top of page

Waveguides for 3D printers and CNCs

11/16/2019 original post

11/19/2019 additional commentary


6/7/2021 added Scanspeak D3004/6600 and R2904/7000 waveguides

9/1/2021 added designs to the Scanspeak and Bliesma T25 pages

6/11/2022 added routing templates and updated designs for use with these templates

12/27/2023 added SB29SDAC and new designs for existing tweeters

Several years ago I started this project to design high performance waveguides for popular tweeters for DIY'ers since the manufacturers don't seem interested in giving us anything worthwhile. At the time 3D printing was not cheap and so I made these usable for CNC also. In just two short years the costs have dropped dramatically and I'm very optimistic that these designs may get widespread use. 


A lot of work was put into this terribly slow and iterative project. If you are interested in seeing all the permutations that lead to the current designs I would suggest reading the threads I started at diyaudio and PETT.  

A note about my measurement method. I use MLS measurements with a reflection free 7ms gate. I USE  NO SMOOTHING. I want to supply the best possible data to users, not wishful fantasies that fall short when put into use. So be wary when comparing these results to others. Mic distance is typically 1m. The earlier measurements were done with 48khz sampling resulting in a 24khz frequency response cutoff, and you may see some crazy spike artifacting right at the cutoff. Since then I've moved to 96khz sampling, so now measurements are good up to 48khz and better capture the breakups of hard domes. 

Much more data and commentary to come, but for now I wanted to at least get up the plots and CAD files. .stl files are for 3D printing, and .step for CNC machining. Variants with and without a phase shield are labeled as such. Typically a hard dome would require a phase shield, and soft domes do not. 

Here are a couple examples of my waveguides with typical crossover filtering in place, since newcomers may not know what to make of the raw, sloping responses of waveguides and horns. The primary advantage is seen in the off axis response. Note how it is a straight line sloping downward. If this tweeter had just the factory faceplate and so no pattern control, the response wouldn't be at all straight, you would see SPL near and an octave above the crossover to be basically the same as the on axis response, followed by a sometimes very large droop above this. This is what causes the abrupt power bulge in the power response of a typical tweeter and woofer crossover. As a result most of the off axis response is not a straight line, but a wavy one where just above crossover the downward sloping response suddenly jumps up in energy right in a very sensitive region for the ear. And we hear this off axis power response nearly as much as the on axis response, so you want the response to ideally be straight lines, that go from flat on axis, to gently sloping as you move off axis. This is what you get with waveguide, as seen below.


Another benefit is matching the acoustic center of the woofer, time aligning them and making for smooth responses that are phase aligned. I've designed the acoustic center depths to match the likely woofer for a a given guide: in general a 5" waveguide would have a .75" depth, matching typical 10-15cm woofers. The 6.5" waveguides are usually around 1.1" deep, about the AC depth of 15-19cm woofers. And lastly, 8" waveguides will be around 1.5" deep, matching the typical 22cm woofer.

I've added routing templates to the website to help woodworking novices. You will need 5/16" OD and 1" OD template guide bushings to use the templates. 

I will update this main page as I make progress with new designs and tweeters.

SB Acoustics SB26ADC/CDC

SB Acoustics SB26STAC

SB Acoustics SB21

SB Acoustics SB29SDAC

Satori TW29TX

Satori TW29BN


Dayton RST28

Bliesma T25B

Bliesma T34B

Scanspeak D3004/6640

ScanSpeak Revelator D2104/7120

Scanspeak D3004/6600

Scanspeak R2904/7000

Peerless/Vifa OT19NC00-04

bottom of page