The foam is rough enough that friction will stop slices of foam from slipping past each other. But if the waveguide surface is smooth then the foam will slide past it.
I've had 1" slices of foam laid against each other in horns for 2 years and they haven't budged relative to each other.
This sort of arrangement isn't suitable for commercial use and the pieces might dislodge when the speakers are moved.
I've had 1" slices of foam laid against each other in horns for 2 years and they haven't budged relative to each other.
This sort of arrangement isn't suitable for commercial use and the pieces might dislodge when the speakers are moved.
In an earlier post you misunderstood my meaning.
I did not say that the amount of higher modes was the difference between the area of a flat piston and a curved one.
What I did say was that the amount of scattering involved in the plane wave compression driver case was proportional to the difference in area between a plane circular piston and a spherical cap that represents a wavefront perpendicular to the duct wall.
If instead of a flat piston we now look at the area of the wavefront produced by a domed piston, the difference between it and the and our spherical cap that is perpendicular to the wall, can be shown to be considerably less.
This is consistent with your observation regarding the scalar products of the two surfaces, i.e. the velocity vectors are much more likely to interact in such a way as to produce a surface that is a close approximation to radially symmetrical equiphase surface, this is because there is close to a mapping of equal orthogonal vectors from the dome surface to the wavefront, the dome surface in turn being a close approximation to the surface produced by a set of equal length vectors emanating from the cone apex.
The flat to spherical cap transition does not have this feature and in fact needs to be converted into an approximation of a radially symmetrical potential field totally by scattering.
It then seems obvious that since the dome tweeter produces what we might call a psudo potential field in the first place, it is in fact inherently closer to an exact one, and less diffraction is involved in the process.
This is the last "hand waving" post I will make on this subject.
I will have to look up stuff that I haven't done for many years to put these observations more rigorous mathematical basis, but I do intend to when I get around to it.
rcw
I did not say that the amount of higher modes was the difference between the area of a flat piston and a curved one.
What I did say was that the amount of scattering involved in the plane wave compression driver case was proportional to the difference in area between a plane circular piston and a spherical cap that represents a wavefront perpendicular to the duct wall.
If instead of a flat piston we now look at the area of the wavefront produced by a domed piston, the difference between it and the and our spherical cap that is perpendicular to the wall, can be shown to be considerably less.
This is consistent with your observation regarding the scalar products of the two surfaces, i.e. the velocity vectors are much more likely to interact in such a way as to produce a surface that is a close approximation to radially symmetrical equiphase surface, this is because there is close to a mapping of equal orthogonal vectors from the dome surface to the wavefront, the dome surface in turn being a close approximation to the surface produced by a set of equal length vectors emanating from the cone apex.
The flat to spherical cap transition does not have this feature and in fact needs to be converted into an approximation of a radially symmetrical potential field totally by scattering.
It then seems obvious that since the dome tweeter produces what we might call a psudo potential field in the first place, it is in fact inherently closer to an exact one, and less diffraction is involved in the process.
This is the last "hand waving" post I will make on this subject.
I will have to look up stuff that I haven't done for many years to put these observations more rigorous mathematical basis, but I do intend to when I get around to it.
rcw
Please leave that to another thread
Significant? Hardly. It is a different issue altogether as well with regard to the unsubstantiated claims made. Even so, whatever impact it would have on a horn or waveguide would be nothing more than the introduction of additional diffraction artifacts as it would interfere with the intended change in profile required to minimize HOMs as I read it here, not a positive attribute. Having an impact is not equivalent to being an improvement.
Let's not drag that irrelevant diversion into this thread.
Dave
soongsc said:
If this is true, the EnABL patterns should show significant measurable effect due to how air moves along changing surfact profiles, which so many people have tried to prove the contrary in another thread.
Significant? Hardly. It is a different issue altogether as well with regard to the unsubstantiated claims made. Even so, whatever impact it would have on a horn or waveguide would be nothing more than the introduction of additional diffraction artifacts as it would interfere with the intended change in profile required to minimize HOMs as I read it here, not a positive attribute. Having an impact is not equivalent to being an improvement.
Let's not drag that irrelevant diversion into this thread.
Dave
rcw said:
I will be very interested in your results once you cease using the ill defined term "scattering" and terms like "obvious" when there is nothing of the sort and apply mathematics which can actually be understood.
