Horn Loading and How To Figure Out When It Ends.

[Trdat]

I have the Dayton B52 1014 horn and although its rated down to 600hz I have read that a horn that size is only good till about 1000hz.

I found a post on the DIY audio with someone who has taken measurements with the horn and the same CD DE250 with up to 40 degrees off axis but I dont now how to tell when the horn loading ends. Not really sure what I am looking at and if I am looking at the right graph. If anyone could help? Below is 40 degrees off axis...

 

[alex-z]

If you want to analyze the horn effect you need raw measurements of the driver before the horn is added. If an anechoic environment is unavailable, gated measurements in a low reflection environment can also suffice.

 

[Trdat]

If you want to analyze the horn effect you need raw measurements of the driver before the horn is added. If an anechoic environment is unavailable, gated measurements in a low reflection environment can also suffice.

Raw measurements of the driver off axis?

But how does the drivers measurements tell us if the horn itself is capable of keeping directivity that low?

 

[alex-z]

Raw measurements of the driver off axis?

But how does the drivers measurements tell us if the horn itself is capable of keeping directivity that low?

You learn the horn capability by doing the A/B comparison of the raw driver vs driver + horn. Same principle as measuring completed speakers. A full set of CEA-2034 data will show the directivity of any driver.

https://www.audiosciencereview.com/...ents-spinoramas-with-rew-and-vituixcad.21860/

 

[puppet]

Understanding Horn Directivity Control (prosoundtraining.com)

 

[Trdat]

Understanding Horn Directivity Control (prosoundtraining.com)

Okay, Ill have to read it and digest it. Thanks.

 

[fluid]

I have the Dayton B52 1014 horn and although its rated down to 600hz I have read that a horn that size is only good till about 1000hz.

I found a post on the DIY audio with someone who has taken measurements with the horn and the same CD DE250 with up to 40 degrees off axis but I dont now how to tell when the horn loading ends. Not really sure what I am looking at and if I am looking at the right graph. If anyone could help? Below is 40 degrees off axis...

A spectrogram will tell you nothing about the polar directivity. You are correct that the B52 1014 begins to lose pattern control around 1K.

DonVK simulated this horn based on the dimensions he measured this will give you a better idea of just what the waveguide does as it is simulated with a constant acceleration source.

https://www.diyaudio.com/forums/multi-way/326926-modular-active-3-progress-post6528840.html

The baffle size and depth of the cabinet will also have an effect on the low end directivity. The combination of all of these things together with the drivers natural response will determine how low it can be used.

Loading is something that is less clear cut with constant directivity waveguides. They are usually designed for polar control. To increase loading usually needs the guide to be more exponential to load at lower frequencies but this results in more beaming at the higher frequency end. The radiation impedance plot shows these sort of effects.

Have a look at this article from Bjorn Kolbrek where the following was taken from part 2
https://www.grc.com/acoustics/an-introduction-to-horn-theory.pdf

"The OS waveguide does not have a sharp cutoff like the exponential or hyperbolic horns, but it is useful to be able to predict at what frequency the throat impedance of the waveguide becomes too low to be useful. If you set this frequency at the point where the throat resistance is 0.2 times its asymptotic value, so that the meaning of the cutoff frequency becomes similar to the meaning of the term as used with exponential horns".

If you were to use this definition on the simulated waveguide it gives a value of ~1100Hz

 

[Trdat]

DonVK simulated this horn based on the dimensions he measured this will give you a better idea of just what the waveguide does as it is simulated with a constant acceleration source.

https://www.diyaudio.com/forums/multi-way/326926-modular-active-3-progress-post6528840.html

https://www.diyaudio.com/forums/multi-way/326926-modular-active-3-progress-post6528840.html

I will look into this and try and understand measurements with some theory.

The baffle size and depth of the cabinet will also have an effect on the low end directivity. The combination of all of these things together with the drivers natural response will determine how low it can be used.

I had no idea that the size and depth of cabinet effects low end loading. Any specifics such as larger cab or deeper cab gives more loading...?

Loading is something that is less clear cut with constant directivity waveguides. They are usually designed for polar control.

Again this is great, at least I know what I am trying to discern.

To increase loading usually needs the guide to be more exponential to load at lower frequencies but this results in more beaming at the higher frequency end. The radiation impedance plot shows these sort of effects.

