WHG,
That simple 2D software should clue you into how simple the curvature of the Oblate spheroid truly is if it is only looking at a plan view of the waveguide in a simple cross sectional view. That is nothing like using Solidworks or even AutoCad software doing three dimensional design. Since these are really simply rotated about an axis and not asymmetrically varying multi-axis designs that is why I think you are making to much of the Bezier curvature. The simple drawing that Allen just put up shows how simple this can be. What I have done has been asymmetrical designs using elliptical cross sections rotated about an x,y axis and expanding in a spherical manner and not simple planar concepts. I think you can understand the difference with the concepts. The LeClech horns though different are much more similar to what I was doing. Very different than what Earl has proposed. I still fail to see why a round horn waveguide is thought to be that difficult?
That simple 2D software should clue you into how simple the curvature of the Oblate spheroid truly is if it is only looking at a plan view of the waveguide in a simple cross sectional view. That is nothing like using Solidworks or even AutoCad software doing three dimensional design. Since these are really simply rotated about an axis and not asymmetrically varying multi-axis designs that is why I think you are making to much of the Bezier curvature. The simple drawing that Allen just put up shows how simple this can be. What I have done has been asymmetrical designs using elliptical cross sections rotated about an x,y axis and expanding in a spherical manner and not simple planar concepts. I think you can understand the difference with the concepts. The LeClech horns though different are much more similar to what I was doing. Very different than what Earl has proposed. I still fail to see why a round horn waveguide is thought to be that difficult?
There is much talk about the smoothness of the transition from the throat to the waveguide.
How smooth does that transistion need to be? What are acceptable tolerances?
Gap between end of throat and start of waveguide?
Difference in angle at throat to waveguide junction?
Radial offset between throat and waveguide?
Surface roughness of the waveguide at or near the throat?
Surface roughness at some greater distance from the throat?
Can anyone put numbers to these? Is anyone prepared to put up numbers?
either in mm, or in %, or some other guideline numbers.
How smooth does that transistion need to be? What are acceptable tolerances?
Gap between end of throat and start of waveguide?
Difference in angle at throat to waveguide junction?
Radial offset between throat and waveguide?
Surface roughness of the waveguide at or near the throat?
Surface roughness at some greater distance from the throat?
Can anyone put numbers to these? Is anyone prepared to put up numbers?
either in mm, or in %, or some other guideline numbers.
Last edited:
So!? What's your point! (Or is this just another attempt at belittling someone.)
I have a full 3D package too, but SmartSketch is the easiest to use of any of the "sketching" packages.
1) I have measured a 1 mm gap to be disruptive
2) a few degrees here would likely be an issue, but this would be hard for me to test
3) again an offset of say 1 mm will be seen in the data
4,5) I have not tested this, but I would say that it would need to be extremely rough to be a factor, the greater the distance out the lesser its effect is likely to be.
In my builds I use a very small amount of filler to smooth any discontinuities at the driver waveguide junction. The finger is very sensitive to any discontinuities - much better than 1 mm. Hence if it feels smooth then it is smooth. Initially it almost never is, after smoothing it is.
Hope this helps.
I agree on the finger test.
it might be considered subjective, but it is extremely sensitive. Maybe too senstive.
I badly lacerated the end of a finger by accidentally sticking it into a bench saw!
It was very neatly pieced together by a wonderfully skilled nurse.
It remained heavily bandaged for 6 weeks. The school kids all laughed at my sticky.
After they took off the bandages, I could feel my fingerprints on the other hand. The effect only lasted a week or so.
it might be considered subjective, but it is extremely sensitive. Maybe too senstive.
I badly lacerated the end of a finger by accidentally sticking it into a bench saw!
It was very neatly pieced together by a wonderfully skilled nurse.
It remained heavily bandaged for 6 weeks. The school kids all laughed at my sticky.
After they took off the bandages, I could feel my fingerprints on the other hand. The effect only lasted a week or so.
Simply No Surprise
I was expecting to find a 3D tool that will get me to a model of a prototype horn that can be reproduced by a 3D printer or NC machining center.
As for the ad homonym, why am I not surprised to find it here? WHG
I was expecting to find a 3D tool that will get me to a model of a prototype horn that can be reproduced by a 3D printer or NC machining center.
As for the ad homonym, why am I not surprised to find it here? WHG
And why am I not surprised at your inability to recognize condescension on your part.
Pot vs. Kettle
OS is the assumed shape of waves propagating in an infinite hyperbolic horn, only when the wavelength of such waves is much larger than horn cross-section dimensions. This issue was addressed by Freehafer and Salmon in their PHD dissertations at MIT in the late 1930's. What I was addressing in my post here was a methodology to terminate such horns. The firestorm of undue criticism that has followed is emblematic a chronic form of the "Not Invented Here" syndrome. Even the ethnicity of the paper authors was brought into question. That gem was then followed by an allegation of condescension by others. And oh sure, the pot calls the kettle black yet again.
