Sound Advice
Presented by Randy Zimmerman
Introduction
Good acoustical design
– Comfortable and productive environments
Systems
– Comfort vs. energy efficiency
Proximity to occupants
– Air terminal units
– Air outlets
Designers must understand acoustical ratings in
order to write good specifications
2
What You Will Learn
 Sound power vs. sound pressure
 Sound power determination
 End reflection correction
 Sound criteria
 Determining NC ratings for catalog data
 Specifying in terms of NC
 Specifying maximum allowable sound power levels
 Sound paths
 Mock-up room testing
3
The Sound Room
 Products are tested in a qualified
reverberant chamber (per AHRI
220)
 Reverberant chambers are used for
quiet products
– Low absorption
– Low background
 The reverberant field eliminates all
directionality from a sound source
 Sound levels within the reverberant
field are equal at all points
4
The Comparison Method
Determine the sound power (Lw) by comparison to a
known reference sound source (RSS)
Measure the sound pressure (Lp) of the RSS in order
to determine the room attenuation
Lp = Lw – room attenuation
Lw = Lp + room attenuation
If we know that the RSS creates Lw = 80 dB in the first
octave band (63 Hz), but we read only Lp = 70 dB, we
know that we have 10 dB of room attenuation in that
octave band
Room attenuation is constant
All sound meters measure sound pressure (Lp)
5
The Test Procedure
Set-up any ductwork and equipment to be tested
Remove any unnecessary material from the test
chamber
Turn off equipment and close test chamber doors
Take background sound pressure level
Switch RSS on
Take RSS sound pressure level
Switch RSS off and set test conditions
Record sound pressure levels at various conditions by
changing flow rates, pressures, etc
6
The Decibel (dB)
Because of the great differences in energy (or
pressure) available, the log of the actual
value is used
Reference power is 10-12 watts
Reference pressure is 0.0002 microbars
dB is measured vs. frequency
An infinite number of frequencies, so they are
averaged into bands, typically called ‘Octave Bands’
7
Octave Bands
Octave bands are centered about
increasingly wider frequency ranges,
starting with 63 cycles/second (Hz)
Each band doubles in frequency
Bands are traditionally numbered,
in our industry, as shown
Octave Band Designations
Center Frequency
Band Designation
63
1
125
2
250
3
500
4
1000
5
2000
6
4000
7
8000
8
8
Octave Bands
Fan-powered products usually create their highest
sound levels in octave band 2 (125 Hz), but sometimes
octave band 3 (250 Hz)
Grilles, registers and diffusers create their highest
sound levels in octave bands 4 (500 Hz), 5 (1000 Hz) or
6 (2000 Hz)
Octave bands 4-6 are known as the speech
interference bands
It’s industry convention to report sound data for
octave bands 2-7 only
Sound room size and design can cause problems with
readings in octave bands 1 and 8
9
Decibel Addition Example
To add two decibel values:
80 dB
+ 74 dB
10
Decibel Addition Example
To add two decibel values:
80 dB
+ 74 dB
154 dB (Incorrect)
11
Decibel Addition Example
Correction To Be Added To
Higher Value (dB)
3
To add two decibel values:
80 dB
- 74 dB
= 6 dB
2.5
2
1.5
Difference in Values: 6 dB
1
From Chart: Add 1.0 dB
to higher Value
0.5
80 dB
+ 1 dB
0
0
2
4
6
8
10
Difference In Decibels Between Two
Values Being Added (dB)
81 dB
(Correct)
12
Good To Know
Any sound source 10 dB lower than background
level will not be heard
Add 3 dB (or 3 NC) to double a sound source
– Two NC40 terminal units over an office would
probably create an NC43 sound level
– Two NC20 diffusers in a room would create a worst
case sound level of NC23 (if they are close together)
– Don’t try to add-up dissimilar products in this manner
13
Sound Power Changes
Equation for sound power changes = 10logn
1 Fan on
1 Fan on
1 Fan on
vs. 2 Fans on
vs. 4 Fans on
vs. 10 Fans on
n=2
n=4
n=10
Add 3 dB
Add 6 dB
Add 10 dB
1 Fan on
vs. 100 Fans on n=100 Add 20 dB
50 Fans on
vs. 100 Fans on n=2
Add 3 dB
14
Proximity To Sound Sources
Would you really expect to hear 100 fans running at
the same time?
