I'm not qualified to give detailed advice, but I have looked into this question. I'll relate some of the impressions I have formed (and standby to be corrected!).
In the past several years, the fact that a speaker needs controlled directivity has been getting a lot more attention. One solution is to make it as omnidirectional as possible (e.g., Linkwitz Pluto), or to make it dipolar (e.g., Linkwitz Orion). Note that the priority is not necessarily the choice of the directivity pattern, but that some thought has been given to controlling directivity over a wide frequency range.
Another approach is to use a very good constant-directivity (CD) horn that is well-matched to the LF speaker at the crossover frequency (e.g., Geddes' designs). I understand that a speaker cone does not make a good CD horn and hence coaxials are not considered ideal for controlled directivity. But they can, apparently, do a "decent" job. B&C speakers provides directivity information for its coaxials. The fact that the speaker cone is not a good horn can also make for a "non-smooth" response for the HF driver.
As for the time-alignment properties - when the LF and HF speakers are at different physical locations, time alignment will be different at different angles and there is no one time delay that will make it right at all directions for all wavelengths. So a coaxial would seem to be a good choice in that regard.
Personally, I like the sound of a big coax that uses a compression driver. I worked in a studio where we used Altec 604s for monitors.
good luck on your quest. You may find that using coaxials will represent some trade-offs and you will have to decide which characteristics are more important to you.
Tom
In the past several years, the fact that a speaker needs controlled directivity has been getting a lot more attention. One solution is to make it as omnidirectional as possible (e.g., Linkwitz Pluto), or to make it dipolar (e.g., Linkwitz Orion). Note that the priority is not necessarily the choice of the directivity pattern, but that some thought has been given to controlling directivity over a wide frequency range.
Another approach is to use a very good constant-directivity (CD) horn that is well-matched to the LF speaker at the crossover frequency (e.g., Geddes' designs). I understand that a speaker cone does not make a good CD horn and hence coaxials are not considered ideal for controlled directivity. But they can, apparently, do a "decent" job. B&C speakers provides directivity information for its coaxials. The fact that the speaker cone is not a good horn can also make for a "non-smooth" response for the HF driver.
As for the time-alignment properties - when the LF and HF speakers are at different physical locations, time alignment will be different at different angles and there is no one time delay that will make it right at all directions for all wavelengths. So a coaxial would seem to be a good choice in that regard.
Personally, I like the sound of a big coax that uses a compression driver. I worked in a studio where we used Altec 604s for monitors.
good luck on your quest. You may find that using coaxials will represent some trade-offs and you will have to decide which characteristics are more important to you.
Tom
I would not want to influence your current opinion on this matter and I wish you success. I will say though that every manufacturer has to have a point of view on design aspects in order to achieve a sale which is a holy grail of starting a company in the first place, all else is talk as justification. We know we can have superbly sounding PA systems, home hifi classic rigs, WAW systems (Scott will be happy about this one), large/small co-axial builds.
edit: Hugo, those look like ultimate hoax speakers, lol.
Last edited:
OK, When you think so. People's beliefs and persistent habits are always biggest bottleneck in audio.
DIY'ers avoid them like a fire from points me and you have just mentioned above. There are already simply two or three DIY projects published to the world. It's a pity.
this is THE ultimate:
Voxativ Loudspeakers: AC - XHB Hybrid Field Coil
Voxative are the ultimate in cost, but can you play complex classical music or Pop, loud using them? Not from what I have heard.
Voxativ Loudspeakers: AC - XHB Hybrid Field Coil
Voxative are the ultimate in cost, but can you play complex classical music or Pop, loud using them? Not from what I have heard.
I've built a version of Jon Marsh Minerva studio monitor speaker with these Seas coax drivers. There are measured and simulated results among the build pics here if interested.
DIY - Minerva monitor build | Flickr
These are my best speakers yet for sound and imaging, but then again there are so many drivers and technologies I've not tried yet. Also it is true that the Xover for this driver tends to get quite complex in order to tame all the irregularities, but can be done as seen on final results.
DIY - Minerva monitor build | Flickr
These are my best speakers yet for sound and imaging, but then again there are so many drivers and technologies I've not tried yet. Also it is true that the Xover for this driver tends to get quite complex in order to tame all the irregularities, but can be done as seen on final results.
Last edited:
Your comment is a bit abstract to me but I am interested in what you mean in practice - do you think e.g. these irregularities (which plainly are there) do not matter too much for the sound and if so why would you say so ?
The new KEF Q350 is a modestly priced coaxial speaker. It's xo circuit could be used as a starting point for a 3-way construction. Very easy to do as semiactive with dsp-plate amp. 650$/pair! Don't get scared about the response linked, this is how a really good and natural sounding speaker measures!
KEF Q350 loudspeaker Measurements | Stereophile.com
KEF Q350 loudspeaker Measurements | Stereophile.com
Do you refer to this private small talk that I had with C. Cabasse ? If so I can confirm I really appreciate this kind of opinion expressed when there is no marketing interest in promoting any "sales" driven philosphy (a private talk with an ordinary visitor). On the other hand, with all due respect, people from the industry doing advanced projects have experience diffilcult to find in non-commercial/DIY community.
However, it is not the opinion of Mr Cabasse that makes me think coaxials or fullranges should be a necessary component of a highest fidelity loudspeaker. Yes, there might be plentitude of superbly sounding speakers, particularly for their builders. It is just that my focus is on the technologies/components that allow to achieve the best possible parameters. Not typical parameters but these, as we have been learning recently, are especially important for our perception of sound/music. One of the key things is that the most fundamental parameter - linearity of frequency response - is less important than commonly believed.
