Coaxial vs non-coaxial... (paging DS-21!)...

[lunchmoney]

DS-21's post in a previous thread has me once again wondering about the pros/cons of coaxial vs non-coaxial... I know I asked this before, but it wasn't a very thorough conversation... and he wasn't in on it I don't believe...

From DS-21:

"...anything that gets people past the asinine paradigm of a dome tweeter mounted flush in a baffle with highly audible power response errors due to the inevitable directivity mismatch between tweeter and next driver down in the crossover region is a net plus for the hobby."

In the past (and here) he has vented quite strongly about how coaxial configurations are superior to the non-coaxial configuration that 99% of people tend to make (including myself)...

And he seems to know his stuff, so it's got me curious... tell me again why don't more people make coaxial designs?

I'll go make popcorn. Play nice, kids ;)

[kevinhaskins]

Re: Coaxial vs non-coaxial... (paging DS-21!)...

DS-21's post in a previous thread has me once again wondering about the pros/cons of coaxial vs non-coaxial... I know I asked this before, but it wasn't a very thorough conversation... and he wasn't in on it I don't believe...

From DS-21:

"...anything that gets people past the asinine paradigm of a dome tweeter mounted flush in a baffle with highly audible power response errors due to the inevitable directivity mismatch between tweeter and next driver down in the crossover region is a net plus for the hobby."

In the past (and here) he has vented quite strongly about how coaxial configurations are superior to the non-coaxial configuration that 99% of people tend to make (including myself)...

And he seems to know his stuff, so it's got me curious... tell me again why don't more people make coaxial designs?

I'll go make popcorn. Play nice, kids ;)

Because most of the off-the-shelf solutions suck. Coax has been popular for in-wall, car audio, and low-cost solutions so there just are not many high-quality solutions on the market. Also... you don't magically erase the driver dispersion mis-match. The tweeter is still a 1" or 0.75" device and the midwoofer that it is sitting on is still a 5.25", 6.5" or a 8" device. The dispersion is the same as it is flat on a baffle. What improves, is that you no longer have the path length difference on the vertical axis of measurement. At least not to the degree you would have with devices mounted top/bottom.

You can do a waveguide, to control the tweeter dispersion and improve performance and directivity up high, but there isn't much you can do about the midwoofer. The trick would be to have a device that could crossover low, matching as close as possible the directivity of the midwoofer.

[Æ]

Re: Coaxial vs non-coaxial... (paging DS-21!)...

Coaxials behave more like a true "point source" hence easy to obtain good imaging. Less placement difficulties.

DS-21's post in a previous thread has me once again wondering about the pros/cons of coaxial vs non-coaxial... I know I asked this before, but it wasn't a very thorough conversation... and he wasn't in on it I don't believe...

From DS-21:

"...anything that gets people past the asinine paradigm of a dome tweeter mounted flush in a baffle with highly audible power response errors due to the inevitable directivity mismatch between tweeter and next driver down in the crossover region is a net plus for the hobby."

In the past (and here) he has vented quite strongly about how coaxial configurations are superior to the non-coaxial configuration that 99% of people tend to make (including myself)...

And he seems to know his stuff, so it's got me curious... tell me again why don't more people make coaxial designs?

I'll go make popcorn. Play nice, kids ;)

[dlr]

Re: Coaxial vs non-coaxial... (paging DS-21!)...

Because most of the off-the-shelf solutions suck. Coax has been popular for in-wall, car audio, and low-cost solutions so there just are not many high-quality solutions on the market. Also... you don't magically erase the driver dispersion mis-match. The tweeter is still a 1" or 0.75" device and the midwoofer that it is sitting on is still a 5.25", 6.5" or a 8" device. The dispersion is the same as it is flat on a baffle. What improves, is that you no longer have the path length difference on the vertical axis of measurement. At least not to the degree you would have with devices mounted top/bottom.

You can do a waveguide, to control the tweeter dispersion and improve performance and directivity up high, but there isn't much you can do about the midwoofer. The trick would be to have a device that could crossover low, matching as close as possible the directivity of the midwoofer.