Properly selected diffusers shouldn’t be heard from
more than 10 feet away
Although there may be multiple diffusers in a space,
it’s unlikely that more than one or two are within 10
feet of an occupant
We would only expect to be able to hear a 10 foot
section of continuous linear diffuser from any single
location
15
For High Frequencies
1 dB
not noticeable
3 dB
just perceptible
5 dB
noticeable
10 dB
twice as loud
20 dB
four times as loud
16
For Low Frequencies
3 dB
noticeable
5 dB
twice as loud
10 dB
four times as loud
17
What We Hear
Our ears can be fooled by frequency
– Both tones sound equally loud
65 dB
40 dB
63 HZ
1000 HZ
A Difference of 25 dB
18
Acoustic Quality
Not too quiet
Not too loud
Not too
annoying
Not to be felt
Don’t destroy acoustic privacy
Avoid hearing damage
Don’t interfere with speech
No rumble, no hiss
No identifiable machinery sounds
No time modulation
No noticeable wall vibration
19
OCTAVE BAND LEVEL _ dB RE 0.0002 MICROBAR
NC Curves
80
70
NC-70
60
NC-60
50
NC-50
40
NC-40
30
20
NC-30
APPROXIMATE
THRESHOLD
OF HUMAN
HEARING
NC-20
10
63 125 250 500 1K 2K 4K 8K
MID - FREQUENCY, HZ
20
Typical NC Levels
Conference Rooms < NC30
Private offices < NC35
Open offices = NC40
Hallways, utility rooms, rest rooms < NC45
NC should match purpose of room
Difficult to achieve less than NC30
Select diffusers for NC20-25 (or less)
21
Sound Power Vs. Sound Pressure
Sound power (Lw) cannot be measured directly
Sound pressure (Lp) is measured with a very fast
pressure transducer (i.e. a microphone)
Calculate sound power (Lw) by correcting sound
pressure (Lp) readings in a reverberant chamber to a
known power source
– Reference Sound Source (RSS)
22
Reference Sound Source
Correction device for a
reverb room is the RSS (per
AHRI 250)
– Calibrated in an anechoic
chamber to simulate a free
field condition
– Used in a reverberant field,
so there is a known error
called the “Environmental
Effect”
23
In a Reverb Room
Sound power (Lw) is calculated from measured
sound pressure (Lp) and corrected for background
– Unless product sound is 10 dB above background
RSS is used to “calibrate” the room
Data is recorded per octave band (or 1/3 octave
band if pure tones are anticipated), for each
operating condition
24
Catalog Data
Sound pressure data is collected by a
frequency analyzer that samples microphones
via a multiplexer
Data is collected and sound power recorded
Spreadsheets are used to check the linearity
of data sets
Catalog data is prepared from actual sound
power data sheets using accepted regression
techniques
25
Diffuser Testing
Current test standard for
diffusers
– ASHRAE 70-2006
No significant changes in
many years
26
Terminal Unit Testing
Current test standard for
terminal units
– ASHRAE 130-2008
ASHRAE 130 is currently
under review
– SPC 130
– It will be updated to include
more products including
exhaust boxes
27
Sound Tests
Discharge sound, VAV terminals
– Unit mounted outside room
– Discharging into reverb room
Radiated sound, VAV terminals
– Unit mounted inside room
– Discharging outside reverb room
– All ductwork lagged to prevent
‘breakout’
Diffuser supply/return sound
– Unit mounted flush to inside the
reverb room wall
28
Performance Rating
Current rating standard for
terminal units
– AHRI 880-2011 (effective
Jan 1, 2012)
– Increases discharge sound
levels due to end reflection
– This affects all published
data and selection software
– The boxes will still sound
the same, but now the
acoustical consultants will
be happier
29
Sound Path Determination
Current standard for
estimating sound levels in
rooms
– AHRI 885-2008
– Provides sound path data
from ASHRAE research
– Attenuation factors for duct
lining, ceiling tiles, room
volume, elbows, flex duct,
etc
30
Industry Standardization
AHRI 885-2008 contains
Appendix E
– Recommends standard
attenuations to be used by
all manufacturers for
catalog data
– First presented in ARI 88598
– Makes comparing catalog
NC levels much less risky
31
AHRI 885-2008 Catalog Assumptions
Radiated Sound
Environmental Effect
Ceiling / Space Effect
Total dB Attenuation
Discharge Sound
Environmental Effect
Duct Lining
End Reflection
Flex Duct
Space Effect
Total dB Attenuation
2
2
16
18
2
2
3
9
6
5
25
3
1
18
19
Octave Band
4
5
0
0
20
26
20
26
6
0
31
31
7
0
36
36
3
1
6
5
10
6
28
Octave Band
4
5
0
0
12 25
2
0
18 20
7
8
30 53
6
0
29
0
21
9
59
7
0
18
0
12
10
40
mineral fiber tile
5/8 in thick
20 lb/ ft3 density
5 ft, 1 in fiberglass lining
8 in flex duct to diffuser
2500 ft3 room volume
5 ft from source
The following dB adjustments are used for the calculation of NC above 300 CFM
300 - 700 CFM
Over 700 CFM
2
2
4
Octave Band
3
4
5
6
1
1
-2
-5
3
