Acoustic Horn Design – The Easy Way (Ath4)

[mabat]

I do. Use it as it is...
Come on, this is 260 mm across, let's keep some dignity and don't make it even smaller...

 

[jzagaja]

Or X shape so that 5" MTM isn't covered by waveguide.

 

[mabat]

Vertical BLD is really exemplary.

...but so is the plain old ST260: https://www.diyaudio.com/community/...-design-the-easy-way-ath4.338806/post-6498359

 

[sheeple]

Restraining vertical should be more important than shifting axis performance. And doesn't KVAR looks less resonant than BLD? BLD, however, has a wider horizontal pattern, which I think is not a bad feat.

 

[mabat]

Less resonant? I could show you a resonant device but would have to look somewhere else.

Perhaps the KVAR has just a little too much of directivity for its typical use case, at least that's my impression. This is not a night and day difference though.

 

[sheeple]

... whatever causes the more in 'wiggles' with the BLD plot.

 

[mabat]

Hard to say, that could be the response of the driver "blurred" by the overall higher diffraction level on KVAR. It's much less noticeable in the second set of measurements. These are already minute differences.

 

[sheeple]

Interesting conclusion. I would reason the other way round.

 

[mabat]

Maybe we don't even mean the same "wiggles"...

 

[mabat]

I made a quick hi-res simulation of the two (60 points between 1 - 5 kHz).
Maybe the BLD is really a little bit less smooth after all.

KVAR:

BLD:

Throat impedances (KVAR = blue):

 

[Dkalsi]

Vertical BLD is really exemplary. Let's make it horizontal instead.

Is the overall linearity of the waveguide-driver combo less important than the normalized polar pattern - with the idea that the driver+waveguide linearity can be addressed with passive/active crossover/dsp?

Based on the first glance, the linearity of the driver's frequency response is less impacted by the KVAR version than the BLD, but the BLD has a tighter polar pattern.

 

[mabat]

Is the overall linearity of the waveguide-driver combo less important than the normalized polar pattern - with the idea that the driver+waveguide linearity can be addressed with passive/active crossover/dsp?

Basically, yes. That said, it's of course always better if it's reasonably good enough to begin with.

(Beware of the word 'linear'. I understand you mean 'FR deviation from a flat line' but that's usually not what linear means when talking about systems, where 'linear' means that the behaviour is independent of level (obeys the superposition principle). Which is also a reason we don't have to pay much attention to a particular SPL here - the FR will still be the same, as long as the whole system stays linear, i.e. it's not driven too hard.)

 

[Dkalsi]

(Beware of the word 'linear'. I understand you mean 'FR deviation from a flat line' but that's usually not what linear means when talking about systems, where 'linear' means that the behaviour is independent of level (obeys the superposition principle).

Thank you for this!

 

[Dkalsi]

If normalized polars are helpful, see attached:

1) BLD - Horizontal - 0 - 70 - Normalized for On-axis (0 Degrees) - 1Khz - 10Khz
1) BLD - Vertical - 0 - 70 - Normalized for On-axis (0 Degrees) - 1Khz - 10Khz
1) KVAR - Horizontal - 0 - 70 - Normalized for On-axis (0 Degrees) - 1Khz - 10Khz
1) KVAR - Vertical - 0 - 70 - Normalized for On-axis (0 Degrees) - 1Khz - 10Khz

It's amazing how tight the radiation pattern of BLD-Vertical is (which, as @mabat pointed out, can be rotated to be used as horizontal)

 

[mabat]

Thanks, that's even better than it seemed from the previous graphs.

 

[mabat]

ST260-BLD, looking at the frontal hemisphere - spherical surface with radius 0.5m around the horn:

 

[mabat]

In the latest Ath beta, this is how to add such observation field:

ABEC.SphericalField:1 = {
Radius = 0.5 ; [m]
; Angle = 90 ; [deg]
; Density = 30 ; how many elements per 90 deg
}

 

[mabat]

Axial field -

ABEC.AxialField:1 = {
Inclination = 90 ; [deg]
Radius = 0.4 ; [m]
z0 = 0.0 ; [m]
z1 = 0.5 ; [m]
Resolution = 20.0 ; [mm]
}

 

[mabat]

BTW, this is for 12 kHz -

 

[mabat]

The same for the KVAR.