And from my understanding the B52 is an imbetween which manages both ends well hence why the low end loading is not that great.

Have a look at this article from Bjorn Kolbrek where the following was taken from part 2
https://www.grc.com/acoustics/an-introduction-to-horn-theory.pdf

If you were to use this definition on the simulated waveguide it gives a value of ~1100Hz

1100 is pretty high, I got a bit of a gap between my 15 inch and the low end of my horn/CD. Anyway, good to know thanks...

 

[digitalfrost]

No sure if it's worth the money to you, but when simulating horns it's worth to take a look at http://www.aj-systems.de/indexe.htm

 

[kyle_neuron]

No sure if it's worth the money to you, but when simulating horns it's worth to take a look at http://www.aj-systems.de/indexe.htm

This is a one-parameter model from what I can tell, so you might as well use Hornresp. Akabak is more involved, but being a BEM+LEM solver it’s going to be more accurate if used properly. It’s what’s been used in the DonVK sims shared above.

A license isn’t cheap, but for the hobbyist you can use the trial. It’s only limited in terms of saving results. Some familiarity with CAD tools is also useful, of course. Even if you’re going to ignore the BEM part, the Lumped Element tools don’t have the segment restrictions that Hornresp or other tools do.

 

[fluid]

A license isn’t cheap, but for the hobbyist you can use the trial. It’s only limited in terms of saving results. Some familiarity with CAD tools is also useful, of course. Even if you’re going to ignore the BEM part, the Lumped Element tools don’t have the segment restrictions that Hornresp or other tools do.

For non commercial use Jorg Panzer will give out a full student licence for ABEC/AKABAK and VACS which allows saving of everything. Much can still be done with the trial version but rerunning a 10 hour simulation isn't fun. ABEC is what I use and Don uses too. The latest version is AKABAK 3 which is more of a point and click system moving away from the script based input in ABEC. The learning curve is steep and time consuming but well worth it.

 

[Plcamp]

Understanding Horn Directivity Control (prosoundtraining.com)

That’s a really good article, thanks for posting.

 

[kyle_neuron]

For non commercial use Jorg Panzer will give out a full student licence for ABEC/AKABAK and VACS which allows saving of everything. Much can still be done with the trial version but rerunning a 10 hour simulation isn't fun. ABEC is what I use and Don uses too. The latest version is AKABAK 3 which is more of a point and click system moving away from the script based input in ABEC. The learning curve is steep and time consuming but well worth it.

Yup, he’s great for that if you’ve got a valid non-commercial use for it. I’m very interested what you’re doing that takes ten hours to solve? It must be something that needs a super dense mesh size or accurate model of compression driver phase plug behaviour!

I’ve had great success with model and mesh optimisation, a current project I’m working on has reduced from 15 minute solves to just over a minute - on an AMD Ryzen 5600X - from a change of approach to my geometry and subdomain arrangement. Which is important for the next step…

I don’t think the learning curve is particularly steeper than any of the other tools either. Especially with the latest Akabak, which has a fairly intuitive GUI without restricting the scriptable flexibility of ABEC3.

Even if you’re only simulating things as a lumped element model, it’s a good option to look at due to the flexibility and lack of restrictions on segments.

I think people are perhaps put off by the concept and Akabak’s prior reputation? The example files and help file included are also very good, although I have considered putting together a ‘beginner’s guide’ that extends the great CAD to sim workflow videos for waveguides by Austin Mys on YouTube. With a bit of planning and the right tools, you can revise a mesh for a new sim fairly quickly. Especially for the price compared to a COMSOL license.

 

[fluid]

The long simulation times are always with dense meshes to give them half a chance of being accurate up to 10K. I much prefer axisymmetric waveguides that can use circsym as they are done in seconds.

Full 3D simulations of large objects and complete cabinets over the whole frequency range take time, but some of that will come down to the computer too. Some things can't easily be optimized and brute force high resolution is a simpler solution. Having good CAD drawing skills opens up much more possibilities.

I agree that Akabak 3 is easier to get to grips with being point and click but for me now that I am used to the scripts I don't want to give them up! I've seen and posted about those Youtube videos before they are a really good introduction and there are a couple on the R and D Team website that are helpful too. I've also considered making a how to thread but I suspect it would become quite a time drain so I have resisted