Differences in Klippel NFS Measurements Based on Enclosure Thickness and bracing

[flowjm]

As an aside, is it possible to change the colour map to something more perceptually uniform (such as viridis) when using the Klippel NFS software?

 

[theREALdotnet]

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

View attachment 373243View attachment 373244

Very interesting, not sure I would have predicted that. Destructive interference between the sound emanating from the cabinet walls and the drivers?

 

[Dave Zan]

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?

It has already been noted that the polar plots show it can not be effectively EQed.
There is also another clue.
The "wiggle" shape of the frequency curve is typical of non-minimum-phase response, minimum-phase resonance would typically show as a simple bump.

The subject of minimum phase and non-minimum phase can be a bit nerdy, you will see it discussed in terms of the placement of Poles and Zeroes in the Z plane, and other explanations of no obvious use!
But the TLDR summary is that minimum phase system frequency response deviations are reversible, whereas the extra phase over the minimum (that makes a non-minimum system) means it can't be reversed.
There is a deep connection with the "Arrow of Time" (why time runs only one way, not reversible, despite the physical laws that are time symmetric), which even ties back to the initial conditions for creation of the universe, if you like to read about physics.

The fact that "time" is somehow involved means that the frequency response curve provides a clue but that the phase response curve is even better, hence Amir frequently refers to it to decide if a headphone response can be equalised. Obviously the directivity plot isn't much use there.

Anyway, that clue can help you decide whether to try to fix a ripple, if you don't happen to have a Klippel.

Best wishes
David

P.S.. And thanks to @ascilab for the data, nice work

 

[olieb]

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.

Super interesting!
These graphs show that there are resonances and these cannot be EQed.
If resonances show up in the FR then the effect is probably already quite pronounced. (The sound energy from the resonating parts is of the same magnitude as the direct sound.)
Can you show some waterfalls that would produce much more clearly what is going on in the time domain?
Maybe the range 400-600 Hz is not the only one with resonances.

 

[AsciLab]

Can you show some waterfalls that would produce much more clearly what is going on in the time domain?
Maybe the range 400-600 Hz is not the only one with resonances.

with Bracing

without bracing

 

[AsciLab]

with Bracing

without bracing

 

[Matias]

@amirm this is super interesting, front page material imo.

 

[Thomas_A]

Nice experiment. Are these unstuffed boxes? That resonances would appear the 500-600 Hz region fits with dimensions. Wonder how ply and MDF differ. I see vibrations in the 500-700 Hz for a small MDF enclosure but also non-linear behaviour.

Speaker enclosure vibrations - a few measurements with accelerometer

There are been quite a number of threads dealing with vibrations in speaker cabinets, both in relation to "speaker feet" and vibrations in enclosures themselves. I recently took out my ACH01 accelerometer to do some measurements, both speaker feet and enclosures. There was some initial...

www.audiosciencereview.com

19 mm MDF side panel wf plot (acceleration)

9+9 mm MDF with damping glue, wf plot (acceleration)

 

[ttr]

Very interesting!

I would love to see impedance plot comparison.

 

[somebodyelse]

It would be interesting to see how panel damping compares to bracing. Something like the common automotive butyl/aluminium combination.

 

[Paweł L]

It would be interesting to see how panel damping compares to bracing. Something like the common automotive butyl/aluminium combination.

I think combination of both is best to lower resonances and dissipate energy stored in panels. Long time Michael Dzurko of Audio Concepts, well known DIY company from Wisconsin during 1990s and early 2000s, built a few monitors with Focal and ScanSpeak 7" woofers, Jaguar being the most ambitious of his projects. The cabinets were very extensively braced top to bottom and side to side, something that you would only see in very expensive designs like B&W, Focal Grand Utopia or very, very serious DIY projects.
I used the same midwoofer from SB in two different enclosures. One regular 3/4" MDF, double thick baffle, squarish thing, the other one kind of a test box with sides made of 1.25" particle board, horizontally braced. All panels have vinyl flooring pieces glued with PL400. The baffle is double layered, with rubber foam gasket in between and outer baffle panel bolted to inner baffle, so I can change panels depending on what diameter of frame is used to try different midwoofer-tweeter combinations. Inside is layered with Roxul mineral wool and Dacron. I think the test box although ugly is slightly cleaner and maybe 'faster' sounding.
If I would make another 'over engineered' test box, I would like to try rubber damping on panels instead of the cheap and dirty vinyl flooring left overs. Maybe it would provide better panel damping and dissipate energy better.