At 507 Hz the normalized throat impedances are:
0.008 + 0.260i (smaller WG)
0.032 + 0.264i (larger WG)
So the real part is 4x higher, actually. It that a plausible result? I don't know. It doesn't mean much anyway, none of those two waveguides would be ever used that low, but it's interesting.
It seems that with the larger waveguide there's still an obstacle to the wave propagation that is already almost gone with the smaller one. That's also why the DI is still considerably higher.
0.008 + 0.260i (smaller WG)
0.032 + 0.264i (larger WG)
So the real part is 4x higher, actually. It that a plausible result? I don't know. It doesn't mean much anyway, none of those two waveguides would be ever used that low, but it's interesting.
It seems that with the larger waveguide there's still an obstacle to the wave propagation that is already almost gone with the smaller one. That's also why the DI is still considerably higher.
Last edited:
Yep interesting, looks roughly 6dB increase in power there around 500Hz as well. Make the real impedance 0.128 and get power above 90db at 500Hz?!
Instead of reaching for sharper impedance cut off with the peak (high Q, the exponential horn), one could make the impedance knee low Q and extend the "tail" lower in frequency? Well, having fun reasoning, I don't have a clue how it all plays out
Instead of reaching for sharper impedance cut off with the peak (high Q, the exponential horn), one could make the impedance knee low Q and extend the "tail" lower in frequency? Well, having fun reasoning, I don't have a clue how it all plays out
Last edited:
If you look the GIF impedance plot carefully there is difference in many parts, both the peak and the turn around ~1.5kHz are "smoother" with the bigger device. The highs have roles reversed though, smoother with the smaller device.
I speculate from the impedance plots that you are approximating the "best" profile with these two devices from both sides of it, if the smoothness of the plot is some sort of measure of goodness. The small peak there is in the impedance dictates there be ripple higher up, or vice versa, no peak no ripple, like sharp/shallow filters in frequency domain. Reasoning from this speculation the smoothest impedance plot is when there is no peak at the knee, no need for ripple, smoothest response but no loading what so ever, right? I guess the loading and performance would happen if you could shift the whole throat impedance graph in frequency, or push the platoe higher past 1. I have no idea what this would mean in reality or is there any connection with reality
edit. some "long impedance tails" on conical devices here post #8276
I speculate from the impedance plots that you are approximating the "best" profile with these two devices from both sides of it, if the smoothness of the plot is some sort of measure of goodness. The small peak there is in the impedance dictates there be ripple higher up, or vice versa, no peak no ripple, like sharp/shallow filters in frequency domain. Reasoning from this speculation the smoothest impedance plot is when there is no peak at the knee, no need for ripple, smoothest response but no loading what so ever, right? I guess the loading and performance would happen if you could shift the whole throat impedance graph in frequency, or push the platoe higher past 1. I have no idea what this would mean in reality or is there any connection with reality
edit. some "long impedance tails" on conical devices here post #8276
Last edited:
Can you plot that easily for your optimizer runs? It could reveal some interesting "hidden" data.
Other than that check your comment here: Is it possible to cover the whole spectrum, high spl, low distortion with a 2-way?
Have you tried to find a throat with some loading and then ran the optimizer to parameters that don't change the throat much? If one could select the loading and then find a good pattern control by brute force?
Eagerly waiting what the curly throat phase plug thing effect is on loading (and pattern)
Other than that check your comment here: Is it possible to cover the whole spectrum, high spl, low distortion with a 2-way?
Have you tried to find a throat with some loading and then ran the optimizer to parameters that don't change the throat much? If one could select the loading and then find a good pattern control by brute force?
Eagerly waiting what the curly throat phase plug thing effect is on loading (and pattern)
Last edited:
Ah increased loading on the backside should be utilized in the front somehow? This would require modifications to the front (reflex port to the waveguide) and probably ruin it.
Most ideas are already found out decades ago so probably my ideas are in vain, but it is so much fun can't help posting
Most ideas are already found out decades ago so probably my ideas are in vain, but it is so much fun can't help posting
Well I guess it should. What would it be good for otherwise?
The truth is that all the radiation from the back side of the diaphragm is normally wasted (turned into heat inside the driver). But I don't have a clue what can be done about that.
Last edited:
I've got no idea I haven't thought what is the mechanism why increased loading leads more output. It is probably not something that affects the diaphragm itself but something that affects the molecules the diaphragm excites in which case the loading must have effect on the front? This would go into the compression driver technology which is very mature I suppose.
.. Could there be ducting system between the front and back of the diaphragm within the compression driver so that the waveguide would stay intact? I guess so, a mid reflex compression driver probably some physics would limit the usability, too much air speed or something would be required and it wouldn't work out. Otherwise there would be an example already.
I haven't thought what is the mechanism why increased loading leads more output. It is probably not something that affects the diaphragm itself but something that affects the molecules the diaphragm excites in which case the loading must have effect on the front? This would go into the compression driver technology which is very mature I suppose. .. Could there be ducting system between the front and back of the diaphragm within the compression driver so that the waveguide would stay intact? I guess so, a mid reflex compression driver
probably some physics would limit the usability, too much air speed or something would be required and it wouldn't work out. Otherwise there would be an example already.
