3D printed waveguides

[TN Allen]

"...a modeled result is always second to the real world"

Important point, and good to keep in mind while evaluating response curves. Although response curves are the result of real world testing, they are also only indicators, based upon narrowly defined objectives, which are, at best, only a substitute for real world listening.

Manufacturers need to mass produce WGs that produce sound that appeals to a broad array of listeners. Individuals however, can fashion audio systems to suit their own preferences. We have vast resources at our disposal, and need not comply with one size fits all marketing objectives.

Meeting design objectives developed for mass marketing seems to me to circumscribe our choices for developing audio systems that appeal to our listening preferences, even if outside the conventional "box".

[Ro808]

Agreed Tim.
On simulated models vs. measurements: measured performance is in many cases better, or not so bad, compared to sims. This may be caused by resonances / interferences being damped in the actual driver or any other deviation.

[TN Allen]

What we need is a methodical way to evaluate how a set of speakers sound, that is well integrated in a helical design process, that also includes FEA, accurate milling and turning, and response curves.

The methodical listening part is probably the most difficult to define, but I suspect there are several people on this forum with good ideas for doing this.

[TN Allen]

Brandon and Ro808,

Have you found, or worked out relationships between guide depth, mouth radius or circumference and cross section, to frequency response?

[augerpro]

Mouth diameter governs the lowest frequency of pattern control. Depth mostly impacts the degree of directivity or "focus" of the energy.

[Ro808]

Even though we're not stricly talking horns here, the basic rules from horn theory still apply.
This quote explains things better than I could:

Typically, a loudspeaker includes a driving unit that
is coupled to a horn. The large end of the horn, called the
"mouth," typically has an area large enough to radiate sound
efficiently at a desired low frequency. The small end of the
horn, called the "throat," has an area selected to match the
acoustic impedance and exit diameter of the driving unit and
to reduce distortion of the acoustic signal.
The loudspeaker horn guides the acoustic signal or
acoustic energy into particular directions or regions. The
loudspeaker horn surfaces that constrain and control the
radiation of acoustic energy are commonly referred to as an
acoustic waveguide. The surfaces of an acoustic waveguide in
a loudspeaker typically produce a coverage pattern of a specified
total coverage angle that may differ horizontally and
vertically. The coverage angle is a total angle in any plane of
observation (although typically horizontal and vertical
orthogonal planes are used). The coverage angle is evaluated
as a function of frequency and is defined to be the angle at
which the intensity of sound, or sound pressure level (SPL ), is
half of the SPL on the axis (the reference axial direction is
usually normal to the throat of the driver).

A horn for use with a loudspeaker may include an
entrance disposed at a first axial end of the horn and configured
to receive a driver. A mouth may be disposed at a second
axial end of the horn opposite the entrance. A contoured
surface may extend between the entrance and the mouth.
A cross sectional shape of a coverage pattern of
audible sound emitted by the loudspeaker coupled with the
horn may be independent of a shape of the entrance and a
shape of the mouth.

[TN Allen]

In practical terms, and just as a reference example; based upon what you know of WG design, what effect would you predict for a guide 8" in dia. X 2" deep, with 1.475" throat diameter fitted to an RS 28 tweeter at the surround, rather than fitted to the standard face plate. The cross section radius would nominally be 4", but might vary plus or minus .1".

[natehansen66]

It's been awhile since I've had much to do on this forum but I check in on this thread now and then. Right up my alley! I've recently been getting (back) into ABEC3. It's made by the same guys that put out AxiDriver and Akabak. It's a BEM acoustic modeling software. It should be excellent for designing waveguides....which is my intended use........and allows you to model the diaphragm and enclosure/waveguide in cad to be used for the solver. I'm just working with a simple woofer/sealed box model but hope to be modeling a wg I have built in the coming months to see how the results compare. A full featured student version can be had for free with the caveat that it not be used for anything commercial.

[augerpro]

I'll have to check that software out, thanks Nate!

[natehansen66]

It's fairly complex of course. You need to turn the cad model into a mesh with different sw, then scripts are needed to point the solver in the right direction. I can point you to a few threads with some instructions of you want. I'd be willing to have a go at modeling existing designs if you have step files. May be a while before I'm capable of that though.

[augerpro]

Nate are you familiar with Fusion 360? I can export as .step but that is the whole model. If I can export as the native .f3d file you can manipulate it all you want (turn off the mounting flanges for example) to make it easier to import into ABEC3.

[natehansen66]

I've been learning my way around it....actually working on a wg model in 360 right now. Can you upload here or do you want my email?

[augerpro]

Here is the .f3d for the elliptical C design from post #303. You can turn off the phase shield if that would be easier, I have measurements with and without it, so doesn't matter for comparison.

https://www.dropbox.com/s/tuwtg53eyxd116h/5x.75%20C%20v5.f3d?dl=0

[Ro808]

Theoretically, it should be possible to predict (power) response, D(irectivity)I(ndex), horn cut-off frequency etc. based on the mathematics for a specific horn type (conical, exponential, tractrix etc.) and one or more derived target curves.
However, theory for these classic horns is based on the (supposedly plane) wavefront produced by compression drivers, which is fundamentally different from dome/ringradiator tweeters.

From studying research material and other sources of information it becomes clear waveguides for dome tweeters are developed through an iterative design, development and prototyping process in which FEA optimization is a key element.

Previous experiments with waveguides on this and other forums have shown nice results. I think, a logical next step has to be some kind of method in which BEM/FEA is included.

The following text is from a research paper on the "Development of a method to optimise the geometry of a horn to give a specified smooth beam width":

"The geometry of the horn is parameterised and the source superposition technique used to calculate the beamwidth.
An investigation is made of a geometrically simple hom profile consisting of an essentially conical horn with a radiused entry at the horn throat and a radiused flare at the horn mouth. The ability of this geometiy to achieve the desired nominal beamwidth is investigated, as is the the effect of throat radius on the performance of the system.
More complicated geometry parameterisations are investigated and a Bezier spline based geometry is found to be flexible enough to define a shape that approaches constant beamwidth behaviour, although it may not be able to find a desired nominal beamwidth. This geometry parameterisation is then solved repeatedly for a wide range of lengths and throat dimensions, and a method developed to enable an optimum design to be quickly found."

[Ro808]

In my view, these 2 parts from my previous posts have particular significance for tweeter waveguides:

"The small end of the horn, called the "throat," has an area selected to match the
acoustic impedance and exit diameter of the driving unit and to reduce distortion of the acoustic signal."

"A geometrically simple hom profile consisting of an essentially conical horn with a radiused entry at the horn throat and a radiused flare at the horn mouth."


The latter, of course, doesn't rule out other geometries, but it's probably easier to start with a simple profile.