3D printed waveguides

[augerpro]

DanP> with the smaller 5"x1" waveguide you mentioned, how much did you investigate throat angle changes via changing the flare radius? How directive was it at say, 2.5khz?

Gonna go run and get my Transducer Labs tweets for measurement and I'll make another post to get this ball rolling.

[augerpro]

So let's kick this off and see where it goes, eh? Here are some design criteria for the first waveguide that I think would be meeting a real need:
1) designed to mate with a 12-15cm woofer between 2-3khz
2) looking at some examples of SS and SB woofers, 2.5-5dB spread from 0 to 60 degrees at 2.5khz, with the sweet spot probably being 3.5dB.
3) 5" diameter and 1" deep, but would like to investigate this to optimize to the -3.5dB @ 2.5khz dispersion, and time alignment needs (have to research that a bit and take some measurements)
4) only looking to have pattern control to no lower than 2khz. If we reach lower while meeting the dispersion and time alignment that's great, but not a priority (although LR2 alignments would be helped by pushing it)

Things to experiment with in this design:
1) round vs elliptical
2) throat angle (probably by changing curve radius)
3) relaxing the perfect match to baffle face to free up possible throat angles if the measurements show this to be more worthwhile
4) depth and width to optimize to the dispersion and time alignment requirement above

Here are the measurements for two tweeters I have on hand.

SB Acoustics SB26ADC-C0004
Although you lose the phase shield when removing the faceplate, I found the dome was glued right to the body, it's not replaceable. So no worries with misaligning the dome assembly when attaching the waveguide.
Diameter of surround 1.30"
Height of dome .12" (kind of had to eyeball that)
Distance of PS from dome not more than 1/8"
PS diameter .47"
4 equidistant mounting holes 2.04" across from each other

Transducer Labs N26C-A
Diameter of surround 1.37"
Faceplate cannot be removed, so you need to make room for the screws on it, I measure a slot .3" wide by at least .75" long being able to accommodate the screw heads. The slot would have to start by 1/16" from the edge of the throat

[DanP]

DanP> with the smaller 5"x1" waveguide you mentioned, how much did you investigate throat angle changes via changing the flare radius? How directive was it at say, 2.5khz?

Gonna go run and get my Transducer Labs tweets for measurement and I'll make another post to get this ball rolling.

I didn't investigate with the smaller guides, just the 8.5" guides, but I cant imagine the effect would be different on smaller guides, just higher frequencies would be affected. And to reiterate, I really don't think flare radius or angle will change anything besides a deviation from optimum. The sound wave is starting the size of the dome and expanding in a controlled way to the size of the mouth. It has the span of the depth to make this transition. Those are the parameters that will determine the maximum performance in the arenas we care about (directivity, boost, AC alignment and maybe a few others). If rhe wave is to be transformed in ONLY these ways, it should be disrupted as little as possible along the way.

There are only so many options for flare geometry and they really don't differ that much from each other. On one extreme, you have a straight profile (conical) with a small roundover at the mouth, which results in the most shallow angle possible at the throat. The other extreme starts at 90 degrees at the throat and makes whatever sharp bend necessary to get to the mouth edge in the small space available. Neither of these extreme flares is ideal due to wave expansion rates, excessive reflections, and disruptions due to rapidly changing radii. In the middle is a smooth flare that has the optimum expansion rate, minimal opportunity for reflections and no changing radii. Deviating from this "optimum" flare a little will yield little change from optimum. In other words, I don't think fudging the 180 mouth/baffle transition just to change the throat angle will do anything but deviate from optimum. It should ALL be as smooth as possible. Tangent is smooth. Constant radius is smooth. The throat/tweeter transition needs to be smooth. The initial throat angle shouldn't make any difference as long as it facilitates a smooth transition all the way to the baffle.

If you want more or less directivity, you just go larger or deeper and continue to let those parameters determine the optimum flare radius and throat angle. I'm not sure what the directivity of a given WG would be at a given frequency, but I have found that waveguides and woofers that are about the same diameter and depth match directivity quite well.