What (I think) you may be trying to say is that in a conical waveguide with an equally subtended dome the HOM are created at the surface and their content does not change with propagation - this is true. Contrast that with the flat piston on an OS waveguide which starts out with zero HOM, but they are created as the wave propagates, which is also true. But in no way does this imply that the HOMs that reach the mouth will be greater or less than one another in either case. You HAVE TO do the math to show this one way or the other. I did that math, as I said, and it is published in my book and as an appendix on my web site.
BUT, there are two practical limitations to your point. First the dome has to subtend the same angle as the cone and this requires a special dome shape for every cone angle, and such domes generally do not exist, so real world implimentation is limited to impractical. And this technique can only produce an axial symmetric pattern. However, the compression driver on an OS waveguide has neither of these limitations and very functional real world devices are readily available.
I believe that your use of the term "scattering" comes from the fact that the HOM problem can be formulated as a sort of S-Matrix or Scattering Matrix formulation (see Chapter 6 of Audio Transducers). BUT, I would like to point out that this is an error since the S-matrix is formally a quantum mechanical technique and its analogy with the waveguide mode problem is incomplete. Thus, it is ill advised to use this formulation as it is only likely to cause confusion, as it does here. No where does anyone use the S-matrix analogy for the duct mode problem as you have been trying to do. (Except the Benade paper and the other one, but neither of those papers was looking at the waveshape problem but the "tuning" problem, i.e. reproduction waveguides versus musical horns.)
No. I actually wish people can point out anything I might have missed so I can really learn. At least that's what we have learned to do while working on projects with companies in the US. Now that I look back and see how discussions evolve here, it's interesting to see that deffensive attitude that was originally thought to be unique to Chinese culture is really not that way at all. I'm still learning.
soongsc said:
Based on my experince in Asia, you are distinctly non-asian. You take a position and dig in your heals. You often make statements that are unfounded, but claimed as fact simply because they are what you believe. I enjoy answering questions, but sometimes your posts are hard for me to respond to because you don't seem to do your homework and find out the facts. If you are so interested in HOMs then read what has been written about them. Most of what you say about them is incorrect, but correcting you just leads to other misunderstandings and we don't seem to get anywhere.
RCW
I would also like to point out another difference in the dome/conic horn and the OS waveguide. In the OS waveguide the HOM contributions will be frequency dependent (as well as the radiation via the modal impedance). In the dome/conic situation the modal contribution will be independent of frequency, although the propagation of these modes will be frequency dependent depending on the modal impedance. Thus, it is more than possible, even likely, that the two forms will be equal at some frequency with one better above and the other better below. This would be a critical aspect since the one which had the lower HOM at the higher frequencies would clearly be my preference, and I am guess that would be the OS waveguide. Anyone who has heard my systems can attest to the extremely clean HF response that they have, something which no other tweeter has ever exhibited.
I would also like to point out another difference in the dome/conic horn and the OS waveguide. In the OS waveguide the HOM contributions will be frequency dependent (as well as the radiation via the modal impedance). In the dome/conic situation the modal contribution will be independent of frequency, although the propagation of these modes will be frequency dependent depending on the modal impedance. Thus, it is more than possible, even likely, that the two forms will be equal at some frequency with one better above and the other better below. This would be a critical aspect since the one which had the lower HOM at the higher frequencies would clearly be my preference, and I am guess that would be the OS waveguide. Anyone who has heard my systems can attest to the extremely clean HF response that they have, something which no other tweeter has ever exhibited.
I guess the only way to confirm whether I'm asian or not is to meet face-to-face some time.gedlee said:
My basic point of view is that the source of HOMs can be from driver (diaphragm or other parts), horn/guide structural modes, and waves from impedance mismatching at the ends of the horn/guide. You are saying that this is unfounded?
soongsc said:
Well I would guess that you are Asian, just not typical Asian. I have nothing against Asians, of course, my son is 1/2 Asian and my Chinese Mother-in-Law was the nicest woman that I have ever met.
To answer your question, basically, you are almost completely wrong. The driver can cause HOMs, but this is not the dominate cause, structural modes are not even a factor (except in perhaps very weak waveguides and even then they would not be significant) and the end conditions of the device is a small effect. The HOM are created within the waveguide itself, a factor that you don't seem to recognize, and which is, by far, the greatest factor. Beyond this you need to read the literature until you have a grasp of the concepts since what you keep replying is incorrect.
