Sitting here working and thinking about all the negative phenomenon with the coaxial concept brought up in this discussion and others. Trying to come up with lil solutions that can lean on each other to form a complete answer. I feel I have it, a one answer to all the questions raised so far and if I have answered this right, it seems within my capability to make it at home and prepare it, so others can make one too. This belongs to the community, I am only connecting the neck bone to the hip bone. Prolly another concept that I will end up flying solo in monologue
I posted this drawing in a previous post:
This is a real driver I am working on. It has hidden suspension for both the dome and the cone. The waveguide is extended by around 40mm by an additional trim ring. The dome is FEA optimized with a reinforcement and seems to perform really well. I am planning to BEM-simulate it as well. The goal is to get the 32mm dome to extend pretty low, and the midrange paper cone to extend very high while at the same time fit the dome dispersion really well. The domes dispersion is further improved by its extension outside the coil, making it load the waveguide better. The two share a very powerful radial motor so efficiency is likely to be very high.
I feel that I have adressed the important issues with this design, but maybe I have overlooked something?
The midrange design is based on my article in this issue of AudioXpress:
https://audioxpress.com/news/active...roduct-development-in-audioxpress-august-2023
This is a real driver I am working on. It has hidden suspension for both the dome and the cone. The waveguide is extended by around 40mm by an additional trim ring. The dome is FEA optimized with a reinforcement and seems to perform really well. I am planning to BEM-simulate it as well. The goal is to get the 32mm dome to extend pretty low, and the midrange paper cone to extend very high while at the same time fit the dome dispersion really well. The domes dispersion is further improved by its extension outside the coil, making it load the waveguide better. The two share a very powerful radial motor so efficiency is likely to be very high.
I feel that I have adressed the important issues with this design, but maybe I have overlooked something?
The midrange design is based on my article in this issue of AudioXpress:
https://audioxpress.com/news/active...roduct-development-in-audioxpress-august-2023
I meant going beyond a discussion to actually starting a community project. I have a concept. I have had many but there has never been any takers for collaboration
Your design seems solid mechanically, and acoustically, as much as I can understand these that. Keep in mind, while I can make many types of structure, I don't know things like maths and equations nor do I know about acoustic modelling
How do you feel about feeding the lurkers? Guys who are chiming in are in the right spirit, there are a lot more just leaching
Your design seems solid mechanically, and acoustically, as much as I can understand these that. Keep in mind, while I can make many types of structure, I don't know things like maths and equations nor do I know about acoustic modelling
How do you feel about feeding the lurkers? Guys who are chiming in are in the right spirit, there are a lot more just leaching
It's jumping out to me, what you may have missed
Also with solving the reflex system negatives, I have a concept and waiting on parts to build a proof of concept through trail and error as I couldn't raise any interest from the maths types folks
Same with the thing about tools for speaker building. I have a viable design for a driver hole tool and a panel cutter that address all the negatives of using a router and table saw. Again, I'll get there through intuition and testing but couldn't raise any interest
I hope that the current discussion has checked off your boxes for whatever you were wanting to know. (There are at least one or two here that seem to have seized the discussion for their own purposes. Must be a Myers-Briggs thing...
)The reason why it's important to distinguish minimum-phase defects in amplitude response is because it's a fatal flaw in driver design if it's not. Also, polar coverage consistency hasn't been discussed in detail, but is at least as important a subject as on-axis amplitude response--since EQ can correct most minimum-phase amplitude response issues, but nothing but a horn can correct polar response problems.
All of these internal cancellation and polar coverage consistency problems are typical for direct-radiator type coaxial drivers (cone diaphragm with horn-loaded HF driver inside the magnet structure of the woofer cone). If you also look at dual-diaphragm compression drivers, I think that the issues of self-cancellation are very much lessened relative to coaxial direct radiator cone drivers, and you get consistent polar coverage, to boot, with a single straight-sided horn, and inaudible modulation distortion, with the possibility of flat phase response, too. The prices are about the same for the coaxial direct radiating drivers.