Dan

[DanP]

Things to experiment with in this design:
1) round vs elliptical
2) throat angle (probably by changing curve radius)
3) relaxing the perfect match to baffle face to free up possible throat angles if the measurements show this to be more worthwhile
4) depth and width to optimize to the dispersion and time alignment requirement above

Here are the measurements for two tweeters I have on hand.

SB Acoustics SB26ADC-C0004
Although you lose the phase shield when removing the faceplate, I found the dome was glued right to the body, it's not replaceable. So no worries with misaligning the dome assembly when attaching the waveguide.
Diameter of surround 1.30"
Height of dome .12" (kind of had to eyeball that)
Distance of PS from dome not more than 1/8"
PS diameter .47"
4 equidistant mounting holes 2.04" across from each other

Transducer Labs N26C-A
Diameter of surround 1.37"
Faceplate cannot be removed, so you need to make room for the screws on it, I measure a slot .3" wide by at least .75" long being able to accommodate the screw heads. The slot would have to start by 1/16" from the edge of the throat

So you want a 5" diameter x 1" deep waveguide for each of these tweeters? I can do the one for the SB tweeter from your description above, but I think I'll need a drawing because I'm not understanding the slot for the screw heads. Also, we'll need someone else to do the CAD on elliptical guides if you want to explore that type of WG. It's way too time consuming in my maladapted CAD system.

BTW if you've got tweeters laying around, I have WGs laying around that I won't ever use. I'm not sure what they'd fit, but you're welcome to them if you want.

Dan

[TN Allen]

I think I have an oval guide similar to that in the photo of the 3D printed guide with the 6" rule. The throat opening is 1.3". I milled several of these for the Vifa OC25... dome without a faceplate. It will need sanding, but could be an inexpensive place to start.,

Given a little more information, I'll do the CAD file for the small oval guide.

[augerpro]

Here are some pics that will help make my description clearer

[augerpro]

If someone has a known good RS28A they could loan out, I'd like to add it to the list. Also got my eye on the Tang Band 25-1983 with inverted dome, and Satori TW29RN.

[DanP]

Here are some pics that will help make my description clearer

How tall are the screw heads?

Dan

[augerpro]

So you want a 5" diameter x 1" deep waveguide for each of these tweeters?

Dan

Yeah let's start here. With diameter and depth set, how much freedom do you have with the curve radius (trying to visualize if there is room to play or if you are locked in)? What do you think of trying to incorporate a phase shield into the design? Also, for the prototypes it might be useful to make the mounting flange 1/8" thick only so they can be surface mounted and the edges taped over to minimize diffraction. I don't have a garage anymore so I can't fab individual flush mounted baffles like I used to.

[augerpro]

TNA> oval would be awesome! One question, you would use an ellipse or super ellipse, anyone know the properties of each and which would be optimal? I think in one of the Geddes papers in the dropbox link he mentions the when going from circle to ellipse you would use a mouth size that has an area equivalent to the circle mouth area. And at "slice" along the length of the horn flare also.

[augerpro]

How tall are the screw heads?

Dan

Hmmm, just guessing (do you need exact?) probably 3/32"

[augerpro]

Some food for thought, second one is Revel's newest design (although on a lower tier speaker):

[DanP]

Yeah let's start here. With diameter and depth set, how much freedom do you have with the curve radius (trying to visualize if there is room to play or if you are locked in)? What do you think of trying to incorporate a phase shield into the design? Also, for the prototypes it might be useful to make the mounting flange 1/8" thick only so they can be surface mounted and the edges taped over to minimize diffraction. I don't have a garage anymore so I can't fab individual flush mounted baffles like I used to.

Pretty much locked in on the radius unless you're purposely trying to introduce irregularities.

I can incorporate any phase shield you can describe to me, or any other throat geometry for that matter.

