ISO140-4 TESTING

[Scotty]

Hi All

I was wondering if anyone could shed some light on something for me. I
have done numerous ISO140-4 Sound Insulation Tests on paritions. The
standard discusses a procedure which should be implemented if the
background noise in the receiving room is less than 10 dB below the
source and backgrond combined (measured in receiving room with source
running in the source room).

Where the background noise is less that or equal to 6dB below the
source and background, the standard says (Clause 6.6 last paragraph)
the following:

"If the difference in levels is is less than or equal to 6dB in any of
the frequency bands, use the correction 1.3dB corresponding to a
difference of 6dB."

Now I have always thought that this meant that you should subtract
1.3dB from the source/background combined measurement value at the
frequency which is less than or equal to a 6dB difference. I have
recently had this approach questioned. As the wording within the
standard is not particularly clear, can anyone shed light on whether
this is the correct approach?

Also, if the background measured is actually higher that the source/
background combined, is 1.3 dB still subtracted from the source value?

Any help would be greatly appreciated.

Regards

Scotty

[Answerman]

The group is now sci.physics.acoustics. Google is not yet carrying the
new group, but it is carried by eternal-september. You can sign up for a
free account at:
http://www.eternal-september.org/

Post by Scotty
Hi All
I was wondering if anyone could shed some light on something for me. I
have done numerous ISO140-4 Sound Insulation Tests on paritions. The
standard discusses a procedure which should be implemented if the
background noise in the receiving room is less than 10 dB below the
source and backgrond combined (measured in receiving room with source
running in the source room).
Where the background noise is less that or equal to 6dB below the
source and background, the standard says (Clause 6.6 last paragraph)
"If the difference in levels is is less than or equal to 6dB in any of
the frequency bands, use the correction 1.3dB corresponding to a
difference of 6dB."
Now I have always thought that this meant that you should subtract
1.3dB from the source/background combined measurement value at the
frequency which is less than or equal to a 6dB difference. I have
recently had this approach questioned. As the wording within the
standard is not particularly clear, can anyone shed light on whether
this is the correct approach?
Also, if the background measured is actually higher that the source/
background combined, is 1.3 dB still subtracted from the source value?
Any help would be greatly appreciated.
Regards
Scotty

[Angelo Campanella]

Post by Scotty
Hi All
I was wondering if anyone could shed some light on something for me. I
have done numerous ISO140-4 Sound Insulation Tests on paritions. The
standard discusses a procedure which should be implemented if the
background noise in the receiving room is less than 10 dB below the
source and backgrond combined (measured in receiving room with source
running in the source room).
Where the background noise is less that or equal to 6dB below the
source and background, the standard says (Clause 6.6 last paragraph)
"If the difference in levels is is less than or equal to 6dB in any of
the frequency bands, use the correction 1.3dB corresponding to a
difference of 6dB."
Now I have always thought that this meant that you should subtract
1.3dB from the source/background combined measurement value at the
frequency which is less than or equal to a 6dB difference. I have
recently had this approach questioned. As the wording within the
standard is not particularly clear, can anyone shed light on whether
this is the correct approach?
Also, if the background measured is actually higher that the source/
background combined, is 1.3 dB still subtracted from the source value?

The math is:

Room sound energy = transmitted source sound energy plus background sound
energy

10*log(10^(Lr/10)=10*log[10^(Lt/10)+10^(Lb/10)] Eq. (1)

Lr = room sound level in dB
Lt = Transmitted source sound level in dB
Lb = room background sound level in dB

Note that this is an energy relation, hence the decibel multiplier and
divisor are both 10 and not 20 (proof left to the student).

We want solve for transmitted source sound energy. Eq.(1) is interesting,
but we can't just swap terms around to get Lt. We have to decide both sides
by ten, then take the antilog, then we can swap terms around. Having done
that, the result is

transmitted source energy = Room sound energy minus background sound energy

10*log(10^(Lt/10)=10*log[10^(Lr/10)-10^(Lb/10)] Eq. (2)

Now we can study what occurs when Lr-Lb is exactly 6 dB. Let Lr = 56 dB and
Lb = 50 dB.

Then Lt = 10*log[400,000 - 100,000] = 10*log[300,000] = 54.77 dB.

You can see that the difference is 1.23 dB... someone gerenerusly rounded
that to 1.3...

(Note that if instead Lt was 56 dB, Lr would be
10*(log[10^(Lt/10)+10^(Lb/10)] = 57.0 dB...)

======================

Back to your question whether only 1.3 is allowed when background is
greater, the answer is "yes". Under this circumstance, we know that the
partition will not be represented by users of the result (architects,
manufacturers and developers) as being greater than it really is.

The reasoning for this goes as follows:

1- The reason that the 1.3 dB value becomes fixed is a matter of how
background sound can perturb the precision of measurement under a variety of
circumstances. If more signal is taken out (more that 1.3 dB), it
artificially increases the apparent partition sound isolation. But we know
that commonly, the background can be variable without our being aware of it.
for one thing it is out of our control. (If we could control it, we would
quell it immediately; then we are left with only the uncontrollable noises;
back to square 1.)

2- If the acceptable margin of the test sound over background is reduced
to, say, 3 dB, this artificial result is even more likely.

3- It then becomes a matter of brinksmanship; How many more test results
do you want to pass to save time and money at the risk of reduced confidence
in the results. Since it is desired by many if not all users that the
overall confidence limit be smaller than a decibel, one should not risk more
than a fraction of a dB on account of background uncertainty.

I may not have answered your question well... Suffice to say the choice
of 1.3 dB max correction is a matter of opinion, chiefly ISO TC43/SC2...
It's often used in the US as well. Some committees have limited this aspect
to be 10 dB above noise... to get the background correction uncertainty to
be much, much less than a dB...

Ange

www.campanellaacoustics.com