How to Make a New Wave Biradial Horn

And to your three points - aren't # 2 and 3 the same thing?

I guess a horn could be 'round' (have a round mouth) without being symmetric, if it had uneven expansion near the throat, like the 18 Sound horns, or butt cheeks, like some JBL horns.

There are a few square mouthed horns that have done similar e.g. the P.AUDIO PH-220* mouth is 220mm x 220mm, and dispersion is meant to be 90° x 75°

What would you say about the SEOS WG ?... they are not specifically designed to operate correctly with all Compression drivers, they seems to be a very aleatory choice ?

To my eye, the SEOS horns look nice, and very reasonably priced.

Apparently the throat angle was optimised for the DE250/DNA-360. Nothing aleatory about it :)

I found some lists of exit angles**, and there are a several other drivers that are close to that ...and you can always DIY adjust the horn. That is, if I wanted to mate a SEOS to a driver with a narrower throat angle, I'd add an adapter plate to lengthen the SEOS throat slightly, and then re-grind the (composite) throat to get a smooth transition.

*The RCF H100 is the same horn, in plastic.

**why are these not on spec sheets?
 
To my eye, the SEOS horns look nice, and very reasonably priced.
Apparently the throat angle was optimised for the DE250/DNA-360. Nothing aleatory about it :)
I found some lists of exit angles**, and there are a several other drivers that are close to that ...and you can always DIY adjust the horn. That is, if I wanted to mate a SEOS to a driver with a narrower throat angle, I'd add an adapter plate to lengthen the SEOS throat slightly, and then re-grind the (composite) throat to get a smooth transition.

the loudspeaker unit parameters consitancy seems very critical with a passive filter desing, the noches have to be accuratly centered ont the peaks and there are many ways to destroy the equilibrium.
The amplitude and the Q values of the corrections are hard to fix and will also harm the phase if you don't really know what you are doing.
A compression is not a direct radiator, the errors have a tremendous impact on the sound.
 
What would you say about the SEOS WG ?

I would say that it has all the compromises that one would expect from its design. It is small and as such cannot control below about 1500-2000 Hz, which is too high for my taste. It is even smaller vertically and as such the vertical control fails even lower than the horizontal and as such lacks control through the crossover region where the lobbing occurs making it worse. It has almost no control along the diagonals due to its rectangular nature. Finally, the mouth flare is not sufficient to eliminate reflections. All of these are compromises accepted by the designers in order to get a smaller/cheaper device.
 
I would say that it has all the compromises that one would expect from its design. It is small and as such cannot control below about 1500-2000 Hz, which is too high for my taste. It is even smaller vertically and as such the vertical control fails even lower than the horizontal and as such lacks control through the crossover region where the lobbing occurs making it worse. It has almost no control along the diagonals due to its rectangular nature. Finally, the mouth flare is not sufficient to eliminate reflections. All of these are compromises accepted by the designers in order to get a smaller/cheaper device.

Do you share the drawings of your original design ?
Here is a tessellation of the one that i've drawn, but the dimensions shall not be right...
 
It is small and as such cannot control below about 1500-2000 Hz, which is too high for my taste

Many people have taken polars that, if I'm reading them correctly, do not concur with this. The 12 and 15 look like they are fine (horizontally) right down to the crossover point.

Seos 12:
At last, the SEOS12 Measurements - AVS Forum | Home Theater Discussions And Reviews

SEOS 15:
My living room "make over" (aka the "surrounded by bass" project) - Bass Gear - Data-Bass Forums

It is even smaller vertically and as such the vertical control fails even lower than the horizontal and as such lacks control through the crossover region where the lobbing occurs making it worse.

The SEOS vertical polars are indeed a lot more blobby, but I don't understand lobing well enough to know how much pattern control is needed to get benign lobing :(

I was under the impression that a good design will 'steer' the lobes so that there are no big notches on and around the listening height, but I don't really understand how this is calculated / how the lobes are 'steered' (I guess I should read up + noodle around with some simulating software until the concept clicks a bit better).

Finally, the mouth flare is not sufficient to eliminate reflections.

The horizontal transition from the horn angle to the baffle seems very large / smooth, so I assume you mean that the vertical roundover should (ideally) be larger?

If so, wouldn't the flipside of increasing this (the compromise), be that you'd increase C-to-C spacing?

