Jmmlc said:
Hi ...
But the point is that impedance IS NOT frequency response. While they may be related you cannot say that they are equivalent.
Figure 5 is interesting - thanks - as this is exactly what I am looking for. I would have expected a lower resonance frequency, but I can see how the rigid canal termination would double the resonance. Its usually at about 3 kHz. Then one might expext a correction for a closed ear insert phone to be a dip at about 7 kHz and a peak at about 3 kHz. The peak is already in the transducer, but not the 7 kHz. Very interesting.
Earl,
Who said that frequency response and impedance are the same?
Surely not me.
Best regards from Paris, France
Jean-Michel Le Cléac'h
Who said that frequency response and impedance are the same?
Surely not me.
Best regards from Paris, France
Jean-Michel Le Cléac'h
Figure 5 is very interesting and one would think that earphones have this built in. Perhaps not. And if it is different enough from ear to ear, it may explain why some folks like certain 'phones, other not.
(I once bought a pair of Grado - sent them back presto!)
We are gettting OT, tho, aren't we?
(I once bought a pair of Grado - sent them back presto!)
We are gettting OT, tho, aren't we?
I just started to listen to this new waveguide. It does reveal some problems which sound like it's related the mismatch of impedance with the amplifier, so I'm going to plug it some more components just to flatten the impedance. No measurements yet, but it does sound like there is good potential. 45 deg angle toe-in seems to be the right angle so that amplitude does not change. There is no back curl or baffle that merges with the waveguide lip, but the diffraction effect is VERY noticeable. I think this will show up in the measurements as well.mige0 said:
Reading the Newell and Holland book, it was pointed out that a good class A amplifier is necessary to keep crossover distortion to minimum, I can now understand how important that is having listening to this new waveguide, it's just so sensitive in revealing little things.
Re: Re: Keele¡¦s Asymptotic Model
Dr. Geddes,
Thank you for answering my questions and sharing some history on Keele's work. I had no idea Keele used diffraction devices in his work. Indeed, it would seem that Keele's formulas are optimistic then. If I work backwards using your 15" 90 degree OS waveguide and the 2/3 angle at approximately 1KHz, the Kk constant would have to be somewhere around 52 X 10^3 instead of the suggested 29.707 X 10^3. Does this sound more realistic?
The only other question I would have is the mouth round over accounted for here? How much of the 15" diameter mouth on your waveguide is pure OS, and how much is the round over? At what diameter does the round over begin? If the round over is not accounted for then the Kk constant may come in at a lower numeric value.
Rgs, JLH
My original post on this subject--> http://www.diyaudio.com/forums/showthread.php?postid=1858334#post1858334
gedlee said:
Dr. Geddes,
Thank you for answering my questions and sharing some history on Keele's work. I had no idea Keele used diffraction devices in his work. Indeed, it would seem that Keele's formulas are optimistic then. If I work backwards using your 15" 90 degree OS waveguide and the 2/3 angle at approximately 1KHz, the Kk constant would have to be somewhere around 52 X 10^3 instead of the suggested 29.707 X 10^3. Does this sound more realistic?
The only other question I would have is the mouth round over accounted for here? How much of the 15" diameter mouth on your waveguide is pure OS, and how much is the round over? At what diameter does the round over begin? If the round over is not accounted for then the Kk constant may come in at a lower numeric value.
Rgs, JLH
My original post on this subject--> http://www.diyaudio.com/forums/showthread.php?postid=1858334#post1858334
I've never tried to put my work in terms of Keele's so I don't know about the numbers.
The radius of the 15" waveguide is about 4", but it is a "fillet" at the 15" point. Hence the true radius is more like 18" at the outside and about 14" at the inside. The best way to "model" this (see my book) is to use 15" as the mouth and then convolve the polar response with a gaussian smoother. This was shown in the book to be theoretically correct as well as yielding results that are accurate.
The radius of the 15" waveguide is about 4", but it is a "fillet" at the 15" point. Hence the true radius is more like 18" at the outside and about 14" at the inside. The best way to "model" this (see my book) is to use 15" as the mouth and then convolve the polar response with a gaussian smoother. This was shown in the book to be theoretically correct as well as yielding results that are accurate.
"How much of the 15" diameter mouth on your waveguide is pure OS"
I believe that OS refers only to the shape of the transition from throat to WG, which is conical (not OS as it is often called).
I believe that OS refers only to the shape of the transition from throat to WG, which is conical (not OS as it is often called).
noah katz said:
Hi Noah
That's not really true. The OS becomes conical at larger axial locations as all waveguides do. The equation is asymptotic to a straight line at a given angle - namely conical. Hence, basically, OS and conical are indistriguishable for large axial locations. Its true that they most differ at the throat, but my waveguides are OS all the way out to the mouth radius.