Chris
Last edited:
Allow me to be vague because I can't explain the maths
In boat design, you can make a ship longer to allow more payload, this will have a minus and a plus. A longer vessel will have greater hull speed but will be heavier
Rules like this appear everywhere, and I am sure there is in audio as I can observe the pattern when modelling in Winisd or HR
To get a significant result, one has to change something, at the cost of something else. To get a usable, significant result, one will have to allow a change in another perspective to adjust any. The key is to make that allowance also useful so that it is a positive allowance rather than a negative
If I knew the maths, I could present it without my concept seeming stupid yet again
In boat design, you can make a ship longer to allow more payload, this will have a minus and a plus. A longer vessel will have greater hull speed but will be heavier
Rules like this appear everywhere, and I am sure there is in audio as I can observe the pattern when modelling in Winisd or HR
To get a significant result, one has to change something, at the cost of something else. To get a usable, significant result, one will have to allow a change in another perspective to adjust any. The key is to make that allowance also useful so that it is a positive allowance rather than a negative
If I knew the maths, I could present it without my concept seeming stupid yet again
It's these kinds of attitude that makes me say, hey man, enjoy the status quo with what's currently available to you. We can make a community driver together, or I can just be happy with the Pioneer coax and save myself the bother, there are too many ideas and only one deathI hope that the current discussion has checked off your boxes for whatever you were wanting to know. (There are at least one or two here that seem to have seized the discussion for their own purposes. Must be a Myers-Briggs thing...
Funny thing is that I did! Read from start, I am out here man, so not worth thisStart your own thread.
There are things that I think still need to be discussed relative to the the OP's initial problem statement, it seems to me (at least). I'd personally like to see that conversation play out a bit more before switching gears to "DIY coaxial drivers", since it's been made clear that a thread already exists on that subject. It's nice to understand what's not good and what is good with current coaxial (direct radiator) drivers. All of the questions below I think remain based on the original question posed, and is why I took time to post here.
It's always interesting to explore the likely business reasons why so many coaxial direct radiating drivers are made and sold today. Is this because the builders or users simply don't know how to deal with horn-loading the LF cone properly? Is it because users ignore the polar coverage consistency needs (i.e., the need for flat--not inclined--directivity index plots)? Is it because users don't understand the effects of bass/midrange modulation distortion on the sound quality of these drivers? Tannoy makes good coaxial drivers that have poor bass/midrange directivity control and modulation distortion issues, but relatively excellent HF directivity control and modulation distortion. Is this why they haven't been more successful in the current marketplace? Is it price, only?
The comment that nc535 said about horn-loading these type of drivers is something that I think is also within bounds of the OP's intent (I would hope): why would someone take a direct radiating-type driver and try to make it into a compression driver for an MEH? It seems counterproductive to me, considering the availability of dual-diaphragm compression drivers at about the same price point. It certainly isn't because of their upper end SPL response for PA duty. Is it only the simplicity of placing a single driver in a box, then making the multiple electrical connections and more complex crossover dividing networks?
Chris
It's always interesting to explore the likely business reasons why so many coaxial direct radiating drivers are made and sold today. Is this because the builders or users simply don't know how to deal with horn-loading the LF cone properly? Is it because users ignore the polar coverage consistency needs (i.e., the need for flat--not inclined--directivity index plots)? Is it because users don't understand the effects of bass/midrange modulation distortion on the sound quality of these drivers? Tannoy makes good coaxial drivers that have poor bass/midrange directivity control and modulation distortion issues, but relatively excellent HF directivity control and modulation distortion. Is this why they haven't been more successful in the current marketplace? Is it price, only?
The comment that nc535 said about horn-loading these type of drivers is something that I think is also within bounds of the OP's intent (I would hope): why would someone take a direct radiating-type driver and try to make it into a compression driver for an MEH? It seems counterproductive to me, considering the availability of dual-diaphragm compression drivers at about the same price point. It certainly isn't because of their upper end SPL response for PA duty. Is it only the simplicity of placing a single driver in a box, then making the multiple electrical connections and more complex crossover dividing networks?
Chris
Last edited:
I see those questions as market strategy, user choices/taste etc. While they may be interesting, I focus on the actual technical properties and challenges, and how they can be solved.
The dispersion pattern of a loudspeaker with cone + dome coaxial is different from what you have in a baffle mounted tweeter, or even a waveguide mounted tweeter. One can argue that one is better than the other, but it is more a question of design goal than a quality challenge in coaxial drivers.
Modulation can be easily calculated, and there are relatively clearly defined limits to what level of modulation we can actually hear.