In regards to the complexity of the matter, that only occurs when we attempt to describe how an acoustic wave behaves inside the confines of a horn body. Where that notion gets very interesting is when such waves are carrying a music signal. WHG
OS is the assumed shape of waves propagating in an infinite hyperbolic horn, only when the wavelength of such waves is much larger than horn cross-section dimensions. This issue was addressed by Freehafer and Salmon in their PHD dissertations at MIT in the late 1930's. What I was addressing in my post here was a methodology to terminate such horns. The firestorm of undue criticism that has followed is emblematic a chronic form of the "Not Invented Here" syndrome. Even the ethnicity of the paper authors was brought into question. That gem was then followed by an allegation of condescension by others. And oh sure, the pot calls the kettle black yet again.
In regards to the complexity of the matter, that only occurs when we attempt to describe how an acoustic wave behaves inside the confines of a horn body. Where that notion gets very interesting is when such waves are carrying a music signal. WHG
Just to be clear it was not the ethnicity that was questioned but the nationality. Chinese, as a race are very intelligent people, to wit the work being done in Hong Kong and Taiwan, but the education system in the PRC is in shambles, having been wreaked some decades ago during the Culteral evolution, and struggling even now to regain some measure of competence.
My wife is Chinese (with a PhD), hence my son 1/2 Chinese, I have more than 1/2 dozen Chinese nieces and nephews, one even a Prof at Harvard, one step brother, a PhD graduate of MIT, who is now provost of HKTU and another step brother PhD who is head of Linguistics at HKU. These are not dumb people, but not one of them was educated in the PRC.
Concurrance
OK Earl, I can agree to that distinction. My wife, while not Chinese, is from a Pacific Rim country, is also well degreed and possesses an intellect to match as well. No argument from me here on that issue. Bill
OK Earl, I can agree to that distinction. My wife, while not Chinese, is from a Pacific Rim country, is also well degreed and possesses an intellect to match as well. No argument from me here on that issue. Bill
WHG,
Termination is something that was not a consideration in horns for so long, it has been shown by Earl and others to be much more important than was earlier a consideration. Horns just ended at the 1/4 wavelength dimension and that was that. I can't honestly answer if there is an ideal termination radius, I would think it is frequency dependent on horn mouth size and required lowest frequency output. If you have access to a 3d cad system then I don't think you should have any real problems with doing the design no matter which expansion rate or horn shape you so desire. I guess I think more in a 3D fashion of wave shape and volume when I think about a waveguide, not as a simple point cloud or planar wavefront. As far as the joint between the driver and the end of the horn that should be air tight, meaning you would typically use some form of gasket there, not just a driver mounted raw onto the back of the horn. It could be as simple as a thick paper gasket or some form of rubber, whatever can give you a good air tight seal and also not cause a discontinuity at the joint. A simple square cross-sectional O-ring could work but would need a recess into the back or the horn to work and getting one that matched the angle of the exit of the driver would be an issue as you would need to mold your own to have an angle on the ID of the O-ring to match the driver. Even with a large radius at the end of a horn you have to consider that as I said the typical horn is only 1/4 wavelength so are your still going to end up with an omnidirectional wave response at the lowest usable frequency, so is there really a perfect solution, I don't believe so, it is still a compromise but much better than leaving the old flat front to the horn and having a sharp diffraction at that edge.
On the Chinese side of things I have to say I work with many Chinese and to say they are uneducated is simple not so these days. There are plenty of Phd engineers and others in China, perhaps many are educated in other countries but plenty of them do return there to live and work. You just have to be careful with what you are willing to disclose, there is a real problem with stolen intellectual property, that is just something that may take another generation or more to stop when it start to really affect their own industries from within. I'll stay out of the political side of that mess.
Termination is something that was not a consideration in horns for so long, it has been shown by Earl and others to be much more important than was earlier a consideration. Horns just ended at the 1/4 wavelength dimension and that was that. I can't honestly answer if there is an ideal termination radius, I would think it is frequency dependent on horn mouth size and required lowest frequency output. If you have access to a 3d cad system then I don't think you should have any real problems with doing the design no matter which expansion rate or horn shape you so desire. I guess I think more in a 3D fashion of wave shape and volume when I think about a waveguide, not as a simple point cloud or planar wavefront. As far as the joint between the driver and the end of the horn that should be air tight, meaning you would typically use some form of gasket there, not just a driver mounted raw onto the back of the horn. It could be as simple as a thick paper gasket or some form of rubber, whatever can give you a good air tight seal and also not cause a discontinuity at the joint. A simple square cross-sectional O-ring could work but would need a recess into the back or the horn to work and getting one that matched the angle of the exit of the driver would be an issue as you would need to mold your own to have an angle on the ID of the O-ring to match the driver. Even with a large radius at the end of a horn you have to consider that as I said the typical horn is only 1/4 wavelength so are your still going to end up with an omnidirectional wave response at the lowest usable frequency, so is there really a perfect solution, I don't believe so, it is still a compromise but much better than leaving the old flat front to the horn and having a sharp diffraction at that edge.
