How to Make a New Wave Biradial Horn

cXyfrLx.jpg


txix8Kz.jpg


js4vKur.jpg


Here's a few "classis" biradial horns.

UNYgvmu.jpg


While pondering the JBL M2 and the JBL progressive transition waveguides (1) it occurred to me that there's a way to make something similar to those, but not quite the same.

One thing that's cool about this method is that it's exceptionally easy. Some of my waveguide designs have taken as much as sixteen hours to draft, but these are fairly easy and approachable to someone new at 3D. (If you've seen my 3D printed synergy horn designs, I have to manually 'drill' every single bolt in three dimensions and that takes fo-rev-er.)

(1) JBL M2 for The Poors
 
JWIjQD3.png


0qRr77B.png


WqeJI3L.png


As illustrated above, we create four hemispheres and we bond them into a single solid. The diameter of each hemisphere should be approximately as large as you want your waveguide to be. For instance, I'm using a 25cm wide hemisphere to achieve a 25cm wide waveguide. (Again, this is approximate.)
 
Ot3ro3v.png


Here's one of my favorite tricks when I make waveguides. If you space four solids by 25.4mm apart, combine them, then smooth them, you get a 1" round throat. This is because the process of smoothing the solid turns the square throat into a circular throat.

15nNGYR.png

You can see this same trick used in the QSC waveguides to get an absolutely perfect transition from round at the throat to rectangular at the mouth.

Once you fiddle around with this, you'll see there are ways to do this with a throat that's perfectly matched to the exit of the compression driver. For instance, you can bring the solids closer together than you normally would, smooth the throat, and then thrust a cylinder through the throat to match the compression driver.

tJOIQiF.png


2ixDd9W.png


Here's the throat after smoothing. And then I trimmed the excess parts of the hemisphere.

You could leave it at this, and have something similar to the older biradials:

U2FugXU.jpg


But most will want something more compact...
 
  • Like
Reactions: 1 user
pg6lRdD.png


What if you took an oblate spheroidal waveguide

tfzMzBT.png


And a biradial horn

ncIGVui.png

Then merged them together?


Note that this one could be tweaked quite a bit. You could have the biradial curve merge with the OS waveguide curve at any point on the biradial curve. (Because the biradial curve is a full 90 degrees.)
 
As noted in post #3, most people will want a waveguide that's as small as possible.

8HuB2qn.png

Here's a biradial horn that measures 25cm x 25cm x 16.5cm. (9.8" x 9.8" x 6.5")

3nOG4Nx.png

Here's how that horn looks if we just truncate the entire face of it. The size has been reduced to just 12cm deep. That's 27% shallower than before.

fNfDXUl.png

If we smooth that horn, we get this. And now it's starting to look a lot like a progressive transition waveguide. Perhaps the PT waveguides are basically biradial horns that have been truncated and smoothed?

attachment.php

JBL Progressive Transition Waveguide for comparison. Note the 'sharp' cutoff where the horn curve has been truncated.
 
  • Like
Reactions: 1 user
Patrick,

Good food for thought. Thanks for posting. Which software do you use for this modelling?

Another reason for doing it is if no commercial options were available. Like I'm exploring use of a 2" CD on a constant directivity horn that can be used down to 400Hz or so, and with a rectangular form factor to allow tweeter above.
 
Why not using something someone already made?

The thing that got me interested in the progressive transition waveguides was that they combined the incredible polar response of an oblate spheroidal waveguide, but with superior impulse response.

IE, if you want absolutely stellar polars, it's difficult to beat an OS waveguide. For instance, I've never measured anything that has better polars than the big QSC waveguide.

I built a LeCleach horn a few years ago and it had excellent impulse response. But the PT waveguides seem to be a mix of both: excellent polars and excellent impulse response.

As for why I 3D print them, well that's because I like Synergy Horns!

I use Autodesk 123D for all my stuff.
 
An externally hosted image should be here but it was not working when we last tested it.

Here's the waveguide that you get if you use four cones instead of four hemispheres

An externally hosted image should be here but it was not working when we last tested it.

If you slice the waveguide in half, you can see that the profile is fairly similar to oblate spheroidal. Basically the walls are largely flat and there's a huge roundover at the mouth.
 
The thing that got me interested in the progressive transition waveguides was that they combined the incredible polar response of an oblate spheroidal waveguide, but with superior impulse response.

I don't get how a good polar response does not provide an equally good impulse response.

Why not just take your sphere and rotate it around the circular aperture, like a donut. Your design will have failing polars along the diagonals. The rotated sphere won't.
 
I don't get how a good polar response does not provide an equally good impulse response.

Good point. I may have been misinterpeting the impulse response. Pardon my ignorance, but it just occurred to me that you'd have to equalize both horns to the same frequency response to do a proper comparison of the impulse. Is that right?

IE, if I measure two horns without equazling them flat, and then look at the impulse, I'm comparing apples to oranges?

"If the transfer function of a system is given by H(s), then the impulse response of a system is given by h(t) where h(t) is the inverse Laplace Transform of H(s). A less significant concept is that the impulse response is the derivative of the step response."

Why not just take your sphere and rotate it around the circular aperture, like a donut. Your design will have failing polars along the diagonals. The rotated sphere won't.

jrx212_jrx212.jpg


I'm largely tinkering with M2 and PT style waveguides, because the PT waveguide measures so nice. Both have 'beaks' which should widen the polar response on the X and Y axis, at the expense of the orthogonal polar response.