The dispersion pattern of a loudspeaker with cone + dome coaxial is different from what you have in a baffle mounted tweeter, or even a waveguide mounted tweeter. One can argue that one is better than the other, but it is more a question of design goal than a quality challenge in coaxial drivers.
Modulation can be easily calculated, and there are relatively clearly defined limits to what level of modulation we can actually hear.
So do I, but I also wonder about why those drivers are around in such quantity and diversity. That's the "systems engineer" in me to understand the market and driving economic factors in the engineering problem space. It's all one system: the engineering and its business space.
Chris
You are missing that people are sitting in the far field for listening and there on and off axis sound summ together. And - these drivers behave off axis!
What you on axis are strong chancellations like reflections from the surround or some corners in the design. Off axis these are not present.
There are only 2 companies at the moment who perfected the Coax design - KEF with their newer drivers and Genelec with the One's. Both not available for DIY.
As I really like Coax speakers I did some experiments and listened to some DIY Coax systems - they are not made for on axis near field monitoring, you hear the problems in this situation. But in the room, 2-3m distance it works well.
It still bugs me that no manufacturer has figured it out to do a coax similar to KEF ...
I've come across a lot of those while searching for coaxial drivers before. I don't know why they bother. Sure, you still get the advantage of nearly co-incident acoustic centres, but that doesn't seem like a worthwhile trade.
Aside from the much more expensive pro coaxials using compression drivers, there are a few of the type with dome tweeters that measure much better. For instance the Omnes Audio CX 3.1, which I measured as per the attachment. Aside from the nasty resonance from the cone at 6.5kHz, the response is fairly regular.
Aside from the much more expensive pro coaxials using compression drivers, there are a few of the type with dome tweeters that measure much better. For instance the Omnes Audio CX 3.1, which I measured as per the attachment. Aside from the nasty resonance from the cone at 6.5kHz, the response is fairly regular.
Attachments
The KEF LS50 has a pretty comprehensive set of Klippel measurements here (courtesy of Erin with his at-home Klippel nearfield measurement system ). It does very well as long as you don't play it very loud. It's harmonic distortion begins to rise rapidly above 86 dB/1m on-axis, and its compression distortion really begins to take off above 96 dB/1m on-axis, especially in terms of thermal heating (i.e., long-term compression) which is not terribly surprising. It starts to lose directivity control at ~700 Hz, and loses it completely at ~250 Hz. It's a small direct-radiating loudspeaker that's got a very good coaxial driver performance, but, as Mr. IamJF has reported just above, the drivers are not available for DIY.
It's interesting that one of my more capable (in terms of audio engineering and listening) acquaintances bought a pair of LS50s and was happy playing with them in his main listening room. I suppose this is due to their "monitor-like" performance (time domain). This guy also has a pair of the latest La Scala AL5's, a pair of 1st-Gen Jubilees with TAD drivers, and a fairly capable horn-loaded sub made by Klipsch (KPT-1502-HLS), so he's got a fair amount of gear and performance to select from (being a dyed-in-the-wool Klipsch fan). I don't know if he still owns said LS50s but I do know that his discussion of them has dropped off since he talked about them at length on the K-forum website.
It would be nice to have a link or two to said threads (instead of a browbeating on how ignorant we are)...
I recently noticed that any threads over 3 years of age are now locked for further replies--FYI.
Chris
Last edited:
I come from fullrange/augmented-straight-through F.A.S.T. where the highs need to be flattened and lows boosted -- without exception. So even for near-field the coax is no worse in theory.
My brief practical experience with the Tannoy P8 coax diy -- decent not great; this thread explained its limitations. But near-field, out of dozens I only have one pair of speakers with the extension of multi-way and imaging/coherence of fullrange (vintage bicone Supravox TLonken). No choice.
The 15" mentioned earlier was a desktop naked driver pair cloaked in thick wads of stuffing material (like two elves). Hence the need for coax.
My brief practical experience with the Tannoy P8 coax diy -- decent not great; this thread explained its limitations. But near-field, out of dozens I only have one pair of speakers with the extension of multi-way and imaging/coherence of fullrange (vintage bicone Supravox TLonken). No choice.
The 15" mentioned earlier was a desktop naked driver pair cloaked in thick wads of stuffing material (like two elves). Hence the need for coax.
Last edited: