Hi mabat,
Kindest regards,
M
You are awesome, I do not care what people say about you. 😉
The driver is JBL 2450SL, the attached drawing lists it as 38.1 (1-1/2). I rechecked with calipers and it is indeed 1.5 inch - or 38.1mm for us Moravians. 😀
Kindest regards,
M
Hi mabat,
I forgot - the phase plug ends at the face-plate of the driver; therefore, the exit angle is 0 deg.
Kindest regards,
M
I forgot - the phase plug ends at the face-plate of the driver; therefore, the exit angle is 0 deg.
Kindest regards,
M
As a side note, I'm not sure this is the whole causality, actually. If there was no expansion in the phase plug itself, 0 deg would be the right continuation, but do we know this expansion? I've never seen it mentioned anywhere. If there was a non-zero expansion, the horn should probably continue with the same rate (at least locally), but what to do if the phase plug is designed to produce a flat wavefront at the same time?
I'm no Moravian, but I know what a millimeter is 🙂 Thanks, it could easily be slightly different.
Last edited:
Hi mabat,
But what about a driver with a different arrangement? The first one I saw was the RCA 14XX driver. It is essentially a truncated cone with tangential slots, the smaller diameter surface of the truncation ending at the exit. The tangential slots expand from the larger cone diameter surface towards the smaller diameter surface.
The other question - and this may be controversial, does the angle matching matter? Let us consider a conical exit, mated to the same angle of OS wave-guide. There is still a discontinuity at the matched surface.
Another question - assuming that there is a flat wavefront at the exit (as claimed for the JBL driver) - is the 0 deg angle not the correct for OS wave-guide per Geddes?
Kindest regards,
M
I have always wondered about this. The accepted notion of the exit angle is the (usually) cylindrical or conical section following the end of the phase plug.
But what about a driver with a different arrangement? The first one I saw was the RCA 14XX driver. It is essentially a truncated cone with tangential slots, the smaller diameter surface of the truncation ending at the exit. The tangential slots expand from the larger cone diameter surface towards the smaller diameter surface.
The other question - and this may be controversial, does the angle matching matter? Let us consider a conical exit, mated to the same angle of OS wave-guide. There is still a discontinuity at the matched surface.
Another question - assuming that there is a flat wavefront at the exit (as claimed for the JBL driver) - is the 0 deg angle not the correct for OS wave-guide per Geddes?
Kindest regards,
M
All I can say is that this should be ideally designed/optimized as one single system. Otherwise, you just have to work with what you have, and in that case this ideal may be simply unreachable. (The small Peerless driver, for some reason, allows this ideal to be actually reached.) Exit angle mismatch between a driver and a horn/waveguide can have only a small to negligible effect on directivity, but at the same time can pretty strongly affect the overall "evenness" of the frequency response due to the throat reflection (which can, in principle, be equalized, though). Not sure about the higher order effects, but I suspect these will be typically high enough in frequency for not to matter much. To me, having a smooth directivity with as least diffraction as possible (as observable in a normalized graph), is the primary goal. The rest is manageable on the signal level.
Last edited:
Now I only hope I remember it right, but a flat wavefront is in fact not optimal for OS waveguide. I think was shown by Earl Geddes a long time ago - it's in his papers or the book. By a phase plug design, an even more suitable wavefront could be -in theory- constructed (with decreasing velocity towards the edge, IIRC), but that's far above the capabilities of most of us.
Last edited:
Last edited:
I also have the first A460 printed and waiting. I would need someone for assembling these things... 🙂
It was really an easy print, seems to fit nicely so far. (It was a little less than one 1kg spool of filament per waveguide.)
It was really an easy print, seems to fit nicely so far. (It was a little less than one 1kg spool of filament per waveguide.)
I would suggest simply something you're familiar with and what gives nice and accurate prints on your printer. I haven't done any research regarding mechanical or acoustical properties, as I still do mostly prototypes, and for that I'm quite happy with the cheapest PLA (the above is PLA Meta made by SUNLU, virtually warping free). After epoxy coating, I believe these could do just fine. I still tend to make it very light, basically hollow.
Hi mabat,
But, I had been wrong before.
Kindest regards,
M
Indeed, but I do not think that there is enough data and/or understanding how to do that, your efforts notwithstanding.
Hm, I certainly to not have your mathematical prowess, but what I gathered from Geddes' papers was that he was trying to find a shape that would allow a flat wavefront (presumed to be at the exit of a compression driver) to a wavefront at the selected wave-guide coverage angle with the least possible disturbance.
But, I had been wrong before.
Thank you, very much, I though that it would be for the wave-guide in post post 15,578 as we discussed. Well, I will go with the Gen 2 then. Will it work both on the 460 and the 520?
Kindest regards,
M
I've just checked Earl's papers again (what a tremendous amount of work), and indeed, it can be done better than a flat wavefront for driving an OS waveguide. It's clearly explained in his work.
Maybe you could even try to put a small amount of foam into the phase plug rings, an increasing amount for the larger ones.
Maybe you could even try to put a small amount of foam into the phase plug rings, an increasing amount for the larger ones.
Ah, then we misunderstood each other. For a 1-1/2 inch throat it's a bit different situation and I doubt it can be done better than with the current Gen2 (above). Yes, it will work basically the same either on the 460 or the 520.
Dear Mabat. What do you think about a midbass waveguide that holds down its pattern down to 120hz which is close what most rooms have schroeder frequency is around and where one can cross to a sub without much problems. It can be for 8" to 12" drivers. How large would it be with ATH and how optimized can it be with corresponding drivers? If this works there can be a 520gen2 and that midbass waveguid. Then just add sub to ones room and preference.
ps. I have a cf:110hz tractric horn with 1 meter diameter in my room and it adds so much dynamic realism to all music.
ps. I have a cf:110hz tractric horn with 1 meter diameter in my room and it adds so much dynamic realism to all music.
Here's a direct comparison of the A460 (1") and the A460G2 (1.5" adapter), as BEM simulated:
So the differences are on the order of 1 dB at 10 kHz.
(SP is the Sound Power curve, as always 😉)
So the differences are on the order of 1 dB at 10 kHz.
(SP is the Sound Power curve, as always 😉)
Last edited:
And the same LW-normalized, i.e. with the listening window averaged response (as defined in CEA2034) being flat.
(The same data in two different graphs. Still searching for my new presentation standard 🙂)
(The same data in two different graphs. Still searching for my new presentation standard 🙂)
Last edited:
I don't know, go for it, but I wouldn't. It would need to be huge, like at least 2.5 meters, which seems like nonsense to me.
I think the important question here is "compared to what?". Is it the efficiency/sensitivity, or the directivity being controlled better at the lower end? With a 1m waveguide you're almost at Schroeder. I can't imagine that an even bigger would have an advantage, but I don't know.