• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Differences in Klippel NFS Measurements Based on Enclosure Thickness and bracing

ascilab

Active Member
Audio Company
Forum Donor
Joined
May 30, 2024
Messages
175
Likes
1,455
01.png


I plan to compare measurement differences using the Klippel NFS based on wood thickness and the presence or absence of bracing.
Three different boxes will be used for the test. The drivers to be tested are the SB Acoustics SB17nbac and the test boxes are currently complete, and measurements will be taken using the Klippel NFS either today or tomorrow.
 
All else being equal, incl. box volume? I’m looking forward to the results.
 
All else being equal, incl. box volume? I’m looking forward to the results
The external size of the enclosures is the same for all. Enclosures made of 5/8" birch plywood are approximately 5-8% smaller compared to others. Additionally, in the case of enclosures with added bracing, the internal volume decreases by the amount of the bracing.
 
The difference between a 3/4" birch plywood enclosure and a 5/8" birch plywood enclosure is not significant, so the 5/8" birch plywood was excluded from the comparison.
I will provide a detailed comparison of the differences between enclosures made from the same thickness of birch plywood, with and without bracing.
The measurements were completed today, and I will organize and upload the data tomorrow.
 
Could this be fixed with an EQ?
I guess the question is rather: if we try to fix this with an EQ then the frequency curve might get fixed but what happens to the waterfall graph (resonances) for example?
 
Very cool experiment!

What's considered industry standard 'best' material? What I mean is not something exotic or very expensive, but something available that's considered best practice to use. I remember something about HDF, but it's just a vague memory and I'm very far from a subject matter expert.

How was the bracing design chosen? Is this design also typical or industry best practice? Or would it be better to model or test to find the resonant modes and design the bracing based on those results?

I understand that optimizing the bracing or varying material type was likely outside of this design of experiment. Mostly I'm just curious about an image of what next or next-next steps in this kind of design process looks like. :)
 
Could this be fixed with an EQ?
I guess the question is rather: if we try to fix this with an EQ then the frequency curve might get fixed but what happens to the waterfall graph (resonances) for example?
As it doesn't appear on all axes (and thus changes also the directivity indexes) equalisation wouldn't lead to an identical result as its "monodimensional".
 
I'd like to add something but don't wish (honestly) to risk spoiling the fun!

Plywood seems to have random voids in its construction, which can give varied results when tested for resonances. No idea if the best birch or marine grade ply gets around this, but a maker who should know a thing or two about 'birch ply' (cough) doesn't rate it so highly any more and prefers thin-wall MDF instead as it's apparently far more consistent - but of course this is when making batched of twenty or more pairs at any one time rather than a DIY single pair project.

Just sayin' ------ Any chance the OP could try a thin-wall MDF and see if it's better or not. It may still need some damping panels on it and maybe bracing too depending on the design, but maybe it can't be *safely* worked at home.

Good luck with wherever this research is taking you :)
 
Conventional wisdom is bracing matters much more than material thickness. You have provided evidence that supports this. I wouldn't make a cabinet over 0.20 cubic feet without bracing, so you're results don't change anything for me.

But, I would have been much more interested in seeing the 5/8" material WITH BRACING compared to the 3/4" material WITH BRACING. (And even 1/2" material WITH BRACING.) Because I do almost always use 3/4" material, but if you showed that the difference was insignificant from 1/2" or 5/8" material, THAT would change how I build some cabinets.
 
Exotic materials and super thick walls are not necessary for good results.

We use 16mm (~5/8") router grade MDF (~700 kg/m3) with significant / strategic bracing, and find this to work very well without any significant resonance problems, while keeping the weight down.

Note that our products are relatively compact, so there's no large surface area without bracing anywhere. We also add additional strengthening where needed. Some examples:

Both baffles on the 10D are double walled, so the two sides that have drivers are 32mm.
The Inkognito has an extra ring around the driver opening, so the area where the bolts are fastened are 32mm.
The Manta has a double baffle (32mm) on the lower part where the 12" driver is, and of course additional bracing in both upper and lower chamber.

How the cabinet is dampened is also important. All our speakers have felt covering all surfaces, while the subwoofers use open cell foam.

This approach allows high rigidity and limited resonances with relatively thin walls.

Inkognito as an example:
Here we see two out of four braces crossing the cabinet from all sides. There is also additional bracing spanning across all four corners (behind the foam in the middle of the picture), and finally an isolated amplifier chamber that provide additional stiffness on that side.

1717528847873.png


The Manta with cross bracings in both chambers. The ports are also covered with an internal "grille" (not visible in this picture) that are mounted to the walls across the ports, providing additional rigidity. Finally the rear is double walled to isolate the amplifier.
1717529372565.jpeg
 
Last edited:
SPL-Horizontal-without-Bracing.gif
SPL-Vertical-without-Bracing.gif

In the absence of bracing, the off-axis response does not show uniform attenuation, as illustrated in the graph above.
On the other hand, when bracing is present, these issues disappear, as shown in the graph below.

SPL Horizontal_bracing.png
SPL Vertical with Bracing.png



The frequency range where problems occur is also visible in the contour plot, as seen below.
Contour-Plot---Horizontal---without-Bracing.gif
Contour-Plot---Vertical---without-Bracing.gif



When bracing is present, these issues do not appear in the specified range.
Contour Plot - Horizontal - with Bracing.png
Contour Plot - Vertical - with Bracing.png



Also, this can be easily observed in the polar plot.
Without bracing, a dip occurs at 553Hz, showing attenuation of sound on-axis, with strong sound occurring on the sides where resonance happens.
Similarly, at 465Hz, where a peak is observed, the opposite issue is evident.
Polar-Plot-(normalized)---witout-Bracing.gif
Polar Plot (normalized) - witout Bracing - 1.png




When bracing is present, the graph for 465Hz and 553Hz shows a stable response, as seen below.

Polar Plot (normalized) with Bracing.png
Polar Plot (normalized) with Bracing - 1.png
 

Attachments

  • Polar Plot (normalized) with Bracing - 1.png
    Polar Plot (normalized) with Bracing - 1.png
    84 KB · Views: 38
View attachment 373239View attachment 373240
In the absence of bracing, the off-axis response does not show uniform attenuation, as illustrated in the graph above.
On the other hand, when bracing is present, these issues disappear, as shown in the graph below.

View attachment 373241View attachment 373242


The frequency range where problems occur is also visible in the contour plot, as seen below.
View attachment 373243View attachment 373244


When bracing is present, these issues do not appear in the specified range.
View attachment 373245View attachment 373246


Also, this can be easily observed in the polar plot.
Without bracing, a dip occurs at 553Hz, showing attenuation of sound on-axis, with strong sound occurring on the sides where resonance happens.
Similarly, at 465Hz, where a peak is observed, the opposite issue is evident.
View attachment 373247View attachment 373248



When bracing is present, the graph for 465Hz and 553Hz shows a stable response, as seen below.

View attachment 373250View attachment 373251
Excellent stuff. So this is a pretty clear indication (if there was any doubt) that EQ would not fix the issue!
 
I plan to compare measurement differences using the Klippel NFS based on wood thickness and the presence or absence of bracing.
Three different boxes will be used for the test. The drivers to be tested are the SB Acoustics SB17nbac and the test boxes are currently complete, and measurements will be taken using the Klippel NFS either today or tomorrow.
Great thread!

If you have harmonic distortion data from the measurements, that also would be interesting to see.
 
Excellent stuff. So this is a pretty clear indication (if there was any doubt) that EQ would not fix the issue!
That's correct. This issue cannot be fixed with EQ.
 
Back
Top Bottom