2
-2
-7
7
-1
-1
32
Certified Performance Data
AHRI Program
– Directory of Certified
Product Performance
– www.ahrinet.org
– Random samples subjected
to annual third party lab
testing
– Verifies that performance is
within established test
tolerances
– Failures result in penalties
– Voluntary program
33
The dBA Scale
Used for outdoor noise evaluation
Also used for hearing conservation measurements
Basis of most non-terminal sound ratings
34
NC Specifying
Specifying and unqualified NC value is an
‘open’ specification
Specifying an NC with specific path
attenuation elements could result in
acceptable sound quality
It is far preferable to set maximum allowable
sound power levels than to specify NC
35
Example
Octave Band Level_ dB RE 0.0002 Microbar
80
NC rating given is NC-30
since this is highest point
tangent to an NC curve
70
NC-70
Sound Power
60
NC-60
Sound Power less 10 db
in each band
NC-50
50
40
NC-40
30
20
NC-30
Approximate
threshold
of human
hearing
NC-20
10
63 125 250 500 1K
2K
MID - Frequency, HZ
4K
8K
36
Example
90
NC rating given is NC-45
since this is highest point
tangent to an NC curve
NC-70
80
70
NC-60
60
Octave Band Level
dB RE 0.0002 Microbar
NC-50
50
40
NC-40
30
NC-30
20
NC-20
Approximate threshold
10
of human hearing
63
125
250
500
1K
2K
4K
MID - FREQUENCY, HZ
8K
37
Example
90
80
Both noise spectrums would be
rated NC-35, However, they would
subjectively be very different!
Typical grille noise
at a distance of 10FT
(high-frequency)
Typical fan noise from
adjacent mechanical
room (low-frequency)
Octave Band Level_ dB RE 0.0002 Microbar
NC-70
Approximate threshold
of human hearing
70
NC-60
60
50
NC-50
40
NC-40
30
NC-30
20
NC-20
10
Mid - Frequency, HZ
38
NC vs. RC
NC rates speech interference
and puts limits on loudness
NC gives no protection for low frequency
fan noise problems
NC stops at 63 Hz octave band
RC includes the 31.5 Hz and 16 Hz octave band
RC rates speech interference and defines
key elements of acoustical quality
39
Room Criteria
(RC) Curves
90
High probability that noise
induced vibration levels in
light wall and ceiling structures
will be noticeable. Rattling
of lightweight light fixtures,
doors and windows should
be anticipated.
Region B
Moderate probability that
noise-induced vibration will be
noticeable In lightweight light
fixtures, doors and windows.
Octave Band Sound Press. Level, dB
Region A
80
A
70
B
60
50
40
30
C
RC
50
45
40
35
20
Threshold
of audibility
10
ADAPTED FROM 2009 ASHRAE FUNDAMENTALS HANDBOOK - ATLANTA, GA
30
25
Octave Band Center Frequency, HZ
40
Two Parts of RC
Example – RC 40 N
The number is the speech interference level
The letter tells you speech quality
–
–
–
–
(N) = neutral spectrum
(R) = too much rumble
(H) = too much hiss
(V) = too much wall vibration
41
RC Number Calculation
Average of level of the noise in the
octave bands most important to speech
– 500Hz Octave band = 46 dB
– 1000Hz Octave band = 40 dB
– 2000 Hz Octave band = 34 dB
– RC = (46+40+34) / 3 = 40 dB
42
RC Letter Determination
Plot room sound pressure on RC chart
Determine rumble roof
– 5 dB greater then low frequency
Determine hiss roof
– 3 dB greater then high frequency
R - room sound pressure crosses rumble roof
H - room sound pressure crosses hiss roof
V - room sound pressure goes into vibration zone
N - room sound pressure does not cross
43
Rumbly
Spectrum (R)
80
Octave Band Sound Press. Level, dB
Measured data is outside the
reference region by >5 dB,
below the 500 Hz octave band,
therefore the noise is likely
to be interpreted as “rumbly”
90
PSIL=(38+35+29) / 3 = 34
70
60
50
40
30
20
10
RC-34(R)
Octave Band Center Frequency, HZ
44
Rumbly & Induced
Vibration (RV)
90
A
80
Even though the PSIL
Is only 33 dB, the
noise spectrum
falls within regions
A & B indicating a
high probability of
noise-induced
vibration in lights,
ceilings, air diffusers
and return air grilles
Octave Band Sound Press. Level, dB
70
B
60
50
40
30
20
10
PSIL= (38+32+29) / 3 = 33
RC-33(RV)
Octave Band Center Frequency, HZ
45
Neutral
Spectrum (N)
Measured data must
not lie outside the
reference region by
>3 dB, above the 1000 Hz
octave band
80
70
Octave Band Sound Press. Level, dB
Measured data must
not lie outside the
reference region by
>5 dB, below the 500 Hz
octave band
90
60
50
40
C
30
20
10
PSIL=(38+35+29) / 3 = 34
RC-34(N)
Octave Band Center Frequency, HZ
46
Hissy
Spectrum (H)
90
80
70
Octave Band Sound Press. Level, dB
Measured data is
outside the reference
region by >3 dB, above
the 1000 Hz octave band,
therefore the noise
is likely to be
interpreted as “hissy”
60
50
40
30
20
10
C
PSIL = (35+36+34) / 3 = 35
RC-35(H)
Octave Band Center Frequency, HZ
47
Who Uses RC?