Partially correct, meaning partially wrong. The dispersion of a coaxial tweeter is nothing like that of the same tweeter on a flat baffle. That would say that a horn/waveguide is totally ineffective. We know that's not the case, so the question (more than one, really) is, what is the impact?

lunchmoney, for most solutions on the market available to DIY, it's not good. The only coaxes that don't totally muck up the tweeter are way too expensive for most to consider. The directionality is controlled to some degree, but the detrimental effects of typical ones far outweighs the benefits in most cases. One need only go to DIYAudio and read through threads on Geddes' OS waveguide to see just what conditions arise.

In the end, coax tweeters in smallish "horns" (read midwoofers) are not fixable. I've tried. Decent sound is possible, but the linear distortions due to a bad profile of the "horn" and possibly as well as its dynamic modulation of the tweeter output (Doppler distortion) just cannot be corrected. Believe me, I've tried. Just look at the responses from the Insignia at site.

The idea of using as low an Fc as the tweeter can handle is a good idea for smooth off-axis response (horizontal only, of course) of standard drivers. At this point multiple drivers is still the best solution unless one has enough money to get a well-designed coax. IMO, of course.

dlr

[brianpowers27]

Re: Coaxial vs non-coaxial... (paging DS-21!)...

I like the idea of using a 4" coaxial with a larger woofer. The 4" could be mounted behind the baffle to attempt to align the AC. A 4" woofer has pretty decent dispersion.

I guess there arent a lot of uber-clean options out there but the dispersion and the relative phase of off axis signals may just make up for a smaller loss of distortion cleanliness.

[PunkSweeper]

Re: Coaxial vs non-coaxial... (paging DS-21!)...

I like the idea of using a 4" coaxial with a larger woofer. The 4" could be mounted behind the baffle to attempt to align the AC. A 4" woofer has pretty decent dispersion.

I guess there arent a lot of uber-clean options out there but the dispersion and the relative phase of off axis signals may just make up for a smaller loss of distortion cleanliness.

YES. IMHO, the optimal 2-way is a mid tweeter and a woofer, woofer crossed over lower than 100hz possible (to get as close as possible to point-source, with 100hz, being omni directional, blah blah blah) and the ideal 3-way is coaxial mid/tweet and woofer, to diminish Doppler modulation in the treble by limiting mid excursion.

[Face]

Re: Coaxial vs non-coaxial... (paging DS-21!)...

At this point multiple drivers is still the best solution unless one has enough money to get a well-designed coax. IMO, of course.

Who would you recommend, Tannoy, KEF? How about this from SEAS? http://www.madisound.com/catalog/pro...oducts_id=8494

[dlr]

Re: Coaxial vs non-coaxial... (paging DS-21!)...

Who would you recommend, Tannoy, KEF? How about this from SEAS? http://www.madisound.com/catalog/pro...oducts_id=8494

I wouldn't touch the Seas, especially at that price. They haven't done anything over their original unit to improve the "horn" profile. It will just have a lot less motor distortion, but also be more difficult due to that monstrous peak at 5.5K.

I'm not sure what I'd try, I haven't looked carefully. Tannoys do look good and the Thiel and Partner available in Germany look good, just way too expensive.

As an example of just how destructive a bad "horn" is, look at the Insignia tweeter removed from the unit and measured on a 2m x 2m baffle:



he above has no smoothing applied at all. Now go look at the measurement of the tweeter in-driver and in-box:



This one says 1/3 octave smoothing , but re-displaying in Praxis uses the default unsmoothed, so this should be with no smoothing. Edit: Looking back at my other measurements, this actually is the smoothed version.

dlr

[kevinhaskins]

Re: Coaxial vs non-coaxial... (paging DS-21!)...

Partially correct, meaning partially wrong. The dispersion of a coaxial tweeter is nothing like that of the same tweeter on a flat baffle. That would say that a horn/waveguide is totally ineffective. We know that's not the case, so the question (more than one, really) is, what is the impact?

lunchmoney, for most solutions on the market available to DIY, it's not good. The only coaxes that don't totally muck up the tweeter are way too expensive for most to consider. The directionality is controlled to some degree, but the detrimental effects of typical ones far outweighs the benefits in most cases. One need only go to DIYAudio and read through threads on Geddes' OS waveguide to see just what conditions arise.

