Site Links
Howdy, Stranger!
It looks like you're new here. If you want to get involved, click one of these buttons!
Quick Links
Categories
In this Discussion
Who's Online (5)
- DrewsBrews 12:30AM
- Eggguy 12:30AM
- jholtz 12:30AM
- rjj45 12:30AM
- Tom_S 12:30AM
Above what Frequency does Baffle Edge Diffraction become an issue? . . .
At what point does baffle edge diffraction become significant enough to do something about in the design & construction of a cabinet?
Lower frequencies, say below 300 Hz aren't really directional although baffle step is an issue but not the point of this question - or maybe it is part and parcel to it.
I have a baffle design formulating in my mind but need some real data/parameters to work with.
1). Is it the point where the frequency's wavelength becomes greater than the baffle's width or at some lesser value?
2). Does the shape of the driver's cone (Dome?) play an additional roll in determining this?
3). Does the use of a waveguide reduce or eliminate the need for baffle edge curvature?
IOW - at what point/situation do curved/arched baffle edges become a requirement?
TIA.
Comments
Answer to 1 can be answered by using a diffraction tool such as VituixCAD diffraction tool. Adjust driver size and baffle shape and see the difference it makes for yourself.
Answer to 2 is yes. To simplify, the radiation pattern, ie the directivity affects the amplitude of energy that hits the baffle edge relative to the forward facing energy. More directive translates to less diffraction impact. Size of the driver diaphragm has greater impact in this regard that the "shape".
Answer to 3 is reduce. Waveguides or horns make energy more forward facing, less energy at 90 degree angle relative to on-axis. Similar to #2, waveguides/horns change the radiation pattern.So reduced energy at baffle edge means less impact of diffraction. "Eliminate" only happens if no acoustic energy reaches the baffle edge.
"Requirement" is determined by the designer. Baffle shape, width, curves, facets, waveguides, driver sizes, crossover points, driver separation, all impact overall result. It's up to you to decide how much you care about each factor vs the visual look of the speaker and something that is practical to build with the tools you have available and effort you want to put in to chasing the rabbit towards perfection.
Sharp edge vs large roundover will be very clear in the speaker directivity index, while often having only small impact to overall power response. So the ultimate question can be "How important is DI to you?". Other problem for on-axis designers, is high impact of diffraction can lead to poor crossover design choices.
Thanks, dcibel - your answer to item 2 with respect to driver diaphragm size; I assume a smaller diameter driver will have a wider dispersion pattern than a larger one given the same applied frequencies, is this correct?
Yes, of course, assuming all other factors are equal. You can look at any driver datasheet that includes some off-axis information to see this, compare a small 3/4" tweeter to 25-30mm tweeter. Compare a 4" woofer to 8" woofer.
"The Edge" by Tolvan used to be the go-to reference for simulating diffraction. I don't know if VituixCAD does everything the Edge does. http://www.tolvan.com/index.php?page=/edge/edge.php
The old rule of thumb was that a small diameter roundover didn't really make a measurable difference until about 3/4" radius. I think more recent studies have suggested that even 3/4" doesn't do "that" much and it is really a 1" or greater radius that starts really making a difference. But most people don't have larger than 3/4" radius router bit, so 3/4" and 1/2" are the defaults based on typical woodworking tools.
Yes, it does and more! VCAD allows generation of complete off-axis directivity data, as well as the ability to apply the diffraction to an imported frequency response directly. Accuracy at high frequency is limited due to the simplified piston model, but very accurate at low frequency which is the primary intent. The software assumes the diaphragm is flat and perfect, so any breakup behaviour, waveguide effects, etc are not considered, but you can feed it a half space response like much manufacturer data and get a good idea of response on a real baffle with little effort.
With complete diffraction directivity data exported, it can then be loaded back into VCAD as a driver, then view the spatial data in your favourite method - line chart, polar map, or 3D waterfall .
I have played around with the Edge (Tolvan) but I don't understand the definition of "Edge Sources" nor "Speaker Source Density" as used in the "Baffle Designer" UI - can you be of assistance here, a4eaudio?
I must assume that no edge curvature is employed in the Tolvan Edge modelling app without knowing what those missing definitions are and the the app is limited to height and width of the baffle with square edges otherwise.
Additionally, the "Speaker Geometry" list - is it concerning the shape of the baffle or the driver placed upon it?
I've never figured the Edge out other than a basic baffle. I thought it allowed pretty intricate baffle shapes but I'm not sure, and that's the part that I don't know what VituixCAD does either, because I'm just using rectangular baffles with roundovers.
The Edge models both the driver and edge diffraction as a set of point sources. So source density controls how many point sources make up the simulated radiating area of the driver. Edge sources controls how many different point sources are placed along the diffracting edge to simulate the diffraction overall as seen at the mic position. The more numerous these sources the more computationally intensive is the model.
Speaker Geometry controls the shape of the radiating area you are modeling. Typically you want to choose circular since 99% of the time you are modeling a cone driver. But you could also model a line source or other shape.