NC remains the best way to make product
selections
RC is preferred as an analysis tool
Acoustical consultants will typically report
whether or not equipment meets NC spec but
will describe the resulting sound spectrum in
terms of RC
You should continue to see catalog
application data in terms of NC
48
Terminal Unit Installations
Sound characteristics
Optimal installation
Attenuators
Liners
49
Sound Characteristics
Radiated sound is primary issue with fan-powered
terminals
Discharge sound is primary issue with non-fan
terminals
Fan-powered sound is typically set in 2nd (125 Hz)
and 3rd (250 Hz) octave bands
– Long sound waves
– Harder to attenuate
Discharge sound is easily attenuated with lined
ductwork and flex duct
50
Ideal Terminal Unit Installation
>3D
Lined Sheet Metal Plenum
(Max velocity 1,000 FPM)
Max velocity 2,000 FPM
D
Maximize Height
Above Ceiling
VAV
UNIT
Flexible Connectors
For Fan-powered Units
Ceiling
4' Min.
Lined Flexible Ducts
To Diffusers
51
Attenuators
Single duct
– Equivalent to lined ductwork
Dual duct
– Provides a mixing area for unit, but not much
sound attenuation
Fan powered
– Lined elbow or “boot” may provide 2dB
attenuation by removing line of sight to motor
– Carefully engineered attenuators can provide
additional sound reductions
52
Liners
Different liners in single ducts do not affect discharge
sound much
– Unit is too short for the air to interact with liner
1" liner does not significantly
decrease sound compared to ½"
Foil faced liners add 6-8 dB
Fiber free adds 4-6 dB
Double wall is variable
– Kettle drum effect increases sound, but it is directional
53
Flex Duct
Don’t forget about flex duct
5' of flex can reduce mid frequencies
by 20 dB or more
Flex is better than lined duct or attenuators
in reducing low frequencies
You can have too much of a good thing
54
Diffuser Tests - ASHRAE Conditions
Measured Air Flow
10 equivalent Diameters, min
Pressure
Discharge Velocity
Sound
55
Inlets: 3 Equivalent Diameters - Ideal
~1 NC add to catalog data
Measured Air Flow
Flex Duct, 1 radius bend
3 equivalent Diameters
Pressure
Discharge Velocity
Sound
56
Inlets: Long 90 at Diffuser
~3 NC add to catalog data
Measured Air Flow
Flex Duct
Pressure
Discharge Velocity
Sound
57
Inlets: Hard 90 at Diffuser
~5 NC add to catalog data
Flex Duct
Measured Air Flow
Discharge Velocity
Sound
Pressure
58
Inlet ‘Kinked’
~7-9 NC add to catalog data
Flex Duct
Measured Air Flow
2 equivalent Diameters
Pressure
Discharge Velocity
Sound
59
Summary of Results
Minimum add for flex duct = 1 NC
Worst case add, ‘Kinked’ = 7-9 NC
Air distribution pattern can be greatly effected
– Plaque / Perforated shows most effect
– Multi-Cone / Louvered shows least effect
Results were not the same for all diffuser types
Don’t forget that catalog NC’s are based on
typical offices (-10 dB across all bands)
60
Some Diffuser Solutions
Locate balancing dampers at branch takeoff
Keep flexible duct bends as gentle as possible
– Flex duct is a great attenuator of upstream noise
sources
Keep duct velocities as low as possible
– But over-sizing can result in higher thermal loss
61
Additional Resources
Noise and Vibration Control
for HVAC Systems
– Mark Schaffer, 2005
ASHRAE Fundamentals
– Chapter 8, 2009 Edition
ASHRAE HVAC Applications
– Chapter 48, 2011 Edition
62
Summary
NC remains the preferred sound specification
RC is often used after-the-fact
Specified max sound power levels are safest
Lining materials affect sound levels
Careful selection, design and installation are
required to avoid problems
63