Last edited:
For a given diaphragm velocity, the higher the acoustical impedance loading the diaphragm, the greater will be the power transfer.
Power = (volume velocity ^ 2) * acoustical impedance
(analogous to P = I ^ 2 * R in an electrical circuit)
Thanks David!
The formula says power would increase also with increased volume velocity. But that is probably tied somehow to sound frequency? Or would it be matter of using smaller throat diameter, more compression? Or just use the EQ to increase power until mechanical problems. Well, I should read more. Hopefully everyone got their brains fed and better ideas pop up
The formula says power would increase also with increased volume velocity. But that is probably tied somehow to sound frequency? Or would it be matter of using smaller throat diameter, more compression? Or just use the EQ to increase power until mechanical problems. Well, I should read more. Hopefully everyone got their brains fed and better ideas pop up
Last edited:
Here is nice lecture paper if someone is interested. Not sure if this is any relevance to compression drivers and horns but nevertheless, some basics about acoustic loads. http://electroacoustique.univ-lemans.fr/cours/pdf/grain_42en.pdf
To maximize the horn size within a single print, it turns out it is possible to print with the horn tipped up at an angle, fitting the mouth inside the print volume diagonally. I am working with a CR6-SE printer with a build size of 235 x 235 x 250 mm, and not only can I print a horn with the external dimensions of the PH-2510, adding the klipsch woofer notch allows me to print two mirror-image horns at once, which seems slightly faster than printing two horns serially. Coming up with a shape worth printing even one of is a different matter entirely, of course.
Next I would like to experiment with printing in wood-PLA composite instead of normal PLA, and using the 'gyroid' infill available in the slicer software as external ribbing instead of the ribs which I have modeled explicitly in CAD - I think the gyroid shape is open cell, so perhaps there'd be an advantage there in terms of displacing less cabinet volume. With a coarse enough gyroid pattern, I wonder if it would be possible to pour resin into the open gaps.
The profile of the horn shown is more or less what I showed in my demo of the surface loft command, only instead of an 8mm long cylindrical throat which is easy to see in screenshots, I shortened the cylinder section to 0.1mm. Ultimately this will be replaced with an ath4-generated surface, for now I am just assessing manufacturability and verifying external mounting dimensions.
Next I would like to experiment with printing in wood-PLA composite instead of normal PLA, and using the 'gyroid' infill available in the slicer software as external ribbing instead of the ribs which I have modeled explicitly in CAD - I think the gyroid shape is open cell, so perhaps there'd be an advantage there in terms of displacing less cabinet volume. With a coarse enough gyroid pattern, I wonder if it would be possible to pour resin into the open gaps.
The profile of the horn shown is more or less what I showed in my demo of the surface loft command, only instead of an 8mm long cylindrical throat which is easy to see in screenshots, I shortened the cylinder section to 0.1mm. Ultimately this will be replaced with an ath4-generated surface, for now I am just assessing manufacturability and verifying external mounting dimensions.
Attachments
Ath 4.7.2 released
Ath 4.7.2 now available for download: https://at-horns.eu/release/ath-4.7.2.zip
The biggest change is the added lumped element support. Now we can simulate waveguides driven by "actual" transducers (not only compression drivers but any other as well). All we have to do is to prepare the LE models...
(So far I haven't checked all the demo scripts again, so that's stilll missing at the moment.)
Ath 4.7.2 now available for download: https://at-horns.eu/release/ath-4.7.2.zip
The biggest change is the added lumped element support. Now we can simulate waveguides driven by "actual" transducers (not only compression drivers but any other as well). All we have to do is to prepare the LE models...
(So far I haven't checked all the demo scripts again, so that's stilll missing at the moment.)
Here's a sample script -
Code:
; ST270 - Ath 4.7.2
Throat.Diameter = 25.4
Throat.Profile = 1
Throat.Angle=4
Coverage.Angle = 30
OS.k = 2
Rot=3
Length = 120
Term.s = 1.0
Term.n = 4.0
Term.q = 0.996
Rollback=1
Rollback.StartAt=0.6
Rollback.Angle = 180
Rollback.Exp = 1.5
;Throat.Ext.Length = 12
;Throat.Ext.Angle = 8.5
Source.Shape = 1
LE = generic25
; -------------------------------------------------------
; ABEC
Mesh.AngularSegments = 8
Mesh.LengthSegments = 60
Mesh.SubdomainSlices=
Mesh.WallThickness = 4
; -------------------------------------------------------
ABEC.SimType = 2
ABEC.SimProfile = 0
ABEC.f1 = 200 ; [Hz]
ABEC.f2 = 20000 ; [Hz]
ABEC.NumFrequencies = 100
ABEC.MeshFrequency = 32000 ; [Hz]
ABEC.Polars:SPL = {
MapAngleRange = 0,180,37
}
; -------------------------------------------------------
Report = {
Title = "ST270"
NormAngle = 10
DrvImp_Range = 40
MaxRadius = 150
}
Output.STL = 0
Output.ABECProject = 1