All of these are compromises accepted by the designers in order to get a smaller/cheaper device.

? Anyone who thinks the SEOS 15 is too small & cheap can order a SEOS 18 (or 30), and have it anointed with majik audiophile lacquer :)
 
diyAudio Moderator
Joined 2008
Paid Member
I was under the impression that a good design will 'steer' the lobes so that there are no big notches on and around the listening height, but I don't really understand how this is calculated / how the lobes are 'steered' (I guess I should read up + noodle around with some simulating software until the concept clicks a bit better).
You won't get this from your average crossover simulator, you can work it out geometrically.

If so, wouldn't the flipside of increasing this (the compromise), be that you'd increase C-to-C spacing?
It isn't the case that an elliptical mouth reduces spacing, unless it also compromises the horn.
 
Speaking of different horns, I would like to try this. In any cross section with a plane containing the axis it is a hyperbola terminated by Euler spiral. So it's basically assembled from these two parts. The shape of a mouth can be arbitrary, superellipse seems just fine. If it was a circle, then it would be a spiral-terminated OS. I have no idea how well it will behave acoustically with a non-circular mouth. I've written an utility that computes the coordinates - if anyone is interested.
 

Attachments

  • wg-t2.png
    wg-t2.png
    174.8 KB · Views: 458
  • wg-t.png
    wg-t.png
    142.5 KB · Views: 454
Speaking of different horns, I would like to try this. In any cross section with a plane containing the axis it is a hyperbola terminated by Euler spiral. So it's basically assembled from these two parts. The shape of a mouth can be arbitrary, superellipse seems just fine. If it was a circle, then it would be a spiral-terminated OS. I have no idea how well it will behave acoustically with a non-circular mouth. I've written an utility that computes the coordinates - if anyone is interested.

In the fabrication process a 3D solid generation is useless since the cost requied for a suficent quality part (3D printing) is more elevated than a new luxury car.
A matrix of points in order to generate facets seems more adequate IMHO
 
I was under the impression that a good design will 'steer' the lobes so that there are no big notches on and around the listening height, but I don't really understand how this is calculated / how the lobes are 'steered' (I guess I should read up + noodle around with some simulating software until the concept clicks a bit better).

The main lobe is "steered" both with physical driver spacing on the Z-axis and in the crossover. Secondary lobes should be minimized so as to have a first reflection response as uniform as possible. This is often done using steep slopes. 4th order LR24 is fairly common for DSP users.

Imagine two 4" speakers in separate 4" cubic enclosures sitting on top of each other. Each speaker's sound arrives to you in phase or at the same time, the main lobe is thus horizontal, let's call that 0°. Now move the top speaker back by 4" and you can easily see how the point where the sound arrives in phase (again main lobe) is now shifted up to 45°.
 
the vertical control fails even lower than the horizontal and as such lacks control through the crossover region where the lobbing occurs making it worse. It has almost no control along the diagonals due to its rectangular nature. Finally, the mouth flare is not sufficient to eliminate reflections

I can imagine that, but i need some data to be convinced.

I have no idea how well it will behave acoustically with a non-circular mouth

You seems to be intelligent but you can't modelize exactly the response with your brain... me neither.

have it anointed with majik audiophile lacquer :)
I don't see any measurement about the diagonals.

--------------------

I've recently sent a mail to a sympathetic Italian manufacturer in order to obtain a 3D polar measurement of a horn, here is the answer :
"We didn’t receive any request of 3D diagrams before now.
In our opinion our horn have quite simple design, its are not so complicated, vertical and horizontal pattern are sufficient to understand the behaviour"


So, if vertical and horizontal pattern are sufficient to understand the behaviour, it seems that the whole world is idiot.
 
Last edited:
Hi. In fact these are the same data. I suppose I should start a new thread...

672389d1522592804-wave-biradial-horn-celo_420x1100x44s-png


674082d1523299304-smooth-transition-horns-t_wg_rr_951-png


The more that I tinker with the JBL IC waveguides, the more I think the (older) JBL PT waveguides may be superior :)

attachment.php


In particular, my ABEC sims really demonstrate that fast transitions screw everything up. I'm not convinced that diffraction slots are 100% bad, but if you're going to use one it better be realllll smooth.