Acoustic Horn Design – The Easy Way (Ath4)

That was normalized, i.e. effectively an EQ used such that the response at selected angle (on-axis in this case) is exactly flat. It clearly describes the WG "alone", without the properties of a particular driver. In the Pet007's data this is the left lower chart. The upper left is the raw data.

That is correct - measurements posted were normalized.

@Matab - could I make a suggestion? Is there any chance the mounting holes can be shifted by 90 degrees (or two additional 1/2 holes added) - so they print on the bed of the printer?

I know in my case I was just taping the sides together, but my concern is whether the two sides will still come apart (despite glue/epoxy) after the pressure of the bolts is applied. See included pic.

BTW - the Black is printed on Pursa by my neighbor - a successful print.
The Yellow was printed on my Ender 3 Pro - an unsuccessful print. I'm not very savvy with 3D printers or modeling and am completely new to it.


Lastly - and I know this is going to sound ridiculous - but I'm going to add that even the frequency sweeps used to measure the horn sounded quite pleasing. I hate taking the measurements, especially of compression drivers. But I've taken measurements of the drivers on QSC clone waveguide, SEOS waveguide, and the Wayne Pi waveguides - and this horn was the most bearable of them all. Anyways ---- let's get back to objective data :).
 

Attachments

  • unnamed (3).jpg
    unnamed (3).jpg
    148.1 KB · Views: 416
...dome tweeters are still a pain to integrate into a WG... :p

Which tweeter(s) are you referring to?

Some work better than others. The surround and - to a lesser extent - the shape / size of the dome play an important role in possible interference with the waveguide walls.
A traditional dome tweeter that pairs well with waveguides is the ScanSpeak D2604/833000.
I'm seriously considering this driver for a Living Voice IBX/OBX clone.

2 comparable tweeters with factory-fitted waveguides, but different surround-waveguide interface:
1. Morel CAT 378
2. Wavecor TW030WA11
 

Attachments

  • Morel_CAT_378_(Photo_3).jpg
    Morel_CAT_378_(Photo_3).jpg
    110 KB · Views: 350
  • Wavecor TW030WA11.jpg
    Wavecor TW030WA11.jpg
    291.6 KB · Views: 342
Last edited:
Is there any chance the mounting holes can be shifted by 90 degrees (or two additional 1/2 holes added) - so they print on the bed of the printer?
Sure, that's no problem - www.at-horns.eu/release/ST260-19-half-rotated.zip

Lastly - and I know this is going to sound ridiculous - but I'm going to add that even the frequency sweeps used to measure the horn sounded quite pleasing.
That's why we do it - to sound pleasing during the measurements. Who listens to the finished stuff anyway? :D
 
I've been simulating these:

  1. Dayton ND25FN-4
  2. SBAcoustics SB19ST-C000-4
  3. I'd like to try the Peerless DA25BG08-06, but I haven't found accurate measurements of the dome/surround yet.

I'm aware there are better options at higher price-points, but I'm trying to maximize the performance of cheap/accessible drivers.

So far it's been pretty easy to get good results from 1k-14kHz, but the second half of the top octave is tough. (There's an argument to be made for ignoring that region, but I'd like to push the envelope with sims before I spend any money)
 
Mabat - one question:

I'm trying to investigate the impact of the Rollback.Angle parameter in the freestanding CircSym mode, but no matter what value I use the end of the rollback looks discontinuous. Any ideas?

Code:
Rollback = 1
Rollback.StartAt = 0.6
Rollback.Angle = 235
Mesh.AngularSegments = 8
Mesh.LengthSegments = 100
Mesh.SubdomainSlices=
Mesh.RearShape = 1
Mesh.WallThickness = 5.0; [mm]
 

Attachments

  • rollback_angle.PNG
    rollback_angle.PNG
    17.9 KB · Views: 366
Lastly - and I know this is going to sound ridiculous - but I'm going to add that even the frequency sweeps used to measure the horn sounded quite pleasing. I hate taking the measurements, especially of compression drivers. But I've taken measurements of the drivers on QSC clone waveguide, SEOS waveguide, and the Wayne Pi waveguides - and this horn was the most bearable of them all. Anyways ---- let's get back to objective data :).

I've probably done a thousand measurements of various projects. And the moment I hear that pink noise sweep, I know if it's good or not.

It's such a bummer when I work on something all weekend, do that first measurement in REW, and before the measurement even appears on screen, you can tell if it's good just by the sound of the sweep.
 
Thanks again. I think that better would be to normalize to 10°. If you look at the sim (#4969), just above 5 kHz there is a slight ripple on-axis. If normalized to 0° this ripple will show in all the other polars instead - that's not describing the overall behaviour very well. I know it's always difficult with normalization. Maybe you could simply show the raw data as well.
 
Normalized to 10 and raw

Sure, Matab.

See additional pics. Now normalized to 10 degrees, and raw.

Again - the two sides have not yet be glued - just taped together.
 

Attachments

  • Polar - 10 deg Ref.jpg
    Polar - 10 deg Ref.jpg
    193 KB · Views: 425
  • Polar - 10 deg Ref - Raw.jpg
    Polar - 10 deg Ref - Raw.jpg
    212.3 KB · Views: 476
  • unnamed (4).jpg
    unnamed (4).jpg
    172.2 KB · Views: 506
  • unnamed (5).jpg
    unnamed (5).jpg
    205.6 KB · Views: 478