I think we are getting somewhere with the communication.gedlee said:
Well, my wife is known as the toughes female manager in her company. It was interesting to see interaction with a female manager from the US of similar character.What you are saying about the driver diaphram role in HOM generation is small as well? I can agree that phase plug of most compression drivers may have less significant effects. But certainly not the diaphragm modes.
I think you can agree that there are lots of quite weak wave guides out there. If I remember correctly, you mentioned that the wider angle the guide/ the more significant the HOMs? This is in line with the fact that wider guides get structurally excited easier, which results in concentric modes in a similar nature as cone breakup modes of direct radiating drivers. How can you say that the HOMs generated from such are less significant since the vibrating surface is quite large?
Asymmetrical verses axisymmetrical horns
I'd like to see the issue of vertical directivity explored in more detail.
I wasn't going to mention this in Earl's thread, but when the subject was brought up, I thought I'd weigh in. I'd like to see this subject get some attention rather than being dismissed or ignored.
Earl, you may remember at the Lone Star Audiofest in 2004, I tallked with you about making asymmetrical horns. I was in the planning stage of a project to make wood horns on a CNC machine, and since your horns were CD, they were an attractive option. I believe you proposed a PS shape for asymmetrical coverage.
I've been making DI-matched speakers for a long, long time. To me, that means matching the horizontal angles but it also means keeping the vertical angles within the nulls. All speakers with vertically stacked drivers have nulls, so it makes sense to me to limit the acoustic energy to an angle inside the nulls as much as possible.
Uniform horizontal coverage is important, but walls aren't the only reflective boundaries. In fact, the ceiling is probably the most reflective boundary.
Spectral balance off-axis in the horizontal isn't trivial, but it is certainly much easier to do than getting spectral balance off-axis vertically. To me, that's the reason to pay attention to the horn's vertical angle and to limit it to the null angle as much as possible.
That's why I was puzzled to see an axisymmetrical horn used in your Summa, a speaker with uniform directivity in the horizontal plane. I remember asking you why you used a round horn, and as I recall, you basically said you liked asymmetrical flares best, or something to that effect.
I was used to seeing round horns in loudspeakers where uniform directivity wasn't a design goal. Tractrix "salad bowl" horns and others like them became pretty popular around 2000, seems like. Proponents of horns like that usually rely heavily on room treatments, because they really an need anechoic environment to sound right. But speakers with uniform directivity generate a reverberent field with the right tonal balance.
The illustrations I made earlier are pretty good at showing what happens to vertical directivity with various CD horns. The one that desribes a 90 degree round horn looks almost identical to the vertical DI plot you show in your Summa paper. Except, of course, your chart shows DI and mine is angular coverage, so the charts are reversed - mine goes up where yours goes down and vice versa.
I've also posted measurements of actual loudspeakers. They are raw datasets with no smoothing applied and they show on-axis as well as off-axis in both the horizontal and vertical planes. The measurements were done outdoors on an LMS system swept from 20-20kHz.
I have found charts of your speakers that show response from 200Hz up, smoothed to 1/3 octave resolution. Actually, now that I think about it, I've only seen data on one speaker, the Summa. The charts show horizontal off-axis response, buit I don't think I've seen charts of vertical off-axis response.
I don't want to make this a loaded question, because we both know the polars aren't kind when you get close to the null angles. Your curves would look good up to about +/-20 degrees, and then the nulls would cause response to drop around 1kHz. Since your tweeter horn has tall vertical coverage, response will rise above 1.2kHz or so and remain high all the way up into the top octave.
The thing is, you don't have just one flare to work with, you have a family of them. So you don't have to see this as a matter of "yours" verses "theirs". It can just as easily be a comparison of one of your horns against another one of your horns.
The horn you proposed to me in 2004 had an oval mouth to provide an asymmetrical coverage angle. This would allow it to be closer to the woofer, widening the null angle. It also would reduce vertical energy, ideally limiting most of it to angles smaller than the null angle. To me, that's attractive because the horn's vertical pattern matches the null angle from baffle spacing. When the vertical angle drops in the crossover region, it stays low.
The only reason I didn't choose your horn, by the way, is that I'm a DIY kind of guy. I made a horn that works very well for me. It's like the H290 I use in many of my speakers, but larger so it has pattern control at lower frequency. The mouth exit is radiused too, like having a CD flare with a tractrix mouth. But I expect your asymmetrical flare would have worked very well too.
