Cone profiles

[Ludvig]

Can someone sort out the benefits/drawbacks of different cone curve profiles? Datasheets talking about straight, exponential, curve linear profiles and so on tells me nothing.

[JonMarsh]

IMO, it comes down to what the designer get's from the profile, and what you're looking for. The arguement is often made that a curvilinear profile will give a more gradual decoupling and spread the breakup modes out over a wider frequency range. Certainly cones like that made with soft materials do exactly that- turn somewhat to mush over a wide range of frequencies.

Myself, I prefer inverted dome profiles, like the HiVi M8a, M6a, and the many Accuton drivers. This is a very stiff, pistonic way to make the cone, and can result with careful design in a lower Q peak.

I wouldn't be concernd too much about claim one way or the other- just how the driver works as regards frequency response linearity, linear distortion and ETC plots, and the Q of the upper breakup modes.

Other factors are basic motor linearity in the lower bass as well as midrange, and of course that other dominating issue, cost.

What are you looking for in midwoofer performance?

[Ludvig]

What are you looking for in midwoofer performance?

Thanks for info.

I'm running a dipole with RS180 at the moment and is very happy with the performance. I am however very interested how many of the pro sound drivers would perform in a careful dipole design for home use. It would be funny to know how their high efficiency and low compression affects the sound compared to typical hifi drivers. Distortion wise I think they are pretty much the same if you go for higher quality pro sound drivers. Both camps know the tricks with focused magnetic fields, copper/alu shorting rings, T-poles and linear suspensions. Left is the cone that always seem to be made of different kind of papers. Paper cones seem to not be so high regarded in the hifi camp. I have never made a direct A/B comparison, so I think I have to try some pro sound drivers and make my own measurements and opinions.

[JonMarsh]

There's nothing like your own experience and observations- beats reading about stuff anyday. Still, the basics, IMO, are pistonic behavior and low energy storage. Not hard to measure, and sometime possible to discern other ways. Tricky thing to watch out for is early modes that manifest as notches or dips. Far field response may look fairly clean; nearfield and impedance curve isn't.

Examples:

Eton 8-800; main peak at 3.2 kHz, but nearfield dip and impedance curve notch at 1400 Hz.

RS225: similar; nearfield notch at 1500 Hz, with impedance curve problem.

RS180: nearfield notch at 2500 Hz, first cone mode. Less pronounced than the above.

SS 7" and 8" have a cone discontinuity typially at 800 Hz which shows up in the impedance curve, though not as a deep notch in nearfield.

If you can't setup for sine burst measurements for ETC, you may find nearfield measurements very helpful, and that can be done with relatively inexpensive tools.

~Jon

[AJINFLA]

Hi Jon,

would you be interested in taking a look at these? : http://www.htguide.com/forum/showpos...4&postcount=68
That's an Alu cone mid forming a WG for a titanium CD. Sens is in the 93db range (higher for the CD I imagine). I'm still waiting on the T/S info from JBL.
A bit different from the usual suspect home drivers IMHO. Unless you consider the TAD coax attainable .

Cheers,

AJ

[noah katz]

Jon,

"Myself, I prefer inverted dome profiles, like the HiVi M8a, M6a, and the many Accuton drivers. This is a very stiff, pistonic way to make the cone, and can result with careful design in a lower Q peak."

Since the cone is accelerated back as well as forth, what is the advantage of inverting it?

"It would be funny to know how their high efficiency and low compression affects the sound compared to typical hifi drivers. Distortion wise I think they are pretty much the same if you go for higher quality pro sound drivers. Both camps know the tricks with focused magnetic fields, copper/alu shorting rings, T-poles and linear suspensions."

Yes. There are some very nice (on paper at least) Eighteensound drivers, high efficiency and breakup free (though rising) response, like this one

http://www.eighteensound.com/index.aspx?mainMenu=view_product&pid=256

They make some 6", but the 8 is the smallest with a shorting ring.

The little I've been able to google on them (18Sound, not that particular driver) has been positive.

