soongsc said:
As I said, until it is verified that the phase can not be reduced to a single curve by adding or removing a small delay I don't believe that this is anything but a different delay.
The bottom line is that there are many ways to treat a cone, particularly a metal cones, to add damping which will reduce resonances at breakup, or add mass to change the characteristics of the breakup. When ever something is added to the cone the mechanical properties are changed. If you apply a material in a 'disruptive' pattern then what is done is the ideally homogeneous material is now made locally inhomogeneous. With breakup there are two types: Radial or concentric mode and circumferential or bell modes. Radial modes occur in all cones and are related to the wave length relative to the center to surround distance. In a homogeneous cone there is noting you can do to eliminate these resonances other than add damping (like poly cones) and/or try to control reflections from the surround and the point of VC attachment. These locations are positions of impedance mismatches and they lead to reflections. Circumferential modes are basically mechanical instabilities. In an ideal inhomogeneous cone they would not appear unless some destabilizing factor was introduced. Real cones always have some inhomogeneity because of the nature of things. In particular there are always inhomogeneities at the VC/cone and cone/surround junctions. Introducing some type of additional inhomogeneity in these areas could change the behavior. I don't doubt that. But I don't think it’s anything magical by applying it in a particular pattern. I suspect a squiggly line of the same material, or a different material, applied to your cones would also work. As you pointed out, the process you used isn't a magic bullet:
But I am off topic here. Back to your phase observation, as I said above, before I would make any ascertains about the differences I would look to see if the phase can be collapsed to a single curve by altering the delay. I believe we discussed this before and found that it could be. It's even possible that the treatment you added changed the acoustic center slightly due to the small added mass.
It seems that we are in agreement regarding this aspect. I can aonly speculate that the change in phase is caused by disruption of wave flow over the surface because the more inward the pattern is applied, the more phase lead. If I ever have the chance to do the test again, I probably would try a few mics at different angles to see of the radiation pattern changes or not.john k... said:
Well for all I understand the Enabl pattern is meant to alleviate the impedance mismatch at the surround end by terminationg the line properly; here two mechanisms come to mind, stiffness and mass, to that effect. As to why phase shoud be different I don't know, but this may depend on the interaction with the way the FFT bins fall before/after treatment.
The slit cones of ScanSpeak are another way to address the cone modes, here by intriducing internal impedance mismatches if I "read" the intention correctly (never saw a white paper on it) and probably shift the resonances upwards in frequency, possibly to a region where they can be damped effectively.
The slit cones of ScanSpeak are another way to address the cone modes, here by intriducing internal impedance mismatches if I "read" the intention correctly (never saw a white paper on it) and probably shift the resonances upwards in frequency, possibly to a region where they can be damped effectively.
Not convinced on EnABL, but the SS approach works
I suspected that they were stiffening the cone to reduce flexing, but without having to treat the entire surface, minimizing any added mass. If they shift the resonances as you say, and that seems logical, then the arrangement is distributing those resonances to minimize the influence at any given frequency. Seemed pretty ingenious to me when I first saw them.
Likewise, the new surrounds that are treated similarly (Seas? I forget the manufacturer) must work in a similar fashion. I tried treating surrounds with damping in a pattern, but aligned radially, to no avail.
Dave
MBK said:
I suspected that they were stiffening the cone to reduce flexing, but without having to treat the entire surface, minimizing any added mass. If they shift the resonances as you say, and that seems logical, then the arrangement is distributing those resonances to minimize the influence at any given frequency. Seemed pretty ingenious to me when I first saw them.
Likewise, the new surrounds that are treated similarly (Seas? I forget the manufacturer) must work in a similar fashion. I tried treating surrounds with damping in a pattern, but aligned radially, to no avail.
Dave
I got the idea as to what the slits *might* be doing, from this page on
fractal drums
No idea if they had this in mind or if they discovered it independently.
Later it occurred to me that it doesn't have to be that deep, it is likely sufficient to break up the path that reflections would take, into smaller subsections, which must then have a higher resonant frequency. Had they wanted to just stiffen the cone, they should have chosen ridges instead of slits. Come to think of it, why doesn't anyone do this? Two birds with one stone - stiffer cone, compartmentalized cone modes. Say a network of ridges, such as you find it stamped on large tin cans to stiffen the side walls, doesn't even increase mass, but it increases stiffness, and likely introduces the same barrier to reflections.
fractal drums
No idea if they had this in mind or if they discovered it independently.
