HVAC-System Acoustics

July 29, 2008
Methods of reducing unwanted sound associated with mechanical systems in commercial buildings

Though intended to provide comfort, HVAC systems often are a major source of noise in buildings. While the primary focus of designers understandably is temperature and humidity control, the more experienced among them know the importance of keeping sound and vibration in check.

This article provides an overview of acoustics as it applies to HVAC-system design.


Sound levels are expressed in decibels (dB). Decibels cannot be added traditionally. For instance, 50 dB plus 50 dB plus 50 dB does not equal 150 dB. It equals 54.8 dB. To add decibels, one can go through the calculations of logarithmic addition or use Table 1, which provides reasonably accurate results.

TABLE 1. Simplified logarithmic addition.

Decibel values must be considered two at a time. Using the example of 50 dB plus 50 dB plus 50 dB:

1) The difference between 50 dB and 50 dB is 0 dB. Following Table 1, add 3 dB to the higher value:

50 dB + 3 dB = 53 dB

2) The difference between this result and the remaining 50 dB is 3 dB. Following Table 1, add 2 dB to the higher value:

53 dB + 2 dB = 55 dB

A difference of 3 dB barely is detectable by humans, while a difference of 10 dB is perceived as twice or half as loud. Therefore, a 60-dB source typically is perceived as being twice as loud as a 50-dB source of the same frequency.

When dealing with undesirable sound, a designer must consider (Figure 1):

FIGURE 1. Source, path, and receiver.

• The source of the sound. In HVAC systems, air-handling units are common culprits. Each of any other components contributing to noise should be considered a separate source.

• The path of the sound. A single sound can have many paths (Figure 2). Different paths present different problems and, thus, require different solutions. This article focuses mainly on the path associated with ductwork.

FIGURE 2. Paths of sound and vibration.

• The receiver of the sound. For this article, that is the person or persons occupying a building.


In general, the human ear is less sensitive to low- and high-frequency sounds than it is mid-frequency sounds. Weighting is a method of reducing the contribution of low- and high-frequency sounds to attain a value corresponding approximately to what the ear perceives. The most common method of weighting is performed using the "A" scale, which produces values in units of dBA.

Table 2 shows "A" weighting for the center frequencies of the eight octave bands typically associated with HVAC-system acoustics. The more negative the "A" adjustment, the less sensitive the human ear is to the corresponding frequency.

TABLE 2. A-weighting scale.

Where HVAC is concerned, "A" weighting more often is used for outdoor sound calculations than it is indoor sound calculations. For indoor sound calculations, Noise Criteria (NC) is the more common method.

The NC method is fairly straightforward. Sound-pressure levels at eight octave-band frequencies are plotted on a family of curves defining the maximum allowable sound-pressure level corresponding to a chosen design goal. An NC rating is determined by the lowest curve exceeding the value at each octave band.

The downside of the NC approach is that the shape of a sound curve is not evaluated as a whole, and lower frequencies (below 63 Hz) are not evaluated at all. Still, the NC method is the most widely used basis for sound criteria for indoor HVAC applications. For office buildings, NC 35 usually is acceptable (Table 3).

TABLE 3. Sound-pressure levels for NC 35.

Room Criteria (RC) Mark II is the method recommended by the American Society of Heating, Refrigerating and Air-Conditioning Engineers (ASHRAE). This is a relatively new method that evaluates an overall sound curve. Using this method can be difficult because information for the lowest octave band is not readily available from manufacturers. (To learn more, see Chapter 7 of 2005 ASHRAE Handbook--Fundamentals.1


Following are guidelines intended to reduce noise associated with HVAC systems:

  • Select the equipment with the lowest sound-pressure levels at operating conditions.

  • Locate air-handling equipment as far from sound-sensitive areas as possible. Avoid placing rooftop units above occupied spaces.

  • Choose locations allowing, for rectangular duct, straight runs of duct for distances 1.5 times the largest duct dimension or, for round duct, 1.5 times the duct diameter before elbows or other fittings are used.

  • Use radiused elbows instead of mitered elbows. If radiused elbows will not fit, use mitered elbows with turning vanes. The more efficient the fittings, the more efficient the system and the less the noise caused by air flowing through the fittings.

  • Reduce air velocity by using larger ductwork, particularly near terminal outlets.

  • Do not angle offsets more than 15 degrees. If a greater transition is required, consider using elbows, as described above.

  • Reducers and enlargers should not exceed 45 degrees for converging flow or 60 degrees for diverging flow.