[joecarrow]

Since the cone is accelerated back as well as forth, what is the advantage of inverting it?

I thought about that for a little while, and it seems to me that it's probably got to do with how the cone fits with the way the basket structures are generally constructed. Long voicecoil formers would be really heavy.

I bet you could make a woofer with a very shallow mounting depth out of a convex dome. Wouldn't that look cool?

[dlr]

I haven't bought into that yet

There's nothing like your own experience and observations- beats reading about stuff anyday. Still, the basics, IMO, are pistonic behavior and low energy storage. Not hard to measure, and sometime possible to discern other ways. Tricky thing to watch out for is early modes that manifest as notches or dips. Far field response may look fairly clean; nearfield and impedance curve isn't.

Examples:

Eton 8-800; main peak at 3.2 kHz, but nearfield dip and impedance curve notch at 1400 Hz.

RS225: similar; nearfield notch at 1500 Hz, with impedance curve problem.

RS180: nearfield notch at 2500 Hz, first cone mode. Less pronounced than the above.

SS 7" and 8" have a cone discontinuity typially at 800 Hz which shows up in the impedance curve, though not as a deep notch in nearfield.

If you can't setup for sine burst measurements for ETC, you may find nearfield measurements very helpful, and that can be done with relatively inexpensive tools.

~Jon

I am not comfortable with mic placement close to the diaphragm for anything other than low frequency measurements, since the wave is not fully established until some distance out (I"m not familiar with particulars of the wave mechanics) and there can also be phase issues.

WRT to other aspects such as cone geometry and the breakup, I'm more and more of the opinion that the better analysis is in the far-field distortion characteristics. If a curvilinear driver and inverted dome driver are measured under similar situations, if one has a significant issue with cone flex, it has to be apparent in a proper set of distortion measurements. If the distortion does not show adverse effects, then the cone geometry can't be the deciding issue.

There are issues with many drivers that may have more to do with the dust caps and the transition into them at higher frequencies. I experimented briefly with a SS 18W/4531 in that regard. That's where some of it resides. The newer slit-cone SS drivers don't have as much of an issue at 800Hz. But they do still suffer from some non-linear FR problems that I'm guessing has more do to with consistency of manufacturing.

It is as much a function of the effectiveness of the design as I see it. I suspect that it's easier to design and construct an inverted hard-cone for consistency. Or a convex cone for that matter, such as the RS52.

dlr

[jdybnis]

My intuition (FWIW) say that cone flop / decoupling is going to show up in off-axis dispersion measuirements. It's not a distortion thing but still part of the sonic signature. You'd need measurements at lots of angles too in order to see the areas of discontinuity. I haven't seen that sort of measurement and analysis done for any driver. probably impractical without an automated turntable.

[dlr]

Distortion on-axis should be a good indicator

My intuition (FWIW) say that cone flop / decoupling is going to show up in off-axis dispersion measuirements. It's not a distortion thing but still part of the sonic signature. You'd need measurements at lots of angles too in order to see the areas of discontinuity. I haven't seen that sort of measurement and analysis done for any driver. probably impractical without an automated turntable.

Some softer cone drivers take advantage of the fact that there is some flexing coupled with higher internal damping that effectively reduces the radiating area such that the the upper limit is higher and the dispersion characteristics are better. It is a tradeoff as are most aspects of any driver.

If on-axis measurements of distortion are good, then I suspect that the off-axis will be similar, though the even-order distortion distribution should change due to differing response. But if it's good on-axis, I see no reason for it to be bad off-axis.

I'm just not sold on the idea that any one configuration is inherently superior. Each has its own set of advantages/disadvantages.

dlr

[jdybnis]

I didn't mean distortion off-axis. Although that is an idea. I was more thinking in power response and directivity changes that affect the sonic signature. As you say some drivers take advantage of this to improve dispirsion. I understand that this is the approach taken with the Jordan drivers. I suppose if this is impemented to perfection it could be a huge boon. But I suspect in most cases the change in directivity over the frequency range of the "cone flop" is non-monotonic (i.e. it goes up, then it goes down, then it goes back up). This would be audible as a coloration. It would change the reflected sound reaching the listener. But I imagine it would be hard to pinpoint just from listening the cause of that sort of thing.