Later it occurred to me that it doesn't have to be that deep, it is likely sufficient to break up the path that reflections would take, into smaller subsections, which must then have a higher resonant frequency. Had they wanted to just stiffen the cone, they should have chosen ridges instead of slits. Come to think of it, why doesn't anyone do this? Two birds with one stone - stiffer cone, compartmentalized cone modes. Say a network of ridges, such as you find it stamped on large tin cans to stiffen the side walls, doesn't even increase mass, but it increases stiffness, and likely introduces the same barrier to reflections.
I don't think it's the same phenomenon
The link shows a geometric shape boundary, not a geometric shape cut into another boundaried shape as the SS driver represents. The concepts may be similar, but I think that the Vifa "NRC" cones are closer to the link idea, though a simplistic one. Vifa cut several straight edges to the outer periphery of the cone that alters the path length from former to the boundary edge. I'm not sure how effective it really is.
Those slits likely serve dual purposes. They alter the path length and distribute it. The slit is sealed with a damping compound as well. This is better, IMO, that simply stiffening the cone. The ridges, as Few shows in the link below, will strengthen, but it also likely adds more mass than the slits and doesn't add any damping.
I think that the Thiel example does add mass. It's not a flat cone. The physical surface area is larger than a flat cone, even if the effective Sd is not altered as I am guessing is the case. I don't see any barrier to reflections at all.
Being a metal cone, the added stiffen should push the first breakup mode even higher, though the limit to effectiveness will be the wavelength in relation to the depth of the cone. It looks like the Thiel approach allows for a much shallower cone. I suspect that's the primary reason for it.
Dave
MBK said:
The link shows a geometric shape boundary, not a geometric shape cut into another boundaried shape as the SS driver represents. The concepts may be similar, but I think that the Vifa "NRC" cones are closer to the link idea, though a simplistic one. Vifa cut several straight edges to the outer periphery of the cone that alters the path length from former to the boundary edge. I'm not sure how effective it really is.
Those slits likely serve dual purposes. They alter the path length and distribute it. The slit is sealed with a damping compound as well. This is better, IMO, that simply stiffening the cone. The ridges, as Few shows in the link below, will strengthen, but it also likely adds more mass than the slits and doesn't add any damping.
I think that the Thiel example does add mass. It's not a flat cone. The physical surface area is larger than a flat cone, even if the effective Sd is not altered as I am guessing is the case. I don't see any barrier to reflections at all.
Being a metal cone, the added stiffen should push the first breakup mode even higher, though the limit to effectiveness will be the wavelength in relation to the depth of the cone. It looks like the Thiel approach allows for a much shallower cone. I suspect that's the primary reason for it.
Dave
The drum is obviously built differently, and it doesn't have to be a fractal for that matter. BUt the reason why the drum does what it does, seems to be similar to me - compartmentalization.
Ridges: I meant something more subtle, like a grid of triangul;ar cross section stamped ridges of minimal height, say 1-2 mm; I would think this might act as a barrier to a wave (impedance mismatch). Of course a slit with damping has other advantages.
Ridges: I meant something more subtle, like a grid of triangul;ar cross section stamped ridges of minimal height, say 1-2 mm; I would think this might act as a barrier to a wave (impedance mismatch). Of course a slit with damping has other advantages.
john k... said:
Like Edison's great efforts to invent the light bulb, Bud has 20-some years of experimentation under his belt finding out what patterns don't work. This one does. And at a level i find quite disturbing (after you have lived for over 40 years getting used to the "hall-of-mirrors", having it (mostly) dissapear is disturbing at 1st, and after that it makes it hard to listen to anything else)
And just as more research has improved the incandesant light bulb, i am sure further research could yield improved patterns. But 1st people need to start paying attention to it.
dave
gedlee said:
My experience with the person in question was similar, and I should mention that it was he that approached me and solicited my technology/design for his application(s).
Personally, I am always surprised and amazed that people who behave in this sort of way seem to not only be sucessful but seem to make quite a lot of money. Obviously, I am a dolt.
_-_-bear
Russell Dawkins said:
IMHO, Emerald Physics had the best sounding speaker at CES this year.