In the end, coax tweeters in smallish "horns" (read midwoofers) are not fixable. I've tried. Decent sound is possible, but the linear distortions due to a bad profile of the "horn" and possibly as well as its dynamic modulation of the tweeter output (Doppler distortion) just cannot be corrected. Believe me, I've tried. Just look at the responses from the Insignia at site.

The idea of using as low an Fc as the tweeter can handle is a good idea for smooth off-axis response (horizontal only, of course) of standard drivers. At this point multiple drivers is still the best solution unless one has enough money to get a well-designed coax. IMO, of course.

dlr

Your assuming that the tweeter is at the bottom of a conical type horn that is the result of a typical cone profile. That isn't the only solution.

There are ways of using the tweeter such that you don't use a moving cone as the waveguide. With proper design, you don't modulate the output of the tweeter. Don't ask me how I know. ;)

[Face]

Re: Coaxial vs non-coaxial... (paging DS-21!)...

I wouldn't touch the Seas, especially at that price. They haven't done anything over their original unit to improve the "horn" profile. It will just have a lot less motor distortion, but also be more difficult due to that monstrous peak at 5.5K.

I'm not sure what I'd try, I haven't looked carefully. Tannoys do look good and the Thiel and Partner available in Germany look good, just way too expensive.
dlr

I've been curious to try a pair of these drivers: http://83.138.162.162/products/273/D..._V1.06_Web.pdf

There are ways of using the tweeter such that you don't use a moving cone as the waveguide. With proper design, you don't modulate the output of the tweeter. Don't ask me how I know. ;)

Something like this?

[kevinhaskins]

Re: Coaxial vs non-coaxial... (paging DS-21!)...

I've been curious to try a pair of these drivers: http://83.138.162.162/products/273/D..._V1.06_Web.pdf

Something like this?

I'm not familiar with that driver but I'd say no. Hard to tell what they are doing inside the pole. Some of these compression driver horns mounted in the pole don't have FR modulated by the midrange/woofer but they don't match up well with the driver they are mounted within necessarily either. They also have fairly ugly FR compared to your typical dome tweeter.

I'm working on something but it will be a couple more months before I either call it a failure or release more information. Right now I'm fairly bullish on the concept. It shouldn't be expensive either as there is only one major tooling cost associated with it.

[dlr]

Re: Coaxial vs non-coaxial... (paging DS-21!)...

Your assuming that the tweeter is at the bottom of a conical type horn that is the result of a typical cone profile. That isn't the only solution.

There are ways of using the tweeter such that you don't use a moving cone as the waveguide. With proper design, you don't modulate the output of the tweeter. Don't ask me how I know. ;)

I never said it was the only solution.

You said "The dispersion is the same as it is flat on a baffle." I said that it is not. That's pretty much indisputable.

I also said "for most solutions on the market available to DIY, it's not good." I stand by that, as this is in line with my first sentence above, most solutions available to the DIY community. I have some KEF drivers that are true midrange/tweeters that are a much better solution. This removes the moving midwoofer diaphragm while simultaneously providing a much better profile, so I know that there are better solutions than the Seas and similar. I'll say again, I wouldn't consider the new Seas as it has the typical drawbacks of the commonly available coaxes.

dlr

[DS-21]

Re: Coaxial vs non-coaxial... (paging DS-21!)...

And he seems to know his stuff, so it's got me curious... tell me again why don't more people make coaxial designs?

Well, first let's make some distinctions. Coax is a large set, and most of them are awful. Let's break coax into five types:

1) Tweeter mounted on a stick or a bridge: bad. Usually seen in car audio but sometimes also in speakers by Linn and others.

(no idea what that is; first hit on google)

2) Tweeter mounted behind the woofer motor, firing through a horn or waveguide that's proud of the woofer cone. I don't know, as I haven't really heard enough from such speakers. In theory, they should be worse because the midrange reflects on the cone, and there should be lots of diffraction from the exposed WG or horn. The best-known examples are the vintage Altec 604 and progeny, but this design seems to be enjoying a resurgence. B&C and BMS have both recently released new ones.

BMS 15C682

3) Coincident drivers, which have a direct-radiating tweeter mounted on top of the polepiece and use the woofer cone as a waveguide. The KEF Uni-Q is the best known. Seas, Thiel, TAD/Pioneer Elite, and probably others also use it. They can work well, at least in small-scale systems. Generally these use small woofers and high crossover points, so they're really only good for nearfield systems or very small rooms. Too much dynamic compression for truly full-bore systems unless they're specially designed as midranges and crossed over fairly high to helper woofers.