For the first years of my experiments with DSP in audio I was fascinated by the unique possibilities that it offers. E.g. very sharp roll-offs, linear phase (FIR-s). It did not sound bad but somehow in listening tests it could not beat the "primitive" analog xovers. I observed continuous quality improvement when focused more on the timing aspects in loudspeakers. It is also backed by growing audio research publications.
On-axis response is only one small thing to consider and unfortunately the easiest thing to spot by unexperienced DIY-ers. This particular driver exhibits very nice performance at another crucial aspects - controlled directivity and even power response in function of frequency. Well-behaving driver in these regards just fills-out typical listening rooms with balanced acoustic power and early reflections are very consistent tonally with direct sound.
Unconditional benefit of coaxial driver is not just the even power response but especially good 'imaging' coming from single sound source. Even more important for this particular Seas coaxial driver is the fact that midrange cone is designed from the beginning as precision waveguide for tweeter. This was overlooked by many designers of coaxial drivers but not for Seas. They intentionally call it 'midrange' and tamed its low-frequency range to 200 Hz or so to avoid modulation of high-frequency band resulting from midbass cone travelling around the tweeter. What is more, special cone shape and 'stealth' upper suspensions minimise diffraction issues and linearise off-axis response to a great extent.
These little diffractive effects are still present at the FR as you called them 'irregularities' especially at 2.2k and at 9k coming from driver's suspensions. They are almost impossible to hear. If it's a concern for you, let's toe-in the drivers 15 degree and problem solved elegantly with additional benefit from broad sweet-spot unachievable for designs with narrow or un-controlled dispersion. Right away main axis diffraction issues quickly disappear and leave you with pure point-source, controlled directivity design of a great high-frequency dispersion unachievable with typical soft-dome tweeters. This drivers possess almost no resonances and energy storage problems which are much more serious to listener than axial diffraction.
Their magnesium cones act as ideal pistons throughout their pass bands. Efficiency is pretty good for the midrange domestic cone so power demand is low and power compression unlikely. Hi-frequency driver works inside of waveguide and it boosts its efficiency in critical range around 2k. Midrange membrane is lighter than in many midwoofers this size, which is good for midrange reproduction as to many beliefs. Parameters are very consistent, this is top-quality production. This driver was tested by Troels Gravesen and Vance Dickason from AudioXpress and they find nothing to complain. What is more, unlikely to typical coaxials, these drivers exhibit amazingly low distorion throughout broad frequency range and no break-up issues in their usual pass-bands is seen. It's a pity that there is so little such an amazing drive units available to DIY'ers.
Last edited:
KEF Q150 is also worth considering, depending on which size of speaker you want. Look at this for inspiration www.audioexcite.com >> Concentric Three – M5
Not sure if this is helpful or not but... The SEAS drivers are close to what KEF did a few decades ago. In fact, SEAS used to pay KEF a royalty. KEF have developed their coaxials further but SEAS do not appear to have done so in any significant way. They are decent coaxials but not state-of-the-art. They are also expensive given the significant on-axis diffraction effects. Reducing this to a low level is one of the distinguishing features of a modern state-of-the-art coaxial.
I have considered using coaxials like the SEAS but the main thing that puts me off is the complexity of the crossovers you seem to need to use to tame unruly frequency response.
For example if you use the PHL1240TWX-M a recommended crossover for just the midrange section uses 11 caps,chokes and resistors then there is another 9 in the woofer circuit.
So my concern is that any natural benefits that come from this arrangement might be negated by having to use such complex crossovers.
For example if you use the PHL1240TWX-M a recommended crossover for just the midrange section uses 11 caps,chokes and resistors then there is another 9 in the woofer circuit.
So my concern is that any natural benefits that come from this arrangement might be negated by having to use such complex crossovers.
http://www.madisound.com/pdf/L12RE_XFC_AppNote.pdf
Trenner & Friedl Sun loudspeaker | Stereophile.com
Not ultimate but not expensive and has had solid reviews and quietly is offered in DIY form.
You could make a 3 way from this, though efficiency would be low.
This could be the ultimate: long thread, read carefully. Not cheap, but fully developed.
Minerva Monitor: "Patience my ***, I'm going to go build something!"
Trenner & Friedl Sun loudspeaker | Stereophile.com
Not ultimate but not expensive and has had solid reviews and quietly is offered in DIY form.
You could make a 3 way from this, though efficiency would be low.
This could be the ultimate: long thread, read carefully. Not cheap, but fully developed.
Minerva Monitor: "Patience my ***, I'm going to go build something!"
Last edited:
This a bit industry insider view that you bring is very helpful for me. As for your eariler interesting info about availability of KEF replacement drivers I have seen somewhere that is rather limited, you need to quote the serial number. Coaxials have very strong advantages that no DSP or all the more no matter how complex analog circuitry can ever imitate. That's why not using a coaxial design means for me right from the start we are not in the ultimate loudspeaker game although it can still be superb sounding
By the way as an example from speaker crossovers my long time DSP experimenting and research has proved that top louspeaker performance requires 1-st order for the low pass crossover realization. So ironically I have a very complex DSP processing only to achieve effectively 1-st order Butterworth LP filtering although without its unavoidable problems when done in analog.That is why an optimum driver design for me would be actually wide band mid-driver and coaxial tweeter. So the more extended the midwoofer response noth to tha base but also usable frequency to the highs, the better. Actually this extended civilised high frequency response also typically means better transient and impulse/step response.