  • Locate volume dampers and other noise-generating devices as far from terminal outlets as possible. Avoid diffusers that utilize an air scoop. Choose terminal devices with low levels of generated noise.

  • Use flexible connections to equipment. Make certain that flexible electrical conduit has a small amount of “droop,” as tight flexible electrical conduit can transmit sound nearly as effectively as metal conduit.


The first step in designing a system for maximum acoustic performance is determining an acceptable weighting system. Once a weighting system is selected, the next step is deciding what levels are acceptable — for example, NC 35 for offices and NC 45 for lobbies. Different areas can have different requirements.

Evaluate the system as designed, starting at the air handler with the manufacturer-supplied sound-power levels. Consider the system in sections. The easiest way is to define a section as the area between duct branches and where sound-power splits occur. Silencers should be considered a section to account for insertion losses and generated noise. Each octave band must be evaluated.

Sound-power splits occur when divided flow fittings are used. The amount of attenuation is based on a ratio of the areas downstream. A tee with downstream areas A1 and A2 would have the following reductions in sound-power levels (L1 and L2):

L1 = 10 × log (A1 ÷ (A1 + A2))

L2 = 10 × log (A2 ÷ (A1 + A2))

If A1 equaled 400 sq in., and A2 equaled 900 sq in., then L1 would equal -5 dB, and L2 would equal -2 dB, the reductions applying to all of the frequencies being evaluated.

Consider any attenuation a section may provide. Even unlined ductwork provides some natural attenuation, and elbows provide end-reflection losses. Subtract these values from the entering sound-power levels. (Straightforward subtraction can be used for this process.)

Next, determine the amount of generated noise, and add it to the remaining sound-power levels using the method described previously. In many cases, the generated noise from fittings is too low to impact the sound-power levels of a section. Remember that if the difference between two sound sources is 10 dB or more, the louder source prevails without anything being added to it.

Repeat this process until you reach a terminal unit. At that point, add the remaining noise in the system to the generated noise of the outlet, which typically is a function of velocity.

End reflection occurs when a duct terminates into a large space. At low frequencies, sound loss can be substantial. Different diffusers and end conditions can result in data very different than those currently available. If a designer is not familiar with this condition, any sound-power losses should be ignored for duct-system evaluation.

Room acoustics can be difficult to evaluate during the early stages of design because so many factors — carpet, acoustic ceiling tile, and furniture, to name a few — contribute to them. ASHRAE provides a method of determining a room's acoustic characteristics at each frequency, based on a receiver's location relative to the duct outlets and the volume of the room. A more conservative approach is to assume a room offers no sound attenuation.

Once a system is evaluated, you can determine how much and at what frequencies attenuation is required. Make sure to evaluate supply and return ducts.


Breakout noise is noise that “escapes” ductwork. Rectangular duct allows more breakout noise than does round duct. Allowing noise, particularly low-frequency noise, to escape a duct system in a mechanical room or on a rooftop can be useful.

Break-in noise is noise that enters a duct system through duct walls. Rectangular duct allows more break-in noise than does round duct. If duct — rectangular or round — is routed near loud equipment, use exterior lagging and heavier gauges.

With lined duct, the thicker the liner, the more effective the attenuation of low-frequency noise.


Medium- and high-frequency noise usually is easier to attenuate than low-frequency noise. Lined or double-wall duct is reasonably effective at attenuating medium- and high-frequency noise. If a design does not allow sufficient attenuation because duct runs are too short, reroute the ductwork, or consider the use of silencers.


How air approaches a duct silencer, how the silencer is oriented to the fan, and where the silencer is located in relation to the equipment-room wall being penetrated are critical. Transitions must be designed so that airflow through a silencer complies with the manufacturer's intent. The best location for a duct silencer is centered on an equipment-room wall. If a fire or smoke damper is required in a wall, a silencer should be located as close to the wall as possible, within the equipment room. Otherwise, sound may radiate from the equipment room directly into exit ducts.

There are three types of duct silencers: dissipative, reactive (no sound-absorbing material in the cavities), and active. When selecting a silencer, consider:

  • Insertion loss — the difference between sound-pressure levels measured at the same point before and after a silencer is installed.

  • Static-pressure drop — the pressure drop across the silencer.

  • Regenerated sound — sound generated by airflow through the silencer.

System effects can significantly impact the performance of dissipative and reactive silencers. Silencers with internal baffles should be located at least three duct diameters from a fan, elbow, branch takeoff, or other duct element. Locating a silencer closer than three duct diameters can result in a significant increase in static-pressure drop across the silencer, which usually increases both fan and silencer airflow-generated noise.