Re: Re: Re: Keele¡¦s Asymptotic Model
Sorry - this misunderstanding may have occured from how I have put it in my paper, refering to the part of the OS contour where the bending / transition of the wave front form plane to spherically happens "the most".
🙁
Have corrected that part.
🙂
Michael
noah katz said:
Sorry - this misunderstanding may have occured from how I have put it in my paper, refering to the part of the OS contour where the bending / transition of the wave front form plane to spherically happens "the most".
🙁
Have corrected that part.
🙂
Michael
" The OS becomes conical at larger axial locations as all waveguides do. The equation is asymptotic to a straight line at a given angle"
Thanks, Earl, I hadn't realized that.
Thanks, Earl, I hadn't realized that.
Re: Re: Re: Keele¡¦s Asymptotic Model
Before I started building waveguides I got a bit lost in the math. There's really not a lot to this, and there's no need to overanalzye the problem. It's basically three steps:
1 - pick your coverage angle
2 - calculate the OS curve on a spread sheet. I have one uploaded to Google documents, it's public
3 - Use the biggest roundover you can manage on the waveguide AND the cabinet
That's about all there is to it.
JLH said:
Before I started building waveguides I got a bit lost in the math. There's really not a lot to this, and there's no need to overanalzye the problem. It's basically three steps:
1 - pick your coverage angle
2 - calculate the OS curve on a spread sheet. I have one uploaded to Google documents, it's public
3 - Use the biggest roundover you can manage on the waveguide AND the cabinet
That's about all there is to it.
Seems to me that drivers exit angle is less significant with Geddes OS waveguide, due to the very short and rounded transistion🙄
tinitus said:
It doesn't matter what contour you are using, if there is a slope change then there will be reflection and diffraction. Not good.
Ok, to me the throath just seems very short and "rounded", with no angles
Obviously Im mistaken
But an OS waveguide doesnt look like free of slope changes
Im not critical, negative or anything
Its just how it looks to me
Maybe Im just gaining some small interest due to I need to find information about exit angle on a certain BEYMA driver
Obviously Im mistaken
But an OS waveguide doesnt look like free of slope changes
Im not critical, negative or anything
Its just how it looks to me
Maybe Im just gaining some small interest due to I need to find information about exit angle on a certain BEYMA driver
tinitus said:
Exactly the point, it has to change slope, but the OS does this in the most gradual way possible.
Maybe you have been discussing this over and over already
It seems to me like a phase plug inside the driver, changes exit angles considerably
In which case the side angles of the exit doesnt really tell much about actual exit angle
But if you use the flare rate exit angle caused by phase plug, it will cause abrupt change in slope rate of the exit troath angle
If manufactor gives information about exit angles, do they deal with this issue, or will such spec be totally unreliable
If above holds any truth, the short steep OS throath transistion begins to make sense
It seems to me like a phase plug inside the driver, changes exit angles considerably
In which case the side angles of the exit doesnt really tell much about actual exit angle
But if you use the flare rate exit angle caused by phase plug, it will cause abrupt change in slope rate of the exit troath angle
If manufactor gives information about exit angles, do they deal with this issue, or will such spec be totally unreliable
If above holds any truth, the short steep OS throath transistion begins to make sense
I think this is right, Tinitus.
For example for a certain Celestion CD-driver an optimum entry angle (of a horn/wg) is 25deg and this is not the driver's wall exit angle (which even is negative in this example), rather it continues the cross-section expansion rate -- which looks right to me as long as we are below the critical frequency when wavelenghts start to be important.
- Klaus
For example for a certain Celestion CD-driver an optimum entry angle (of a horn/wg) is 25deg and this is not the driver's wall exit angle (which even is negative in this example), rather it continues the cross-section expansion rate -- which looks right to me as long as we are below the critical frequency when wavelenghts start to be important.
- Klaus
It is all wavelength dependent. If the waves are much longer than the dimensions involved then the angles don't matter much. But when the wavelengths from the diaphragm get to be comparable to dimension then these angles all matter. So there is no one right answer. What Tinitus suggests is correct at the lower freqs, but its much more complicated than that at the upper freqs. Its comlicated enough that only a numerical sim could actually sort it out. The simplified lumped parameter approach that is usually applied to phase plug design is only approximate at the upper freqs.
gedlee said:
Thats the point.
But for the mechanism of diffraction it isn't IMO.
Meaning - if we set diffraction as the mechanism (or by-product, or reason) of change of wave front shape.
Or maybe better put - at least this pix of wave fronts changing its shape give us a clue that at this very points of "shape bending" diffraction is happening.
The swiss cheese (sound field) outcome is heavily dependant on the frequencies invoelved due to interference patterns happening - but the point is - that diffraction "per se" happens independently from that.
Meaning . you will see it - gradually - all over the freqeuncy range
Michael