golden ratio. I dont get it

[blistard]

The dutch speaker: Studio de Schop ML 1.1 was designed using the Golden Ratio and didn't use any damping. Sounded quite nice too.
http://forum.zelfbouwaudio.nl/viewtopic.php?f=4&t=11178

[Jmmlc]

Originally Posted by speaker dave

Of course. But a dimension ratio of 2 to 1 is clearly the worst possible choice if well spaced harmonics are desired.

You do understand that, right?

Hello Dave,


As you can see 1:1:2 gives very bad result (max interval between 2 following resonances >80Hz/m). But 1 : 1.5 : 2 gives not so bad result (around 30Hz/m).

For sure, while giving a good spread of the resonances frequencies, a n integer value for a ratio may produce some coincident resonances (but this doesn't affect so much the spread it has more influence on the intensity of the resonance).

I don't care in that graph if resonances overimpose one to the other. But for sure evryone should choose a ratio which is not an integer value but taken from the red line passing through the points :

1 : 1.3 : 1.5
1 : 1.5 : 2.0
1 : 1.9 : 2.5
1 : 2.3 : 3.0
1 : 2.5 : 3.5
1 : 3.0 : 4.0


So if you want replace my sentence :

"Good shapes that spread optimally the resonances frequencies seem to be :
1 : 1.3 : 1.5 ; 1 : 1.5 : 2.0, etc."

by :

"Good shapes that spread optimally the resonances frequencies are located on the red line passing through the points : 1 : 1.3 : 1.5 ; 1 : 1.5 : 2.0, etc. but you should avoid integer values of the ratios"

Best regards from Paris, France

Jean-Michel Le Cléac'h

[flg]

Originally Posted by speaker dave

re: Fibinacci...David S.

thanks Dave,

[speaker dave]

Originally Posted by Jmmlc

Hello Dave,


As you can see 1:1:2 gives very bad result (max interval between 2 following resonances >80Hz/m). But 1 : 1.5 : 2 gives not so bad result (around 30Hz/m).

For sure, while giving a good spread of the resonances frequencies, a n integer value for a ratio may produce some coincident resonances (but this doesn't affect so much the spread it has more influence on the intensity of the resonance).

I don't care in that graph if resonances overimpose one to the other. But for sure evryone should choose a ratio which is not an integer value but taken from the red line passing through the points :

Best regards from Paris, France

Jean-Michel Le Cléac'h

Just not following your logic here at all.

You seem to be looking at the spread in Hz between one set of modes and the next. Since moving the modes away from those of one dimension eventually gets you closer to the next higher mode, there are optimums as shown by your traced red line. Still, every discussion I have seen on the subject tries to get modes to fall away from the modes of each other axis.

The goal is clearly not to have any superimposing. (Which is why cubic rooms and integer multiple dimensions are always warned against.)

Your analysis looks at modes of dimension A vs those of dimension B. It then looks at B vs. C. Optimization of those ignores the A vs. C factor. It is a 2D solution to a 3D problem.

The lower modes will be most evenly spaced with the cube root of 2 approach. (1: 1.260 : 1.587)

The table below shows this. It uses those ratios and is normalized to a dimension with 100 Hz for the first mode of the shortest dimension. The first 10 modes are shown.


1 1.26 1.587

100 79 63
200 159 126
300 238 189
400 317 252
500 397 315
600 476 378
700 556 441
800 635 504
900 714 567
1000 794 630

Where even this approach breaks down is that the higher modes start to overlap. For example the 4th harmonic of the "1" dimension is about the same as the 5th harmonic of the "1.26" dimension. (400 vs. 397)

Still this advice is given for room design because the goals are even spacing at the low end and eventualy enough modes per Octave are achieved (modal density is a desirable thing).

In the box case that we are talking about, if it mattered, we could be assured that upper range modes are dealt with via even minor damping and equal spacing of the low modes (as assured here) would be the best result.

David S.

[puppet]

If damping the box removes resonance concerns ... why shouldn't the ratio(s) being discussed lend themselves to minimizing baffle diffraction? Isn't this a bigger deal?

[mondogenerator]

if you build like most, by sinking the baffle, then that baffle is already very very close to your internal ratio anyway. If you build to optimise baffle instead, you can make up the internal ratio to equal the baffle ratio, using damping... It would depend on the size of the box in question but there isnt much of a discrepancy to begin with.

[planet10]

Originally Posted by blistard

The dutch speaker: Studio de Schop ML 1.1 was designed using the Golden Ratio and didn't use any damping. Sounded quite nice too.
forum.zelfbouwaudio.nl • Toon onderwerp - Speakers vervangen uit "Studio de Schop ML1.3"

If the speaker pictured is the one you are talking about, it is not Golden Ratio.



dave

[phivates]

I propose 1:1.27:1.618.

[Keriwena]

Tell your average woodworker you want a cabinet 12" by 15 6/25" by 19 52/125", and he'll be giggling so hard you'll have to call 911.

[puppet]

I see it as 12" x 15-1/4" x 19-7/16" ... this ain't a swiss watch