A 1/8" flange is no problem, but if you can think ahead to how you want to do this, we can add in features (on the back or edges) that could make mounting really easy. Even something as simple as keeping the same mounting hole pattern from WG to WG could make things easier. We can get rolling without figuring this part out, but keep it in the back of your head as you test the first few.

Dan

[TN Allen]

Regarding post #55, if I still have one of those guides, it is an ellipse, 4" X 6", and about 1 1/8" deep. I don't know what a super ellipse is, or what the "slice" is.

The cross sections will be arcs tangent to the mouth plane, and end at a circlular throat.

Much of the discussion and vocabulary regarding waveguides is ambiguous, in my opinion. I sometimes suspect more conventional geometric terms would be more accurate. What, for example, does oblate spheroid describe?

What sort of ideas are there for testing guides for their effect on listening? This is probably the real challenge. Conventional data from Holm or Omnimic may be useful in some ways, but isn't sufficient to explain or quantify why and what guides do to improve the experience of music through guides compared to flat baffles.

Just playing the Devils advocate.

[augerpro]

Regarding post #55, if I still have one of those guides, it is an ellipse, 4" X 6", and about 1 1/8" deep. I don't know what a super ellipse is, or what the "slice" is.

The cross sections will be arcs tangent to the mouth plane, and end at a circlular throat.

Much of the discussion and vocabulary regarding waveguides is ambiguous, in my opinion. I sometimes suspect more conventional geometric terms would be more accurate. What, for example, does oblate spheroid describe?

What sort of ideas are there for testing guides for their effect on listening? This is probably the real challenge. Conventional data from Holm or Omnimic may be useful in some ways, but isn't sufficient to explain or quantify why and what guides do to improve the experience of music through guides compared to flat baffles.

Just playing the Devils advocate.

From wikipeida: "A superellipse, also known as a Lamé curve after Gabriel Lamé, is a closed curve resembling the ellipse, retaining the geometric features of semi-major axis and semi-minor axis, and symmetry about them, but a different overall shape.
In the Cartesian coordinate system, the set of all points (x, y) on the curve satisfy the equation
{\displaystyle \left|{\frac {x}{a}}\right|^{n}\!+\left|{\frac {y}{b}}\right|^{n}\!=1,} where n, a and b are positive numbers, and the vertical bars |   | around a number indicate the absolute value of the number."


You are right that we need more conventional geometric terms, I just don't know what those are But I have a feel for the essential point. For example the reason an OS curve is used for low diffraction is because it is a catenary curve, from wiki again: "Mathematically, the catenary curve is the graph of the hyperbolic cosine function. The surface of revolution of the catenary curve, the catenoid, is a minimal surface, specifically a minimal surface of revolution." As I understand it, it has the minimum surface area connecting two circles, a "perfect" form if you will. This according to Geddes is the best form to prevent diffraction when transforming the plane wave into a spherical wave. So as we transform from a circle to ellipse there is probably some "perfect" or minimal form (geometric and mathematical) that does this. IIRC the Geddes paper mentioned using the function that preserves the surface area of a circle at any distance along the horn wall with an equivalent surface area ellipse. I think

As far as quantifying the data...well the subjective part has been covered in other forums, and is a pretty deep rabbit hole. It would probably derail this thread. Objectively we can look at a couple things. Back in '08 when I measured all those horns we primarily wanted no ripples in the frequency response and an impulse response that was clean and as close to the ideal as possible. Even dispersion throughout the passband is a goal, but this is never perfect, so a lot of it becomes subjective on what is good enough. But maybe I can be specific on what I personally will be looking for. First is that the dispersion matches the woofer at the crossover frequency. Second is that between 30-40 degrees off axis the response trends steadily down from that point, no humps in the 3-6khz area. If this all happens we can meet the "ideal" set out by Harman's perception studies where the Predicted In Room response of the highest rated speakers is a steady downward slope of @ 8dB from 100hz to 10000hz.