conversations I had on this topic, Google thinks flex waves are something for a guitar, so I am at a loss. Frankly speaking, I think you are making things up.
Last edited:
Lucky,
Thanks for that post. I really appreciate what you just said. I started off working with carbon fiber by itself and found that though I could stiffen the cone without any weight penalties the breakup became much worse and the breakups or bending modes were much sharper, they showed very plainly in a waterfall plot and in the impedance curves as very abrupt. The combination of carbon and Kevlar and my secret sauce worked the best. The only one here who would understand my polymer would be Sy and I am keeping my mouth shut on that one.
As far as the cone shape goes I am in agreement that shape does make a big difference and I used the composite material to my advantage as I could do things that would be much more difficult with a paper type of cone. I looked at the energy transfer of the attachment point of the cone and voice-coil former and optimized that for max energy transfer. I also looked at the final angular attachment to the surround so that it would not end up being a floppy attachment point where the cone actually can become part of the suspension if it becomes to flat at that point. Adhesives at the different attachment points is also critical in how they transfer energy and also in how the adhesives at the cone to surround attachment can become reflective in nature if they are to hard or rigid at that point. The adhesive I used there is a major pain to work with, it has very fast cure time and short mix time, a major bit*h but worth the results. Surrounds are an area that are so often overlooked, so many rubber surrounds reflect so much energy back down the cone, that was another area to look at and I am still exploring new surround topologies for extended excursion.
I still have some tricks up my sleeve in cone development but I have gotten to the point that what I have works well. I have enough development time in composites to see things that others just don't know or understand. Always learning and always trying to spread some of that knowledge to those coming after me in that area. What you see today on the new BMW I series of cars comes directly from that early work I did with Bayer over 25 years ago now. Couldn't give that information away at the time.
ps. Lucky,
I know exactly what you are saying with the transverse waves through the composite material, that is a very good description and explanation of what is happening here.
Thanks for that post. I really appreciate what you just said. I started off working with carbon fiber by itself and found that though I could stiffen the cone without any weight penalties the breakup became much worse and the breakups or bending modes were much sharper, they showed very plainly in a waterfall plot and in the impedance curves as very abrupt. The combination of carbon and Kevlar and my secret sauce worked the best. The only one here who would understand my polymer would be Sy and I am keeping my mouth shut on that one.
As far as the cone shape goes I am in agreement that shape does make a big difference and I used the composite material to my advantage as I could do things that would be much more difficult with a paper type of cone. I looked at the energy transfer of the attachment point of the cone and voice-coil former and optimized that for max energy transfer. I also looked at the final angular attachment to the surround so that it would not end up being a floppy attachment point where the cone actually can become part of the suspension if it becomes to flat at that point. Adhesives at the different attachment points is also critical in how they transfer energy and also in how the adhesives at the cone to surround attachment can become reflective in nature if they are to hard or rigid at that point. The adhesive I used there is a major pain to work with, it has very fast cure time and short mix time, a major bit*h but worth the results. Surrounds are an area that are so often overlooked, so many rubber surrounds reflect so much energy back down the cone, that was another area to look at and I am still exploring new surround topologies for extended excursion.
I still have some tricks up my sleeve in cone development but I have gotten to the point that what I have works well. I have enough development time in composites to see things that others just don't know or understand. Always learning and always trying to spread some of that knowledge to those coming after me in that area. What you see today on the new BMW I series of cars comes directly from that early work I did with Bayer over 25 years ago now. Couldn't give that information away at the time.
ps. Lucky,
I know exactly what you are saying with the transverse waves through the composite material, that is a very good description and explanation of what is happening here.
Last edited:
Are we equalizing for a flat/conatant Group-Delay yet?
Delay equalization - Wikipedia, the free encyclopedia.
THx-RNMarsh
Delay equalization - Wikipedia, the free encyclopedia.
THx-RNMarsh
dvv,
Thanks for the answer on the Son-Audax confusion, I always wondered if it was two separate companies. I still have somewhere the Audax catalog that had so many different devices in it, it was literally a book with so many different speakers. Though they used what appeared to be some crude machining on many of their devices and basically just ceramic magnets they really did seem to understand what they were doing.
I guess your model is the hard dome version and I chose the soft dome, your also has that very short waveguide loading which I have found tilts the FR more towards the bottom end.
The drivers that had that I didn't like were from Seas, very similar to the Dynaudio drivers as far as I am concerned.
Thanks for the answer on the Son-Audax confusion, I always wondered if it was two separate companies. I still have somewhere the Audax catalog that had so many different devices in it, it was literally a book with so many different speakers. Though they used what appeared to be some crude machining on many of their devices and basically just ceramic magnets they really did seem to understand what they were doing.
I guess your model is the hard dome version and I chose the soft dome, your also has that very short waveguide loading which I have found tilts the FR more towards the bottom end.
The drivers that had that I didn't like were from Seas, very similar to the Dynaudio drivers as far as I am concerned.
That is what happens in breakup modes, when the cone is no longer behaving as a piston. You may remember when shallow cones were in fashion. The idea that a flat wavefront was desired. Those of course broke up at low excursion. There really is an optimal cone depth for any given size and material to minimize breakup and still have as flat a cone as possible. But as you understand it is a generally linear combining even from a deeper cone, just not as efficient.
Some folks also use an oversize dust cap to get a flatter radiating area.
The classic British work is "A Textbook of Sound" by A.B. Woods. Has a lot of material that is not in H. Olson's books. Really nice bit of work.
Cross posted!
Unless we are talking about a purely conic shaped cone I would say that even in the pistonic range that the cone does indeed flex, it is not a purely rigid cone as you are imagining it. This is the mistake I see made all the time, assuming that the cone is perfectly rigid and that it does not have flexural properties. Just look at the typical attachment point of a paper cone and the former attachment. Any first years mechanical engineer will see how the lines of force have to cause this condition, the angular attachment becomes a lever arm and will bend the cone, no way do you have a perfectly rigid structure there.
No, it happens all the time. The idea that the cone is rigid and behaves as a single body isn't so, in dynamics the cone is relatively flexible and supports vibrations and waves as such. Breakup is just a special case of flex vibration, sometimes resonant, but flex and propagation of motion from inner to outer happens all the time. Few structures are actually rigid in dynamics.
If you need a reference, try Iain G Mann Vibrations and Waves in Physics ISBN 0-521-27846-5 Ch 12 pp 210-232. But you'll have to make the association between how structures like speaker cones, or vinyl cartridge cantilevers behave for yourself.
You may have misapplied this theory, evidence proves you wrong.
In measurements, the acoustic center of a driver hovers around a point somewhere at middle cone depth. This implies that the speed with which the cone rim reacts to voice coil movements is faster than the speed of sound in air. If your idea were correct, the acoustic center would fall behind the cone. I have never come across a driver where this was the case.
- Status
- Not open for further replies.