I spent the end of 2006 building a Unity clone - I'm thinking about spending this winter building something like the Emerald Physics.
If anyone is interested in seeing me do this, let me know. The last thing I need is another set of speakers, but I'd love to collaborate with you guys on a speaker PROJECT. It's a lot more fun with you guys egging me on
Here's what I posted about the speakers a few months back:
http://www.diyaudio.com/forums/showthread/t-104441.html
Don't mind me, I'm having fun following this thread.
Peeking over at the EnABL thread, it has occurred to me that maybe the subjective increase in resolution is correlated with a drop in distortion at frequencies in the breakup region.
The strong love/hate reactions to Lowther and Fostex drivers has everything to do with their quite pronounced breakup/resonance regions, which are much less "damped" and smoothed-out than more conventional, better-measuring drivers. I've never been a fan of Fostex drivers, but I've been surprised by Lowthers several times - they almost remind me of the fabled Decca cartridge in temperament, sometimes wonderful, other times, not nice at all. When they go sour, they go very sour indeed, typically with choral music at high levels. But when they're good, even the electrostat-lovers (like me) have to grudgingly admit there's something pretty special going on.
If the EnABL treatment does little to decrease to the magnitude of the peaking, but reduces the distortion at those frequencies, I'd expect a change in sound character, and possibly, the increase in headroom the auditioners are reporting. More comprehensive measurements are definitely needed. Just changing the good/bad ratio of the Lowthers and Fostexes would be an astounding achievement.
P.S. After all this buildup, yes, of course I want to listen for myself. Only a few more days ...
Peeking over at the EnABL thread, it has occurred to me that maybe the subjective increase in resolution is correlated with a drop in distortion at frequencies in the breakup region.
The strong love/hate reactions to Lowther and Fostex drivers has everything to do with their quite pronounced breakup/resonance regions, which are much less "damped" and smoothed-out than more conventional, better-measuring drivers. I've never been a fan of Fostex drivers, but I've been surprised by Lowthers several times - they almost remind me of the fabled Decca cartridge in temperament, sometimes wonderful, other times, not nice at all. When they go sour, they go very sour indeed, typically with choral music at high levels. But when they're good, even the electrostat-lovers (like me) have to grudgingly admit there's something pretty special going on.
If the EnABL treatment does little to decrease to the magnitude of the peaking, but reduces the distortion at those frequencies, I'd expect a change in sound character, and possibly, the increase in headroom the auditioners are reporting. More comprehensive measurements are definitely needed. Just changing the good/bad ratio of the Lowthers and Fostexes would be an astounding achievement.
P.S. After all this buildup, yes, of course I want to listen for myself. Only a few more days ...
Lynn Olson said:
choral music even not at so high levels (in my experience)
and not only choral music, also strings, especially in case of bright baroque strings orchestra
in those cases the main break-up resonance of Fostex FE206E (I am testing one) which is easily heard on pink noise as whistling goes beyond whistling into the ear-and-brain-piercing howling
simply unlistenable
the main resonance is somewhere between 2500-3000 Hz
1/10 octave notch filter centered at 2700 Hz is cutting out most of it
does anybody know how to fix it completely?
did anybody do any measurements of effects of cone doping in case of FE206E and similar drivers?
best,
graaf
graaf said:Fostex FE206E (I am testing one) which is easily heard on pink noise as whistling goes beyond whistling into the ear-and-brain-piercing howling
simply unlistenable
does anybody know how to fix it completely?
I consider the stock Fostex drivers as raw material -- at this point the only one i can listen to stock is the FE108e∑ (granted i haven't worked my way thru all of them yet)
Most people use the 8" Fostex as exemplers... but the smaller drivers (4-5") are their best drivers overall.
I am working on taming the big ones -- phase plugs are a start -- but i seriously doubt it is possible for them to ever have the midrange magic of the smaller ones.
dave
Hello Lynn,
Don't know if you'd have the time or interest, but just in case...
I will be displaying speakers at RMAF that use a waveguide-loaded compression driver and the TAD 10" alnico magnet woofer in what might be called a Geddes-esque configuration (no refractive foam, though). I'm among those who place radiation pattern very high up on the list of things to get right. The room number is 1100, in the tower. My room has a little alcove where you can take a breather and prop your leg up.
