Testing baffle edge treatments for tweeter diffraction.

[DaveFred]

Since I am starting a new speaker, I wanted to try a couple of different baffle edge treatments to see their effect on tweeter diffraction.

Tonight I finished the four blanks which are a double layer of 3/4" MDF 8"x15".

Tonight I did the boring basic stuff, routed the rebate to recess the tweeter, did the through hole and notched for the lead ears.

I am going to test the following,

- control which will be left square edged

- 7/8" roundover on three sides

- progressively wider 45º chamfer starting a 0" width at the bottom and ending 1" wide going across the top

- 30º chamfer that starts in the middle of where the woofer will be going to about 1.5" at the top and across

Tweeter will be an inexpensive SB Acoustics SB26ST-C000-5

More to follow....

 

[AllenB]

A great example of DIY

 

[Juhazi]

Excellent, some IRL testing and results! Some kind of box would be good to have for off-axis measurements to be realistic.

The Edge can't handle roundovers nor does it have a box

 

[AllenB]

About interpreting the results..

A sim could show two things. One, that the response is smooth on axis because the diffraction effects are spread. This may not be the good result that it appears to be. Sound is still delayed and displaced. The sound also is not smooth on other axes.

That the response on axis is smooth because the diffraction effects are sufficiently reduced. How this is done depends on the application. For example though, a roundover might be a part of the sim.

 

[Juhazi]

It is a well known phenomenom, that speaker's asymmetric postioning on the baffle can distribute peaks and dips well on-axis (axially), but off-axis will show more wiggles - and different left/right.

Roundovers or bevels will have a better effect, they attenuate/soften the secondary edge source.

Some simulators can handle roundovers, eg. Bagby's BDBS and kimmosto's VituixCAD/Diffraction simulator

Basic article | Baffle geometry
Edge Diffraction: Real World Examples | The Audio Annex

 

[DaveFred]

I cut the edge profiles on the three test baffles tonight.

- 7/8" roundover on three sides

- Progressive 45° chamfer from 0" at the bottom to 1" at the top and across

- 30° bevel starting at 0" from mid where the woofer will be to about 2" at the top and across.

The roundover was the easy one.

For the 45° chamfer, I just made a simple jig to hold the end of the board 7/8" up at the bottom of the baffle.

The 30° bevel jig was a little more work, but I have made enough furniture leg tapering jigs before and for this, just made a short length/tall version. It didn't quite land exactly where I wanted it to, but close enough for testing.

I hope to take measurements in the next few days.

 

[DaveFred]

Pics of large bevel...

 

[David Morison]

Woodwork looks great, am looking forward to the results too!

 

[DaveFred]

I decided to build a box behind each baffle so I could accurately test off axis tweeter response to each baffle.

Finished tonight, hope to test in the next few days...

David.

 

[YSDR]

Good to see that effort, but a mid/woofer in the baffle can change the results (slightly), and I think you want to use one, not just a tweeter alone.

Cheers!

 

[Juhazi]

More tinkering based on theory and my own trials:
- all these baffle types are popular and relatively easy to make
- baffles 3 and 4 have remarkably different flat area dimensions and thus should show gross difference to 1 and 2
- tweeter's band 1-25kHz is so high that the box will have minimal effect.
- a woofer in the baffle would make some minor ripples in response appr. 3-8kHz
- below baffle step limit (woofer range) sidewalls make continuence which most likely will change spl response at perhaps 30-90 deg angles, compared to simulation of a baffle without box (or measurement, which would be a dipole woofer then !)

Still impatiently waiting for measurements!

 

[bolserst]

...Still impatiently waiting for measurements!

While waiting for measurements you might find this set I took 10 yrs ago of interest. It compares a standard sharp-edged box with the Joseph Audio and Avalon Acoustics style diffraction treatments. The Joseph Audio was similar to DaveFred's test baffle #4. The Avalon style reduced the diffraction effects just a bit more then the Joseph Audio, with smoother off-axis response in the critical 1kHz - 3kHz range. I couldn't locate pics of the actual test boxes, but I attached images of the loudspeakers I was trying to test the shape of.

*Forgot to mention that the 3dB rise in the top octave of the raw measurements is not diffraction related, but from the the response peak in the uncalibrated mic.

 

[mark100]

Thank you for the shop photos showing jigs and techniques.
Great experimentation...looking fwd to your measurements!

 

[DaveFred]

Here are some results,

Here are the results of my tests on the four baffles at 0º, 15º, 30º, 45º and 60º.

A comparison of the four baffles at each angle,

 

[DaveFred]

And all Five angles on each baffle,

 

[DaveFred]

Oh, and here at the four baffles,

 

[jrh0516]

Interesting that off axis the differences are minor

 

[AllenB]

What remains is how to pull useful data out of this.

Second, and this is the most important, there is no reason to believe that a small visual aberration means that the perception is equally small. It may well be that small diffraction aberrations are far more audible that the FR differences would lead one to believe (my research shows that.) I think that it is important to note that diffraction effects are always delayed in time and that can make a huge difference in perception.(link)

 

[ScottG]

Oh, and here at the four baffles,

It seems like the order is off with 1st and 2nd (from left to right) relative to the order of the results.

 

[Juhazi]

Oh, and here at the four baffles,

Makes me want to see responses with 30 and 45 deg bevels on edges, like roundover.

Here sharp bevels don't seem to make smoothing, but only change dimensions of the flat area (baffle) and shift diffraction negative peak.