So that brings me back to the same question I asked you in 2004. You have both asymmetrical and axisymmetrical horns in your box of chocolates. Both are flavors that fit your mathematical palate. Both use a spheroid coordinate system for radiusing the throat angle to the wall angle. So why choose the axisymmetric one, when you know the nulls will cut its vertical pattern in half?
I'd like to see the issue of vertical directivity explored in more detail.
I wasn't going to mention this in Earl's thread, but when the subject was brought up, I thought I'd weigh in. I'd like to see this subject get some attention rather than being dismissed or ignored.
Earl, you may remember at the Lone Star Audiofest in 2004, I tallked with you about making asymmetrical horns. I was in the planning stage of a project to make wood horns on a CNC machine, and since your horns were CD, they were an attractive option. I believe you proposed a PS shape for asymmetrical coverage.
I've been making DI-matched speakers for a long, long time. To me, that means matching the horizontal angles but it also means keeping the vertical angles within the nulls. All speakers with vertically stacked drivers have nulls, so it makes sense to me to limit the acoustic energy to an angle inside the nulls as much as possible.
Uniform horizontal coverage is important, but walls aren't the only reflective boundaries. In fact, the ceiling is probably the most reflective boundary.
Spectral balance off-axis in the horizontal isn't trivial, but it is certainly much easier to do than getting spectral balance off-axis vertically. To me, that's the reason to pay attention to the horn's vertical angle and to limit it to the null angle as much as possible.
That's why I was puzzled to see an axisymmetrical horn used in your Summa, a speaker with uniform directivity in the horizontal plane. I remember asking you why you used a round horn, and as I recall, you basically said you liked asymmetrical flares best, or something to that effect.
I was used to seeing round horns in loudspeakers where uniform directivity wasn't a design goal. Tractrix "salad bowl" horns and others like them became pretty popular around 2000, seems like. Proponents of horns like that usually rely heavily on room treatments, because they really an need anechoic environment to sound right. But speakers with uniform directivity generate a reverberent field with the right tonal balance.
The illustrations I made earlier are pretty good at showing what happens to vertical directivity with various CD horns. The one that desribes a 90 degree round horn looks almost identical to the vertical DI plot you show in your Summa paper. Except, of course, your chart shows DI and mine is angular coverage, so the charts are reversed - mine goes up where yours goes down and vice versa.
I've also posted measurements of actual loudspeakers. They are raw datasets with no smoothing applied and they show on-axis as well as off-axis in both the horizontal and vertical planes. The measurements were done outdoors on an LMS system swept from 20-20kHz.
I have found charts of your speakers that show response from 200Hz up, smoothed to 1/3 octave resolution. Actually, now that I think about it, I've only seen data on one speaker, the Summa. The charts show horizontal off-axis response, buit I don't think I've seen charts of vertical off-axis response.
I don't want to make this a loaded question, because we both know the polars aren't kind when you get close to the null angles. Your curves would look good up to about +/-20 degrees, and then the nulls would cause response to drop around 1kHz. Since your tweeter horn has tall vertical coverage, response will rise above 1.2kHz or so and remain high all the way up into the top octave.
The thing is, you don't have just one flare to work with, you have a family of them. So you don't have to see this as a matter of "yours" verses "theirs". It can just as easily be a comparison of one of your horns against another one of your horns.
The horn you proposed to me in 2004 had an oval mouth to provide an asymmetrical coverage angle. This would allow it to be closer to the woofer, widening the null angle. It also would reduce vertical energy, ideally limiting most of it to angles smaller than the null angle. To me, that's attractive because the horn's vertical pattern matches the null angle from baffle spacing. When the vertical angle drops in the crossover region, it stays low.
The only reason I didn't choose your horn, by the way, is that I'm a DIY kind of guy. I made a horn that works very well for me. It's like the H290 I use in many of my speakers, but larger so it has pattern control at lower frequency. The mouth exit is radiused too, like having a CD flare with a tractrix mouth. But I expect your asymmetrical flare would have worked very well too.
So that brings me back to the same question I asked you in 2004. You have both asymmetrical and axisymmetrical horns in your box of chocolates. Both are flavors that fit your mathematical palate. Both use a spheroid coordinate system for radiusing the throat angle to the wall angle. So why choose the axisymmetric one, when you know the nulls will cut its vertical pattern in half?