Bracing material?

[Baseballbat]

Will you show any data to proof that the theory isn't comprehensive enough?

I say that we're in the linear part in the lower left of the diagram. We're talking about excursions in the range of a few hundred micrometers, and you keep telling me that this is non-linear?

 

[Ed LaFontaine]

curveball

I see this discussion exists in a rectilinear world.

I propose that diagonal bracing is more effective per unit installed than parallel & perpendicular.

 

[Charles Darwin]

We are talking resonances here, yes?

Aren't resonances by their very nature non-linear?
As far as I understood them the moment the energy input exceeds the systems internal damping the output at the resonant frequency skyrockets.
Basically we need a repetitive energy input impulse, this travels through the system taking a finite time and if this time taken equals the frequency of the input it adds in phase at the output thus adding to the output amplitude.

Or something like that anyway…

 

[Keriwena]

Will you show any data to proof that the theory isn't comprehensive enough?

I tested my current cab yesterday over a range of 24dB, and didn't get more than +- 1dB difference, which I didn't feel was conclusive. I guess it's braced pretty well. When I'm feeling better, I'll drag one of the older prototypes out of the 'boneyard' in the corner of the kitchen, and see what I get. (The howling cab was disassembled for parts, so I can't test that.)

I propose that diagonal bracing is more effective per unit installed than parallel & perpendicular.

This is something else I've been meaning to test. There's a contingent that suggests bracing the long way, which reduces one unbraced dimension to a minimum. I can't decide if they're on to something, or if it's simply that a longer brace uses more bracing material. Diagonal is an interesting idea, as it would be the longest brace possible, yet it also maximises the shorter dimension.

(Myself, I use a non-parallel, quasi-random eggcrate thing with 2"x1/4" oak slats on edge. Probably overkill, but effective. A while back, someone posted a link to a guy in France that does something similar.)

 

[DrBoar]

JRKO
Direct copling to the driver is not the only mode of drive resonances in the cabinet. Putting a small loudspeaker inside a larger box shows that there is quite a lot of vibrations in the (outer) cabinet.

That being said I do agree on your point to sink the drivers vibrations away from the faceplate. I think in the 80 Linn had a speaker with the driver grounded to the integral stand and the box "floating" In my next large speaker I intend to have braces close to the magnet of the bassdriver and then som felt or other compressed resistive material that will help to off load the vibration from the front plate.

 

[Ed LaFontaine]

This is something else I've been meaning to test. There's a contingent that suggests bracing the long way, which reduces one unbraced dimension to a minimum. I can't decide if they're on to something, or if it's simply that a longer brace uses more bracing material. Diagonal is an interesting idea, as it would be the longest brace possible, yet it also maximises the shorter dimension.

I wish I could have broken it down so well. I'm going on intuition, until shown otherwise. I would add there is a greater distance between the centroid of each sub-divided area when the area begins as a rectangle. Any resonant energy has further to travel...and subject to more resistive losses...to excite an adjacent area.

The braces I use are hardwood, ~square on each end and about 7/8" x 2" deep @ the mid-point. I hold them in place with pin nails until the glue sets.

 

[mondogenerator]

i totally agree that diagonal bracing, or randomised bracing are good approaches. I discussed a fanned bracing arrangment (think Martin acoustic guitar internals)of bracing which all meet at the magnet of the driver, i think it was me and dave P10 in a similar thread, a while back. The concept would seem to be the best way to dissipate the energy of baffle and driver inertia over all panels equally. In 3D it would be difficult to realise, but i think the method is efficient. At least in theory, i never quit got around to trying it. However that non linear nature of resonance would mean any benefit gradually diminishes, since the 'damping' in excitation remains more static (structure changes may have less effect on inherent panel damping than the spectra and nature of it). Trouble is, since then and after much thought, i cant help but think it may produce the opposite result. In coupling that energy to all panels evenly, the box as a whole may become more efficient at radiating at least some of the resonances. That was when i got confused and gave up. Yet another task on my to-do lisj that remains outstanding.