View Single Post
Old 29th January 2008
  #16
Gear Nut
 

AnAverageJoe

The problem with the U boats of Auralex is that they give no technical data.
An U boat is an elastic decoupler with a handy U shape to keep it in place.
Auralex gives no technical information whatsoever about dynamic stiffness making it not possible to calculate whatever.

Possible alternatives are CDM, Sylomer, Kinetics stuff and lots if others.

A good floating floor should be around 10 Hz MSM (mass-spring-mass) (lower is better) and not exceed around 16 Hz.

This value for +/- linear springs (the elastic U-boats or any alternative for that matter) is directly related with the deflection of the elastic material under static load.
Hence even without knowing the dynamic stiffness or the physical principles you can test and measure this yourself.

You have a good floating floor if your deflection is around 2 to 3 mm under static load.
Hence ANY elastic material can only be judged within a total construction taking top and floor mass into acount.
The same material can result in a bad or a perfect floor.

You can influence this deflection by in- or decreasing the number of U boats per surface unit, or by in- or decreasing the mass of the top floor.

The relation between MSM and deflection can be rather exact calculated from:

In metric

d = (15.8/fo)^2
fo = 15.8/sqrt(d)
where >>>
d = defection in mm
fo = MSM (mass-spring-mass resonance in Hz)

In Imperial

d = (3.13502/fo)^2
fo = 3.13502/sqrt(d)
where >>>
d = defection in inches
fo = MSM (mass-spring-mass resonance in Hz)

Hence you can easily see that the mass of this top floor, and the spread of the pads or U-boats is very defining.
The basic idea of ANY decoupler is to load them to the maximum technical allowed static load.
Underloading a decoupler will give the false feel of additional safety and more stability, but will cause loss in acoustic decoupling quality which should be avoided if possible.

A rough rule of thumb, in order not to overload such materials is that the deflection should not exceed ca 10% of the initial material thickness. But this can depends a bit on the type of material, the edge factor and so on. E.g. the standard sylomer types will be somewhat less, rubbers can be somewhat more. Best of course is that the supplier provides useful data to work with.
There are 2 main types: compressible decouplers (volume material decreases under load = foam or fibre like stuff) and deformable decouplers (the volume does not decreases or hardly, but the material deforms expanding sideways = rubber like stuff). This latter type doesn't work anymore when you prevent them to expand sideways under load.

Another good rule of thumb is making the top floor as heavy as possible.
That way the relative comparison of temperary loads as e.g. people become less defining.
The more spare (relatively) you have to foresee in order to cover these temperary loads the higher the MSM will become when these loads aren't present + are standard too high (since underloaded) on these spots where such loads aren't present.

I CERTAINLY should put mineral wool (fiberglass or rockwool with low density in that cavity since it does influence STC rather strongly (certainly with a light topfloor). And it also helps to diminish a drumskin effect of such a floor.
If hygro-thermal problems should be involved that's more a matter of the correct use of a good vapour barrier (dampscreen).
A further good improvement is the use of Green Glue which diminishes flanking extremely, makes that floor more dead (drumskin effect) and additionally increases TL somewhat.

About the John Sayers topic I agree with nosebleedaudio.
That topic arose strongly in function of typical drywall separations in a studio setup, where they meant that it has no sense to additionally strengten a link in a chain when the overall TL is defined by the TL of the walls anyhow.

Real high insulations need very low MSMs and can't be obtained with such applications, needing very large deflections.