Probably 
- I saw it then and I'll show you on any OS coordinates you pick...
But I agree that the actual wavefront development is all but intuitive. We see that in the bends of the channels. Which is what led me to the belief that it's possible to do the other things, provided the channels are kept narrow enough.
- I saw it then and I'll show you on any OS coordinates you pick...
But I agree that the actual wavefront development is all but intuitive. We see that in the bends of the channels. Which is what led me to the belief that it's possible to do the other things, provided the channels are kept narrow enough.
Last edited:
For the creation of a spherical wavefront? The patent by Danley for the layered combiner used in the J1-94 uses paths of different length to shape the exit waveform to spherical before loading the horn throat, so I took that to be a given.
Mabat,
I can now see the issue.
My first thought was that all points to the wavefront must have the exact same travel time and since the speed of sound is constant, these path lengths to the origin must all be equal. That's always going to be true.
The error in my thoughts was that I had to think of the origin now as an extended line, in OS it's not a point. This means that the average of the distances (the integral) from every point on the origin to the wavefront must all be equal. Thus sectioning off part of the origin while the above has to be true for any section, it may no longer be true that each section is the same length - given the new origin dimensions. This would explain why the vanes didn't work as I suspected. It also means that what you are suggesting now makes a lot more sense to me.
I can now see the issue.
My first thought was that all points to the wavefront must have the exact same travel time and since the speed of sound is constant, these path lengths to the origin must all be equal. That's always going to be true.
The error in my thoughts was that I had to think of the origin now as an extended line, in OS it's not a point. This means that the average of the distances (the integral) from every point on the origin to the wavefront must all be equal. Thus sectioning off part of the origin while the above has to be true for any section, it may no longer be true that each section is the same length - given the new origin dimensions. This would explain why the vanes didn't work as I suspected. It also means that what you are suggesting now makes a lot more sense to me.
A fun side project is to make a poor mans Celestion driver with this little 2 inch B&C neodymium driver
B&C 2NDF26 2 inch Neodymium Loudspeaker Driver - 30W RMS, 8 Ohm
The problem is that it is both concave and has a rubber surround.
However it could be fun to make a phase plug for it!
B&C 2NDF26 2 inch Neodymium Loudspeaker Driver - 30W RMS, 8 Ohm
The problem is that it is both concave and has a rubber surround.
However it could be fun to make a phase plug for it!
There's a mention of this sort of thing in 'the horn book', with a note that it needs more investigation. It seems that adding layers of this sort of porous mesh at the throat has the effect of making the horn itself appear quite a bit 'longer' acoustically.
I'll be interested to see what results you get. If you have time and materials, perhaps make a few with varying hole dimensions and arraying, then layer them up while documenting the difference for each?
It's not the kind of thing that is easy to model without some serious CPU horsepower for the ultra fine mesh.
Yes, I did some measurements - more than a decade ago. I am not sure that I could dig them up now.
The "take-away" for me was that FR wise the effect was small, a few dB loss at 10 kHz. What stuck with me was that this loss, as viewed in the impulse response, was well after the peak of the impulse, i.e. the "tail".
This foam was deep, about 10", 30-40 PPI, hence very open. Put down in the throat this foam will have much more effect owing to the much higher velocities. Meyer had a patent on something like this.
The "take-away" for me was that FR wise the effect was small, a few dB loss at 10 kHz. What stuck with me was that this loss, as viewed in the impulse response, was well after the peak of the impulse, i.e. the "tail".
This foam was deep, about 10", 30-40 PPI, hence very open. Put down in the throat this foam will have much more effect owing to the much higher velocities. Meyer had a patent on something like this.
That's an interesting point, I wouldn't have guessed that, i.e. that the velocities are higher at the throat.gedlee said:
(That means that the pressure gradient is higher, right. I'm only thinking how to get this information out of the simulated data.)
- Now when I'm thinking about it, it seems so obvious that I don't know why I found it non-intuitive at first. Probably because I read so many times that a horn "transforms high pressure and low velocity into low pressure and high velocity"...
Last edited:
Fascinating... I quote:
" New Results on Perception
In 2003, Geddes, Lee and Magalotti reported on a subjective test that they had performed with compression drivers.
In this test three drivers were tested for the perception of linear and nonlinear distortions. 26 listeners participated in this double blind test.
It was found that none of these listeners could reliably perceive the nonlinear distortions even when these distortions were upwards of 20% THD."
Is there a description on exactly how this DBT was performed? Room, used system, stimuli (music, tones?), duration, repetition etc?
//
" New Results on Perception
In 2003, Geddes, Lee and Magalotti reported on a subjective test that they had performed with compression drivers.
In this test three drivers were tested for the perception of linear and nonlinear distortions. 26 listeners participated in this double blind test.
It was found that none of these listeners could reliably perceive the nonlinear distortions even when these distortions were upwards of 20% THD."
Is there a description on exactly how this DBT was performed? Room, used system, stimuli (music, tones?), duration, repetition etc?
//