Interestingly, I heard that the OSWG doesn't scale to larger size/lower cut off as it looses pattern control up high if you extend the low end. I don't know if anyone can confirm this.
I am looking to make a CD horn to cover 500hz-20khz, for a 1.5 inch compression driver.
The horn needs to be conical though, and would ideally have a roundover on the mouth.
I have yet to find anything that would meet the criteria that I can purchase.
I am not aware of an oblate spheroid "losing pattern control up high" as you increase its size to extend the low end. There may however be a mild bandpass filter effect which reduces the top end energy a bit.
At some point the compression driver will beam enough on its own that the waveguide is no longer controlling the pattern. Imo what frequency this starts to happen depends on the specifics of the compression driver and waveguide.
Earl's data shows his largest 1" throat waveguide still has pattern control at 20 kHz. The challenge is the bandwidth of the compression driver. The B&C DE500 compression driver needs a bit of lift in the top half-octave or so to "keep up" with the 15" woofer he uses. Comparing the top ends of the NS15 with the New Abbey 12 (which uses the same compression driver), the larger waveguide of the NS15 seems to have a little bit less top end.
It's not that Oblate Spheroid waveguides result in less top-end energy; rather, it's that the energy is spread thinner because of the wide coverage angle at high frequencies. Most horns exhibit pattern narrowing in the top octave, so the on-axis SPL is correspondingly higher in the top end. The horn on the JBL M2 has a very wide pattern (for a large-format horn) all the way up, so its top-end energy is spread over a very wide area resulting in reduced sound
pressure level within the pattern. This may have played a role in the arguably fairly modest 92 dB system efficiency.
About eighteen years ago Earl made me a one-off 2" throat Oblate Spheroid waveguide for a TAD TD-4001 compression driver. It DID beam badly in the top octave or so, but I'm pretty sure this was a function of the long, narrow-angle throat inside of the oldschool-format TD-4001, which was long enough to constrain the radiation pattern into a narrow angle at short wavelengths. I do not think a modern short-throat large-format compression driver would have the same issue to the same extent, but it would probably still start beaming at a lower frequency than a comparable 1-inch-throat compression driver.
I plan to experiment with a couple of 1.4" throat Oblate Spheroid waveguides later this year, but nothing large enough for pattern control all the way down to 500 Hz.
Duke