Airflow-generated noise occurs as air flows into and out of a silencer. When static-pressure drop across a silencer is greater than 0.35 in. wg, airflow-generated noise should be evaluated.

Dissipative and reactive silencers can have a rectangular or a circular cross section. Rectangular and circular straight silencers are available in varying shapes. Straight silencers can have side baffles, center baffles, or both. Special fan-inlet and discharge silencers, including cone silencers, and silencers without internal baffles are designed to minimize system effects and attenuate fan sound pressure at its source.

The side and center pods of dissipative and reactive silencers are constructed of perforated sheet metal. The cavities of the pods of dissipative silencers are filled with either fiber glass or mineral wool, which provides good broadband sound-attenuation characteristics. The cavities of the pods of reactive silencers are tuned chambers void of fill material. Because of the tuning associated with reactive silencers, broadband sound attenuation often is more difficult to achieve.

Dissipative and reactive silencers come in several pressure-drop configurations. Insertion loss, airflow-generated noise, and pressure drop are functions of the silencers' design and location in a duct system. These data are measured experimentally and presented by manufacturers. The data should be obtained according to the procedures outlined in ASTM E477-06a, Standard Test Method for Measuring Acoustical and Airflow Performance of Duct Liner Materials and Prefabricated Silencers.

Active duct silencers produce inverse sound waves that cancel unwanted sound. They are effective in attenuating low-frequency, pure-tone, and broadband sound. Active duct silencers consist of a microprocessor, two microphones placed a specified distance apart in or on a duct, and a speaker placed between the microphones and mounted externally, but radiating sound into the duct. The microphone closer to a source of low-frequency noise senses the sound. The microphone signal is processed by the microprocessor, which generates a signal that is 180 degrees out of phase with the sound and transmitted to the speaker. Sound from the speaker interferes with the unwanted sound, attenuating it. The other microphone, downstream of the speaker, senses the attenuated sound and sends a corresponding feedback signal to the microprocessor, so the speaker signal can be adjusted, if necessary.

Active duct-silencer systems have few, if any, components located within ducts. Thus, they can be used to attenuate unwanted sound with little, if any, pressure drop and without the introduction of regenerated sound into a duct. Discrete frequency tones between 40 and 400 Hz typically can be attenuated by as much as 35 dB. Broadband sound in this frequency range can be attenuated by 10 to 20 dB.

Active duct silencers are limited by airflow turbulence and cross modes near the microphones. The microphones detect noise associated with turbulence as “pseudonoise” inhibiting the controller's ability to analyze the sound being attenuated. Thus, active systems should not be used where airflow velocity is greater than 1,500 fpm or there are duct elements or transitions that can generate significant turbulence. Active duct silencers are most effective at attenuating plane waves, where sound pressure across a duct's cross section is constant. At higher frequencies where duct cross modes exist, the sound pressure across a duct's cross section is not constant, and the corresponding effectiveness of an active silencer in attenuating sound is reduced significantly.


For more information on HVAC-system acoustics, see HVAC Sound and Vibration Manual,2 Chapter 10 of HVAC Systems — Duct Design,3 2005 ASHRAE Handbook — Fundamentals,1 and Chapter 47 of 2007 ASHRAE Handbook — HVAC Applications.4


  1. ASHRAE. (2005). 2005 ASHRAE handbook — fundamentals. Atlanta: American Society of Heating, Refrigerating and Air-Conditioning Engineers.

  2. SMACNA. (2004). HVAC sound and vibration manual. Chantilly, VA: Sheet Metal and Air Conditioning Contractors' National Association.

  3. SMACNA. (2006). HVAC Systems — Duct Design. Chantilly, VA: Sheet Metal and Air Conditioning Contractors' National Association.

  4. ASHRAE. (2007). 2007 ASHRAE handbook — HVAC applications. Atlanta: American Society of Heating, Refrigerating and Air-Conditioning Engineers.

For past HPAC Engineering feature articles, visit

As project manager, technical resources, for the Sheet Metal and Air Conditioning Contractors' National Association (SMACNA), Mark Terzigni oversees the development and publication of technical manuals and provides interpretations of SMACNA standards. Prior to joining SMACNA, he was a regional engineer for McGill AirFlow LLC, assisting sales-staff members, customers, and design professionals with technical issues regarding duct construction, duct design, sound attenuation, and leak-testing equipment. He has a bachelor's degree in mechanical engineering from The Ohio State University.