[dlr]

Whether on- or off-axis, if it's not linear response...

I didn't mean distortion off-axis. Although that is an idea. I was more thinking in power response and directivity changes that affect the sonic signature. As you say some drivers take advantage of this to improve dispirsion. I understand that this is the approach taken with the Jordan drivers. I suppose if this is impemented to perfection it could be a huge boon. But I suspect in most cases the change in directivity over the frequency range of the "cone flop" is non-monotonic (i.e. it goes up, then it goes down, then it goes back up). This would be audible as a coloration. It would change the reflected sound reaching the listener. But I imagine it would be hard to pinpoint just from listening the cause of that sort of thing.

then it is distortion of one form or another. Poorly implemented drivers of any geometry are not going to perform well. The issue is for what might be considered the "ideal", how does one geometry implementation compare to the other.

If the off-axis of a tapered cone has better dispersion and the XO is implented well, then in comparison to another driver geometry that has less dispersion, then all else being equal (the elephant in the room), the one that maintains a more linear response in the on-and off-axis will likely result in better power response. As well, this will likely result in better even-order distortion characteristics, since non-linear response (on- or off-axis) is by definition distortion.

It's just that the off-axis is primarily a contributor to the power response, adding room conditions as a complicating factor.

Even if the change in directivity is monotonic, it will still be considered distortion at any point that the response is non-linear. Whether it's a non-monotonic or monotonic doesn't matter so much as the point at which it occurs and how extreme it is. A smooth, monotonic but rapidly changing off-axis response isn't necessarily better than a non-monotonic, but less rapidly changing response. Both are introducing distortion.

There are not many rules that can't be either broken or mitigated in audio reproduction. It's another case of my favorite saying on speakers, "it depends". There are just way too many variables.

[cotdt]

while there is evidence that highly curved cones have higher breakup nodes, wouldn't that also induce "time smear"? since the sound of the edge of the cone would arrive at your ears before the inner portion of the cone. ???

[dlr]

The overall system response is determinant

while there is evidence that highly curved cones have higher breakup nodes, wouldn't that also induce "time smear"? since the sound of the edge of the cone would arrive at your ears before the inner portion of the cone. ???

I'm not sure what you mean with regard to higher breakup nodes. All dynamic drivers are curvilinear of one kind or another. Alll have a point of breakup and distortion, as does any driver for that matter. What you refer to as "time smear" is essentially what happens in breakup. In the passband this is what contributes to the minimum-phase nature.

The response near the surface of a driver is not the same as it will be at some distance out. It takes some specific distance (I don't know the precise details in the math) before the wave becomes fully established, so the output from all points of the driver will tend to "meld", for lack of a better word, in the passband. As the wavelength generated gets smaller (increasing frequency), the time difference is such that the response no longer "melds". Rather, there is a constructive/destructive interence pattern that is seen as the drooping response. At the same time there will be a point at which the physical movement of the driver is not uniform, flexing.

The off-axis usually makes it more pronounced as the geometric component is usually worsened. But the key is to stay within the usable passband.

All drivers, in essence, reproduce sound with what you refer to as "time smear". This is because they are minimum-phase bandpass devices. That is, there will be a time-delay from one frequency to the next. Small delta in the middle of the passband, larger delta above or below and even larger delta between high and low. This is what defines a driver as minimum-phase, but that's another long topic altogether.

Even flat, full-range drivers such as planar electrostatics exhibit this same phenomenon. But I don't think that the ear is as sensitive to this form of distortion. The other distortion far exceeds this and it's inescapable anyway. The only speaker that would not exhibit this would be one this is linear from DC to infinity. Not going to happen.

Long story short, the "time-smear" (minimum-phase response) of an individual driver is not significant enough to be concerned with it, IMO. The overall measured response will take all aspects into account, SPL and distortion. My take is that the latter is more important than any specific individual geometric characteristic of a driver.