My brother-in-law, who lives outside of Nederlands, has a pair of Aerials in his living room to this day. That's a design that has certainly stood the test of time.
What's you present plan for crossover and/or equalization for your Beyond project?
Thanks,
Duke
Don't know if you'd have the time or interest, but just in case...
I will be displaying speakers at RMAF that use a waveguide-loaded compression driver and the TAD 10" alnico magnet woofer in what might be called a Geddes-esque configuration (no refractive foam, though). I'm among those who place radiation pattern very high up on the list of things to get right. The room number is 1100, in the tower. My room has a little alcove where you can take a breather and prop your leg up.
My brother-in-law, who lives outside of Nederlands, has a pair of Aerials in his living room to this day. That's a design that has certainly stood the test of time.
What's you present plan for crossover and/or equalization for your Beyond project?
Thanks,
Duke
audiokinesis said:
Many thanks for the invite, yes, I'll be there. I plan to do a lot of resting, and not carry any bags of literature this time. Stopping-off points with good music are always welcome.
Let's see, that makes Lowther, Bastanis, Emerald Physics and now Audio Kinesis. John Atwood (of Vacuum Tube Valley and Artemis phono preamp fame) will be staying at my place about 25 miles north of the show venue, so you may see the two of us showing up at the same time - or not, depending on what he plans to see.
I was reading the Dr. Geddes thread and noted the 20-element crossover. For those of you who are new at this, a 20-element crossover for a 2-way speaker is an extraordinarily sophisticated crossover, far, far beyond the abilities of anyone trying to "clone" a speaker.
The usual parts count is - let's see, 3 elements for the woofer lowpass, 2 for the inductance corrector, maybe another 3 for a midrange notch filter. A typical highpass would have 3 parts for the filter, 2 for a shelf filter, 2 for inductance compensation, maybe another 3 for a notch filter somewhere. That's still only 18 parts total - well OK, add a couple of resistors here and there for phase-shaping, and that gets to 20.
In reality, I'm sure the actual topology is a non-obvious computer-modelled network, designed, as Dr. Geddes has said, for uniform and well-corrected polar response over the intended coverage angle. I think the folks who want to "clone" the Summa or ESP15 are going to find it very difficult, even with the same drivers and HF waveguide.
As for the new system in this thread, I plan to keep it as simple as I can get away with. The Ariel used the flattest drivers I could find, and the best-sounding version still has a minimum of 7 parts, 10 if you want to be nice to the amplifier. I don't think the new system will be any simpler - I haven't found any drivers as flat as the 5.5" Vifa and the original sticky-dome 9000-series Scan-Speak.
The RAAL ribbon should be simple enough to highpass filter, but I have no idea what the 8 or 12-inch midbass is going to require - that's the hard part, since that's where the phase-shaping that controls the polar pattern is going to be happening, and the requirements for smooth out-of-band performance are most severe.
(Note: this changes if a compression driver + waveguide tweeter is chosen - simplicity goes out the window then, and quite substantial in-band HF EQ, with associated complexity, is required. I think I saw a max figure of 16 dB of subtractive EQ in the other thread - that's a lot, folks.)
As mentioned before, the only true crossover is between the tweeter and midbass, and the starting point will be low-Q 2nd-order filters. It's entirely possible neither the RAAL nor the midbass will tolerate filters with such a gentle shape, so at that point I'll be evaluating sharper crossovers vs alternative drivers. Dunno 'til I get there. I don't like the sound of high-Q crossovers, but the sound of driver coloration, and poorly controlled phase relationships between drivers, is even worse. So there's some tradeoffs here.
The bass-augmentation drivers will have their own amplification, active lowpass filtering, and room/driver equalization. The starting point for the topology are 1st-order lowpass filters chosen to complement the dipole/quasi-cardioid rolloff, followed by steeper filters starting near 1 kHz to remove the augmentation drivers from the midrange crossover region (and keep the vertical polar pattern from getting too complicated). All lowpass functions will be low-Q non-overshooting filters, Bessel family et al.
I have no intention of "forcing" the dipole/quasi-cardioid augmentation drivers to act as subwoofers, and will have a separate monopole subwoofer. I have an existing REL Strata II that can be pressed into service to evaluate the overall balance of the system.