KEF Uni-Q exploded view

4) Concentric/coplanar drivers. These drivers have two flat-panel or ring-radiator drivers that are concentric. The BMS double compression drivers fit this mold, as do the Thiel CS3.7's upper compound driver and the HiVi S1 mid/tweeter. A monitor that was highly reviewed in The Audio Critic that I'd love to hear someday, the Win SM10, used this configuration. I guess, one could sort of say the Quad ESL-63 fits somewhat in this category even though it's a single electrostatic panel, due to the concentric delay lines. The direct-radiating variants of this driver type do not have the power response advantages as #3 and #5.

Thiel CS3.7 driver

5) Dual Concentric drivers. These drivers have the tweeter diaphragm mounted behind the woofer motor, and it fires through a waveguide drilled through polepiece of the woofer, terminating at the woofer cone. The best known example is the Tannoy Dual Concentric. BMS, Beyma, B&C, Radian, Eminence, and others make such drivers, with varying degrees of sophistication.

Tannoy Dual Concentric cutaway

Second, my claim isn't necessarily that coaxes (#3 and #5) are superior. My claim is nothing more or less than the following:

matching the directivity of the tweeter at the bottom of its passband and the directivity of the driver immediately under in and around the crossover region it is a necessary but not sufficient condition for high-fidelity reproduction.

Coincidents and Dual Concentrics are ways to do that end. They have some disadvantages compared to speakers with separate waveguides*, especially in the top octave. They have two advantages: packaging efficiency (especially useful for front trios of home theaters) and symmetrical horizontal/vertical power response. The latter is important if one cares about the sound quality while standing as well as while sitting. Many don't, with good reason, though I personally do. It's more a "how do I use my system" issue than anything else.

As for why people don't do it more...

Because most of the off-the-shelf solutions suck.

This is largely true, and as dlr mentions the ones that are available are seriously expensive. That is why I ended up modifying Tannoy studio monitors (System 12 DMT II) for my living room system, and using stock Tannoy nearfield monitors (System 8 NFM II) of my home systems, rather than going DIY from scratch.

But Kevin, you're in a better position than most to fix that! Especially since right now you're one of the leading purveyors of extremely high performance value priced product to the audio hobbyist right now. (Yes, I still love my Mael-X even though an Aura driver recently replaced in my main system. It's currently rocking the nearfield system.)

Remember the XBL^2 variant of the Eminence Beta10CX Dan Wiggins was talking about for a while? (I put my "ultimate main system" plans on hold for years waiting for those drivers to come out.) How about a 10" Exodus Audio HE10.1-X or 12" Exodus Audio HE12.1-X with a bespoke midwoofer using everything you've learned about XBL^2 motors, perhaps with a BMS compression driver behind it?

In the end, coax tweeters in smallish "horns" (read midwoofers) are not fixable. I've tried. Decent sound is possible, but the linear distortions due to a bad profile of the "horn" and possibly as well as its dynamic modulation of the tweeter output (Doppler distortion) just cannot be corrected. Believe me, I've tried. Just look at the responses from the Insignia at site.

Then again, don't judge a category by a cheap example. Look at some of KEF's recent designs, or Tannoy's. Or the Gradient Revolution, which uses a bespoke variant of the Seas driver.

[dlr]

Re: Coaxial vs non-coaxial... (paging DS-21!)...

Then again, don't judge a category by a cheap example. Look at some of KEF's recent designs, or Tannoy's. Or the Gradient Revolution, which uses a bespoke variant of the Seas driver.

The focus is what's readily available to the DIY community and at what cost. The prime and better known examples, including one very new one, are those of Seas. Just not good.

dlr

[kevinhaskins]

Re: Coaxial vs non-coaxial... (paging DS-21!)...

I never said it was the only solution.

You said "The dispersion is the same as it is flat on a baffle." I said that it is not. That's pretty much indisputable.


dlr

I'll dispute it.

Flat cone, with a dome stuck in the pole will have the same dispersion as one on a flat MDF baffle. You might get a wiggle or two from diffraction on-axis but it has the same overall dispersion as a normal device on a flat MDF baffle.