3D printed waveguides

[augerpro]

EDIT 1/4/2019: current design downloads in .stl and .stp at http://techtalk.parts-express.com/fo...87#post1398387

Throwing out a feeler/challenge to the community here. Someday everyone will have a 3D printer and it would be nice to have some well developed waveguides for designers to use instead of having them all be one-offs by a small handful of guys, or hacking together something sub-optimum off the shelf. I'd like to see a handful of waveguides developed according to, for example, various common design goals: maybe a very shallow design that is just enough to time-align that LR2 build, and give some directivity control; a deeper, more theoretically "correct" design that really pushes the lower crossover point and DI; whatever else the community finds to be a common need. Between myself and some others here we can round up some good tweeters to start off with (I have TL ceramic tweeter, Scan 9130 HDS, and various SB Acoustics tweeters.)

So I'm looking for someone with a 3D printer, and someone that can do the CAD design. Someone who knows there way around horn design would be awesome. My part will be limited to final measurements, providing some drivers, and design input.

It just seems the commercial outfits really only started using waveguides after the DIY community really explored the topic 8-10 years ago, but we still don't have much that is really optimized and widely available. Meanwhile, Revel gets Scanspeak and SB Acoustics tweeters with waveguides and we get bumpkis!

[Wolf]

Pete Schumacher and DanP already did some major testing of this type of thing.

Wolf

[TN Allen]

There are several discussions from the past years on this forum, with test data and photos. Perhaps you've already looked these over. Dave Pellegrene also did testing and posted his data for various waveguides he made.

Here are a few observations based upon designing, milling, and listening to speakers using various waveguides:

Regarding 3D printing, it's a slow process for objects of any size, printing speed will probably improve though. Access to printers and CNC machines can be expensive, even in maker spaces.
CAD for waveguides can be fairly straight forward, although, when the contours involve double guides the process is more complex, and requires fairly expensive software, unless a person can acquire it as a student or educator. Even so, it takes a while and some work to learn to use it sufficiently well to do complex contours. Once the contours are designed, generating a tool path or 3D printer file is relatively easy.

Designing and comprehensively testing waveguides in ways that relate shape and test data would be great, and would be an important tool for further development, especially if combined with crossover and enclosure software. This is probably more complex than the testing usually done with Holm and Omnimic. I don't know how a person would set up to evaluate the effect of waveguides on what one hears. I suspect it would be even more difficult to factor in the subjective responses that may be important to evaluating waveguides, but including subjective responses could add useful data,. Perhaps it could be done methodically, based upon a universal set of characteristics listeners use to evaluate speakers in a given environment.

If anyone wants to discuss doing as augerpro suggests, perhaps I can contribute to the design, 3D printing and milling effort.

[Sydney]

...So I'm looking for someone with a 3D printer, and someone that can do the CAD design....

Wish I could help you there; I find the prospects of this form of additive fabrication exhilarating.
Google image searchs returns some interesting projects.
Filamentum makes a wood composite filament called Timberfill

https://3dprint.com/5188/akemake-first-3d-printed-wood-speaker/

I'd like to see integrated waveguide akin to the Genelec 8351A SAM active monitor.
Good luck

[augerpro]

Wolf> yeah I've used Dave's waveguides. I'm not sure if he's still selling though. I would like to raise the bar with even further optimized guides, and as I mentioned, hopefully more focused design goals so people can find the right guide for their project.

TN Allen> Can you use the same CAD file for both 3D printers and CNC machines for routing baffles? Is there such a thing as a 3D scanner? I read your post a couple years ago and it was inspiring for sure.

Sydney> Yeah Genelec, JBL... That Timberfill is freaking cool.

[Sydney]

... Is there such a thing as a 3D scanner?....

Yup - using lasers.
Even DIY
http://3dprinting.com/how-to/sardaus...nner-3d-print/

http://makezine.com/2015/01/15/5-diy-3d-scanners-to-watch/

http://www.instructables.com/id/Build-a-DIY-Desktop-3d-Scanner-With-Infinite-Resol/

[dcibel]

There's several 3D print shops that will make whatever you want if you supply the CAD file. One just opened up in my town, I also know of Shapeways.

There's also a 3D printer available at Monoprice for the low low price of USD$200 for anyone that wants to get started at home.

[neildavis]

I just got an email notice from ITEAD Studio about their 3D printing services. You send them the STF file and they do the printing: 3D Printing services. The prices on their circuit boards are very low, and the quality is excellent, so I'm guessing that this will be a good way to prototype waveguide designs. If anyone tries them, please report back on your experience.

[TN Allen]

The solid or surface model geometry, or part file, for a waveguide, can be used to generate a tool path for the CNC machine, and a .stl file for a MakerBot 3D printer. There may be other file formats for other 3D printers, I don't know, but have used a MakerBot using a .stl file. The tool path also requires "posting" for the Haas CNC machine. I think though for some CNC routers, this step and file may be somewhat invisible.

Do you have ideas for testing waveguides that would evaluate more of the effects of the guides, and use subjective evaluations as well?

[augerpro]

Do you have ideas for testing waveguides that would evaluate more of the effects of the guides, and use subjective evaluations as well?

Well, I have a pretty well sorted measurement setup, though having sold my house my big test baffle is gone. I would probably use a large cardboard baffle now. I can measure anything the community thinks is relevant. I did this back '08 or so with Geddes' waveguide (and the QSC which became quite popular after my testing), I don't remember what specifically we looked for (need to find the thread) but ripple in FR and impulse response behavior come to mind.

I've posted threads with subjective impressions before, and unless your initials are JB, they don't find much favor here. But I'm certainly open to it.

So to get the ball rolling, let me suggest two examples I'd like to see developed:

1) a shallow wg that would match the directivity of woofers 4-7" around 1.5khz
2) a more theoretically correct (i.e. OS contour) and deeper wg that could match a 7-8" woofer around 1khz
3) experiment with contours similar to JBL's newest stuff in the M2, 7 series, and 3 series

[ani_101]

I still haven't found any good or proper way of determining the waveguide geometry. Once I can find out a way of calculating the geometry, we can write programs for outputting the waveguide points, that can then be input into CAD programs like fusion 360 which cna hten be tweaked (add screw holes, etc) and out put a STL file that can then be printed.

Did not find any good source for calculating waveguide geometry.

[augerpro]

I still haven't found any good or proper way of determining the waveguide geometry. Once I can find out a way of calculating the geometry, we can write programs for outputting the waveguide points, that can then be input into CAD programs like fusion 360 which cna hten be tweaked (add screw holes, etc) and out put a STL file that can then be printed.

Did not find any good source for calculating waveguide geometry.

Off the top of my head: http://sound.westhost.com/articles/waveguides1.htm
http://www.diyaudio.com/forums/multi...aveguides.html
http://www.diyaudio.com/forums/multi...on-thread.html

For the shallow wg there might not be a real complicated contour. Scanspeak appears to get a surprising amount of dispersion control down to 2khz in their 9900 with a very simple (it appears) shallow cone profile.

[augerpro]

Another standard feature I'd like to see is a "containment" wall on the backside that can be filled to form a thick and strong waveguide. Will probably be cheaper and more effective to leave this to the end user than incorporate it into the design. That way they can use fiberglass, epoxy, high tech polymer, etc.

[Wolf]

Brandon, I didn't say anything about Dave Pellegrene. I said DanP and Pete Schumacher, and they did some exemplary work with what Dan printed.

Later,
Wolf

[augerpro]

Sorry Wolf, I saw DanP and thought Dave P since he did so many waveguides. Did Dan or Pete publish there profile?