I hope that wasn't too verbose.

[jkrutke]

Myself, I prefer inverted dome profiles, like the HiVi M8a, M6a, and the many Accuton drivers. This is a very stiff, pistonic way to make the cone, and can result with careful design in a lower Q peak.

If I recall, those Hi-Vi inverted dome woofers use an aluminum stiffening assembly between the voice coil and cone. Without that, I suspect there would be too much flexing near the voice coil the response would turn to mush. That appears to be the secret behind getting those cones to behave. I've never owned any of the larger coned Hi-Vi's, do they have those stiffening assemblies? the 3, 4 and 5" versions do.

There is an interesting article written by Peter Larsen regarding Geometrical Stiffness of Loudspeaker Cones. Peter was the guy behind many Vifa, Scan-Speak and Dynaudio drivers. His software was used on the newer Seas L18 to push the breakup from 4kHz up to 7kHz. This article used to be free from the Loudsoft website, but now it's pay-only from the AES. If I downloaded it free previously, I wonder if it's free to distribute if I have it.

http://www.aes.org/e-lib/browse.cfm?elib=12622

[dlr]

Is it this one?

If I recall, those Hi-Vi inverted dome woofers use an aluminum stiffening assembly between the voice coil and cone. Without that, I suspect there would be too much flexing near the voice coil the response would turn to mush. That appears to be the secret behind getting those cones to behave. I've never owned any of the larger coned Hi-Vi's, do they have those stiffening assemblies? the 3, 4 and 5" versions do.

There is an interesting article written by Peter Larsen regarding Geometrical Stiffness of Loudspeaker Cones. Peter was the guy behind many Vifa, Scan-Speak and Dynaudio drivers. His software was used on the newer Seas L18 to push the breakup from 4kHz up to 7kHz. This article used to be free from the Loudsoft website, but now it's pay-only from the AES. If I downloaded it free previously, I wonder if it's free to distribute if I have it.

http://www.aes.org/e-lib/browse.cfm?elib=12622

http://www.loudsoft.com/my_documents/my_files/ALMA%20Paper%202003.pdf

This is at the Loudsoft site.

[jkrutke]

http://www.loudsoft.com/my_documents...per%202003.pdf

This is at the Loudsoft site.

Thanks, my old link didn't work, and I didn't see the new link anywhere.

Too bad there's not a little more info in there. It's more of a "This is what my software does" type document. It would be fun to play with that software for a while.

I think Figure 10 is close to what some Hi-Vi concave cones use.

[dlr]

Accuton must be similar

Thanks, my old link didn't work, and I didn't see the new link anywhere.

Too bad there's not a little more info in there. It's more of a "This is what my software does" type document. It would be fun to play with that software for a while.

I think Figure 10 is close to what some Hi-Vi concave cones use.

I wish I could afford to sacrifice an Accuton to see how they are constructed. They must be similar.

I was hoping to see more detail in the document as well.

[ThomasW]

I shoot a picture of the backside of a M8a it that would be helpful.

[noah katz]

"All dynamic drivers are curvilinear of one kind or another."

Perhaps I've been mistaken all this time, but I always thought curvilinear was distinguished from a cone, which is straight-sided in cross section.

I believe a cone is stiffer with higher freq but more severe breakup, while curvilinear can be tailored to more gradually decouple (by flexing) the outer reaches of the cone from the inner allowing the latter to extend to higher freq.

[dlr]

I lumped them together

"All dynamic drivers are curvilinear of one kind or another."

Perhaps I've been mistaken all this time, but I always thought curvilinear was distinguished from a cone, which is straight-sided in cross section.

I believe a cone is stiffer with higher freq but more severe breakup, while curvilinear can be tailored to more gradually decouple (by flexing) the outer reaches of the cone from the inner allowing the latter to extend to higher freq.

Yes, I was lumping the two together, since (possibly in error) they are both generaly referred to as cone drivers.

My terminology may be in error.