There is another parameter, the number of corners, that can be manipulated to give you regular geometric shapes, e.g. triangle, square, pentagon, etc. Each time you change the number or corners you must click the Apply button to update the model. As you increase the number of corners the shape begins to look like a circle, and that it one useful way to model a circular baffle withing having to drag all the corner points into shape, which is how you create irregular baffle shapes.
So, the Edge diffraction point sources are like taking off-axis mic sweeps at various angles?
Based on my usage, VituixCAD seems to be the best edge diffraction modeling tool out there. It can model any shape from a circle to a standard rectangle. And it can model any roundover radius, but the radius must be continuous all the way around. VituixCAD cannot model facets, tapered facets, or tapered roundovers. ViltuixCAD cannot model partial roundover situations. The Edge, as I understand it, cannot model roundovers at all.
Excellent information ^ 4thtry.
IOW - there is no point in modelling for edge diffraction on anything more than a flat baffle. Yes?
IMHO all of these diffraction programs are only good to within 1-2dB. That is "in the ballpark" but not all that accurate and errors of that magnitude can be audible. I like to use them to do "what if" type modeling to see the trends, etc. Then I build and measure. Rinse and repeat.
OK, so - "Build and measure" (Charlie) seems to be the only way to determine the outcome - like when I put 1" foam pipe insulation (2" outside diameter) on every shelf, monitor and cabinet-edge to reduce all diffraction between the driver and my ears opened my eyes to how destructive diffraction really is to the audio quality.
Diffraction simulation can give you insight into the response behaviour for different baffle dimensions and driver locations, but the calculation has limited accuracy at high frequency. For real design work, real measurements should be used, where diffraction simulation only really comes into play at low frequency for near field response processing.
Other limitations include limitations of complex shapes and curves. VituixCAD for example will model a flat baffle of any shape, but roundover is assumed on all edges and of the same roundover on all edges. So complex facets or baffle shaping must rely on real world measurements for accuracy.
I think of baffle shape separately from edge diffraction. So without going down that rat hole, frequencies greater than the baffle size don't see the edge. Frequencies less than the baffle size get to see the edge and create a ripple in the response. Anything else on the baffle would also create a ripple in the response at the corresponding frequency / wavelength.
Driver's location on the baffle changes the diffraction pattern because of different distances to the edge. Driver's size and shape changes also change the diffraction.
I think Q3 is a trick question. On axis there's only minor changes in the edge diffraction similar to driver shape changes for a waveguide.
What we really need is some type of sticky, flexible foam type material that could be used to quickly fashion the baffle into multiple shapes for testing. Take your measurement, then quickly re-shape the baffle into another complex shape and re-test. I have tried using modeling clay and mortite caulking, but it takes alot of time to re-shape between tests. Cardboard is easy to cut to various shapes, but adding and adjusting roundovers and facets can be very time consuming.
Horn everything.
https://www.jfcomponents.com/
unfortunately ^ an impractical solution due to space constraints for most of us . . . but I do love the concept.
The Olson paper from the 50's gives a good starting point. I enjoy playing with the edge program. You can make some crazy shapes.
Rigid foam insulation is easy to cut and shape for trial and error testing. It's also light enough that blue painters tape will hold it in position.
David Eisan posted a bunch of comparisons of various roundovers here:
https://www.diyaudio.com/community/threads/roundover-vs-45-chamfer-vs-double-22-5-chamfer-edge-treatments-for-tweeter-diffraction.384218/post-6976148
Awesome and informative link ^, dcibel - thanks.
Post 118 from the link (dcibel ) --> https://www.diyaudio.com/community/threads/roundover-vs-45-chamfer-vs-double-22-5-chamfer-edge-treatments-for-tweeter-diffraction.384218/page-6
Another thing that I am thinking of trying is vinyl window screen material, re-shaped with loose polyfill to create custom baffle shapes. I'd start by building a thin tower enclosure with the midrange and tweeter mounted in the usual positions. Then I'd cut a small 1.25" hole in the middle of a large piece of screen material and glue the screen onto the tweeter with a small 1.25" diameter bead of silicone close to the tweeter dome. Then I could fashion custom wave guides and tapered roundovers by adjusting the amount of polyfill stuffed under the screen material. Everything would be held temporarily to the sides of the tower with bar clamps. Sounds kind of crazy, but it might work. The polyfill would have the added benefit of dampening high frequency reflections.
^ This sounds pretty cool if it works - I think the screen material wont be flexible enough though.
I had been looking at open cell foam sheeting covered with a layer of spandex material like yoga pants are made of and you know how form fitting that stuff is . . .
https://www.amazon.com/spandex-fabric/s?k=spandex+fabric
Place a layer of 3/8" foam over the baffle and another 1/2" on top of it with a stepped axial hole in the center then place the tweeter or mid over the hole using the perimeter to compress the foam/fabric against the baffle and seal by tightening the mounting screws . . .
You could even make a pair of "buns" with a tweeter in the center just for laughs.