On the Chinese side of things I have to say I work with many Chinese and to say they are uneducated is simple not so these days. There are plenty of Phd engineers and others in China, perhaps many are educated in other countries but plenty of them do return there to live and work. You just have to be careful with what you are willing to disclose, there is a real problem with stolen intellectual property, that is just something that may take another generation or more to stop when it start to really affect their own industries from within. I'll stay out of the political side of that mess.
20 yrs ago, (and more), it was "Made in Japan." And before that it was "Made in Germany, Italy, etc."
This is purely a matter of development and economics and psychology. Especially psychology: We are an inventive species, but before we invent, we have to imitate. The more complex and abstract the subject, (or the higher the level of performance, as in music), the more this is true.
The situation outlined in 1st para above is the result of the great upheaval of WW2. The oriental societies were already isolated and then badly disrupted by the war but did have the comparative advantage of low wages to produce imitative products.
Going beyond that gets me into the perilous area of politics and too OT for this thread.
Except to say the future will be claimed by societies with best educational systems - whatever that means.
20 yrs ago, (and more), it was "Made in Japan." And before that it was "Made in Germany, Italy, etc."
This is purely a matter of development and economics and psychology. Especially psychology: We are an inventive species, but before we invent, we have to imitate. The more complex and abstract the subject, (or the higher the level of performance, as in music), the more this is true.
The situation outlined in 1st para above is the result of the great upheaval of WW2. The oriental societies were already isolated and then badly disrupted by the war but did have the comparative advantage of low wages to produce imitative products.
Going beyond that gets me into the perilous area of politics and too OT for this thread.
Except to say the future will be claimed by societies with best educational systems - whatever that means.
Usually a wave does not "see" a disturbance with dimensions much smaller than the wave.
This seems not to apply to the throat of a horn/ waveguide, and I can think of other examples.
I would like to formulate this idea more strictly but I have never been able to express it formally.
Do you have any ideas or references on this?
It would be nice to have an objective metric for the level of precision required.
Best wishes
David
One More Time ...
Recommended uses of the C-Bezier Curve include:
1) Mouth Geometry
for me as stated earlier,
a) the mouth should be formed by a curve having a declining radius (of curvature) not a single radius of curvature.
b) at the tangent point of departure, the radii of curvature for horn body and mouth should match.
2) Throat geometry
for adapter design when an exiting driver and an existing horn need to be matched.
The China issue is not mine, I just responded when it was raised by others.
This is really not the forum to address the surrounding politics. WHG
WHG
Recommended uses of the C-Bezier Curve include:
1) Mouth Geometry
for me as stated earlier,
a) the mouth should be formed by a curve having a declining radius (of curvature) not a single radius of curvature.
b) at the tangent point of departure, the radii of curvature for horn body and mouth should match.
2) Throat geometry
for adapter design when an exiting driver and an existing horn need to be matched.
The China issue is not mine, I just responded when it was raised by others.
This is really not the forum to address the surrounding politics. WHG
WHG
Boundrary Layer Problems
… arise in acoustically small spaces where the layer occupies a comparatively large portion of the enclosed air volume. The thickness (Dv) of this layer may be expressed by the formula
Dv=((2*u)/(p0 *w))^(1/2)
where
u - viscosity of air
p0 - density of air
w = 2 * pi * f – angular frequency
f – frequency (Hz)
Surface imperfections of comparable or larger measure will give rise to a turbulent air flow which will disrupt normal wave propagation.
Dave, there is more to this than what is presented here, so I have attached an article that may be helpful.
Bill
… arise in acoustically small spaces where the layer occupies a comparatively large portion of the enclosed air volume. The thickness (Dv) of this layer may be expressed by the formula
Dv=((2*u)/(p0 *w))^(1/2)
where
u - viscosity of air
p0 - density of air
w = 2 * pi * f – angular frequency
f – frequency (Hz)
Surface imperfections of comparable or larger measure will give rise to a turbulent air flow which will disrupt normal wave propagation.
Dave, there is more to this than what is presented here, so I have attached an article that may be helpful.
Bill
"a) the mouth should be formed by a curve having a declining radius (of curvature) not a single radius of curvature.
b) at the tangent point of departure, the radii of curvature for horn body and mouth should match."
These requirements can be described by the simple mathematics for a hyperbolic curvature. Nothing that hasn't been done for many years.
b) at the tangent point of departure, the radii of curvature for horn body and mouth should match."
These requirements can be described by the simple mathematics for a hyperbolic curvature. Nothing that hasn't been done for many years.
A great paper, clearly showing how the phase plug is not properly designed. These problems were noted in my patents. Also, the assumed plane wave boundary condition at the exit of the phase plug is obviously wrong. A cross mode is readily apparent. A better model would consider the tube to have plane wave modes and cross modes, each of which have very different impedances. I think that the match to reality would be much better if this had been done as the wave front at the end of a phase plug is never really flat - again clearly evident in the data. Personally I think that the inclusion of the visco-thermal aspects is a lesser consideration than the consideration of HOMs at the interface.