Acoustic Horn Design – The Easy Way (Ath4)

the mems i linked has its resonance @40khz and is very easy to use if you buy the linked evaluation kit for 25 euro, direct balanced line level output, and powered with 2AA batteries.
mems is moving very fast..

problem with a probe tube is that you get resonances in the tube, an acoustic lowpass filter due to the area reduction and big loss of sensitivity.

but be my guest to find out all these things for yourself.. 🙂
20230324_132704.jpg
 
Mabat, as someone working with with large scale additive i came across this post and I'm simply too curious: is this your waveguide?
This is a horn that I helped design together with Addit and Kantarion Sound.
They had a Celestion Axi driver that they wanted to make work in a club setting.
It uses a version of the ESP plug combined with an eliptical horn in order to minimize the on axis problems with axi symmetric horns.
The Axi has quite a drop in HF response when beaming and when you spread it - one has to high shelf quite a bit..
I have not heard it yet, but guys say it sounds nice. Will post measurements when I get them.

This is the ESP script:
; --------------------------- ESP related code ----------------------------
plug25 = { ; ESP definition
Dt = 50 ; input diameter [mm]
At = 6 ; input angle [deg]
Ae = 32 ; exit angle [deg]
L = 60 ; length [mm]
Pos0 = 0.42, 0.80 ; input positions of the vanes (matching axi plug)
; 3 rings: Pos0 = 0.500,0.707,0.866
; 4 rings: Pos0 = 0.446,0.632,0.774,0.894
Sk = 0.77 ; meander skew
CP1 = 0,0.33,0.25 ; control point positions
CP2 = 0,0.33,0.25 ; control point positions
EndAngle = 1 ; minimum tip angle [deg]
WT = 1 ; minimum wall thickness [mm]
}
Source.Contours = ::esp plug25 ; source definition



8CABF3F7-6C70-4602-9D21-6BD625973CFE.JPG
 
  • Like
Reactions: Naturlyd
I actually have hands on experience with the IM73A135V01 and I stand by my claim.
The HF is not consistent, at least for the application at hands if flat response is required, some might be great some won't be.
If the goal is just to have an idea about the distribution of pressure or reconstruct the iso pressure curves either (ECM or MEMS) should do.
Any orientation should work if sensitivity is not an issue it must however remain consistent.
Primo makes some great ECMs (I have no affiliation whatsoever with them).

FYI Mems are NOT perfect regardless of of what people think, at either end of the spectrum they are not as consistent as a good ECM.
You can easily spec a 4mm ECM 15Hz to 20kHz +/-2dB reference @1kHz at a reasonable price and MOQ,
you can't with a Mems unless you buy millions of them and pay for the rejects.
Mems advantages are size (only if you go for really small ones), ease of integration and SNR and AOP for reasonably small and cheap units.

Example of Mems (not the one referred to here) but similar results with them.
MEMS Data.png

And 4mm ECM (prototype batch)
ECM Data.png



If the ECM can be selected small enough then that is the best way, some are as small as 1.3mm.
Adding a tube does have some sensitivity issues but in this context I would give it a try with a 6/4mm ECM if already available,
the cost on top of an ECM is negligible and the EE nothing: a resistor a cap and an old school 9V battery is all you need.

If memory serves Etymotic probes extend beyond 10kHz and if sensitivity is not an issue (which should be the case here) way above that.
It is just the opposite of a MEMS which does peak at HF as opposed to decrease...


Example of measurement done with ECM, never had the need to check frequency above 10kHz
Driver Mode.png


EDIT:
Buy the Mems and ECM with soldered wires, these components are fragile and don't react well to too much heat
 
Last edited:
If I recall correctly, there are ways of making prob tube mics that use a second tub off the side to cancel the tube effects. You should look this up as I don't remember any details.

A mics opening orientation can become important at HF were the wavelength is close to the mics size, but below that it doesn't matter.
 
Now I try to figure out what's the problem with the conical extension I made above. It may not be so easy to actually make a smooth continuation to the segment inside a driver...
Here's my conclusion at this point. As the measured response (the axial cancelation) is strongly repeatable under all the conditions I've tried (even heavy misalignments, semi-open chamber, clay put around on the joint, etc.), I have no other explanation that this is simply how the sound wave develops inside the duct. If I try a longer extension, it will be different again... Only an ideally "clean" source would exhibit wavefronts independent of distance.
 
I think that this is all true. I would also think that it is driver dependent and not an inherent acoustical situation from the phase plug and continuation. The diaphragm is likely not a rigid piston at those frequencies. This would cause all kinds of odd effects, but ones that are very important in the >10 kHz range. Thank god this range is not so important.
 
So is it that these modes, so strongly observable near the throat, are then attenuated by an attached horn? ("dampened by radiation" as I've seen somewhere)
- Should I expect to measure lesser deviations, even in the same cut plane, when measured on a actual horn? (Which would seem to break causality but who am I to say.)
 
Last edited:
So is it that these modes, so strongly observable near the throat, are then attenuated by an attached horn? ("dampened by radiation" as I've seen somewhere)
That's not quite correct. Some portion of this driving source (which is what I interpret your measurements to reveal) will be transmitted as HOMs, depending on the cutoff frequencies of the modes. You can think of > 10 kHz as having several excitable modes, more and more with frequency. The modes efficiency peaks just above cutoff and falls from there, so it's like a resonant peak, which it is, a spatially. The mode cut-ins get ever higher in frequency, sharper slopes and peak levels with the order. How the driving mode shape interacts with these fundamental characters of wave propagation will be highly driver dependent. Modes do not "dampen" they radiate and diminishes as 1/r just like the main mode. So they should be audible. They also will not have any axial response changes as the axial response is always a measure of the source average, which is zero in all modes, except the main or zeroith one.
- Should I expect to measure lesser deviations, even in the same cut plane, when measured on a actual horn? (Which would seem to break causality but who am I to say.)
I would guess that if you are measuring before the wavefront has reflected from the plate, then the horn should have no effect. Its only when the entire waveform is looked at would we expect to see the horn.
 
Faital HF108 is known for its matching the horn simulations using idealized sources almost perfectly up to ~12 kHz.
This seems to confirm that (approx. at 0, 5, 10 mm) -

1679753131994.png


- It's quite shallow driver, so there's only a very short time window available...
 
Last edited: