Whether an occupant considers the background noise acceptable generally depends on two factors. First is the perceived loudness of the noise relative to that of normal activities; if it is clearly noticeable, it is likely to be distracting and cause complaint. Second is the sound quality of the background noise; if the noise is perceived as a rumble, throb, roar, hiss, or tone, this may result in complaints of annoyance and stress. The frequency spectrum is then said to be unbalanced.

The acoustical design must ensure that HVAC noise is of sufficiently low level and unobtrusive quality so as not to interfere with occupancy use requirements. If background noise reduces speech intelligibility, for example, complaints of lost productivity can result. Accordingly, methods of rating HVAC-related noise ideally assess both perceived loudness and sound quality (Wang et al. 2013).

Design Guidelines for HVAC-Related Background Sound in Rooms. Table 1 presents recommended targets for indoor background noise levels in various types of unoccupied rooms due to noise from HVAC systems operating at normal/full-load design conditions. Perceived loudness and task interference are factored into the numerical part of the rating. The sound quality design target is assumed to be a neutral-sounding spectrum, although some spectral imbalance is probably tolerable within limits for most users. The criteria used are described in the next section.

The following notes should be considered in using Table 1:

1. NC/RC values listed in Table 1 are not an exact science, but are based on judgment and experience to represent generally acceptable background noise for typical building occupancies as would be normally measured in fully built-out and furnished spaces unless otherwise specified. An acceptable noise level depends on the specific use of the space. Thus, individual projects should choose a design target, and that may vary from Table 1 values by ±5 dB as deemed suitable by the project team. For example, private offices and conference rooms are listed as NC/RC 30, but in executive offices or specialty conference rooms, a noise criterion as low as NC/RC 25 might be warranted. On the other hand, there is not always a benefit in achieving a lower number, as some background noise can improve acoustical privacy between adjacent areas.

It should be understood that sound level measurements for determining compliance with specifications can be obtained under ANSI S12.72 and are generally based on room-average sound readings unless specified otherwise. See “Determining Compliance” subheading which follows.

NC: ASHRAE recommends the historic method for use of this metric, where the rating number is equal to the highest tangent line value when octave band sound levels are plotted against a family of reference curves.

RC: When sound quality in the space is important, the RC metric provides a diagnostic tool to quantify the speech interference level (500-2000 Hz) and/or spectral imbalance (for example a dominance of low-frequency noise).

dBA and dBC are overall sound pressure level measurements with A- and C-weighting, and serve as good references for a simple, single-number measurement. They are also appropriate for specifications in cases where no octave band sound data are available for design.

2. Beware of possible contribution to background noise levels from intrusive non-HVAC sources such as elevators, and traffic (e.g. through windows).

3. An experienced acoustical consultant should be retained for guidance on acoustically critical spaces (below NC/RC 30) and for all performing arts spaces.

4. Some educators and others believe that HVAC-related sound criteria for schools, as listed in previous editions of this table, are too high and impede learning for affected groups of all ages. See ANSI/ASA Standard S12.60 for classroom acoustics and a justification for lower sound criteria in schools. The contribution of HVAC-related sources to the total noise level meets the background noise requirement of that standard if HVAC-related background sound is approximately NC/RC 25. Within this category, designs for K-8 schools should be quieter than those for high schools and colleges.

5. The NC/RC designations relate to reference curves with octave band sound pressure levels that are (1) selected based on appropriate loudness in the speech interference range (500 to 2000 Hz) and (2) show contours for high and low frequencies that are balanced at the same loudness level. Acoustical evaluation based on octave bands and target balanced contours is recommended, because overall dBA ratings do not reflect undesirable contributions of excessive low-frequency noise. The dBA and dBC levels are listed only as approximate references in the case of simplistic measurements, where dBA indicates relative loudness and dBC indicates prevalence of low-frequency noise. Exact specifications should be established by acoustical experts considering occupant sensitivity.

Criteria Descriptions. This section presents ways to rate or measure the sound to determine acceptability. The information should help the design engineer select the most appropriate background noise rating method for a specific project. Current methods described here and in other references include the traditional A-weighted sound pressure level (dBA) and tangent Noise Criteria (NC), the Room Criterion (RC) and RC Mark II, the Balanced Noise Criterion (NCB), and the Room Noise Criteria (RNC). Each method was developed based on data for specific applications; hence, not all are equally suitable for rating HVAC-related noise in the variety of applications encountered. The preferred sound rating methods generally comprise two distinct parts: a family of criterion curves (specifying sound levels by octave bands), and a procedure for rating the calculated or measured sound data relative to the criterion curves with regard to sound quality.

Ideally, HVAC-related background noise should have the following characteristics:

dBA and dBC: A- and C-Weighted Sound Level. The A-weighted sound level (described in Chapter 8 of the 2021 ASHRAE Handbook—Fundamentals) has been used for more than 60 years as a single-number measure of the relative loudness of noise, especially for outdoor environmental noise standards. The rating is expressed as a number followed by dBA (e.g., 40 dBA).

A-weighted sound levels can be measured with simple sound level meters. The ratings correlate fairly well with human judgments of relative loudness but take no account of spectral balance or sound quality. Thus, two different spectra can result in the same numeric value, but have quite different subjective qualities.

Along with dBA, there is also a C-weighted sound level, denoted as dBC, which is more sensitive to low-frequency sound contributions to the overall sound level than is dBA. When the quantity dBC – dBA is large (e.g., greater than 25 dB), significant low-frequency sound is present. It is recommended that when specifying background sound levels in dBA, the dBC is also included in the specification and does not exceed the dBA reading by more than 20 dB.

NC: Noise Criteria Method. The NC method for rating noise (described in Chapter 8 of the 2021 ASHRAE Handbook—Fundamentals) has been used for more than 50 years. It is a single-number rating that is somewhat sensitive to the relative loudness and speech interference properties of a given noise spectrum. The method consists of a family of criterion curves, shown in Figure 5, extending from 63 to 8000 Hz, and a tangency rating procedure. The criterion curves define the limits of octave band spectra that must not be exceeded to meet occupant acceptance in certain spaces. The rating is expressed as NC followed by a number (e.g., NC 40). The octave midband frequency of the point at which the spectrum is tangent to the highest NC curve should also be reported (e.g. NC 40 [125 Hz]). The NC values are formally defined only in 5 dB increments, with intermediate values determined by discretionary interpolation.

Widely used and understood, the NC method is sensitive to level but has the disadvantage that the tangency method used to determine the rating does not require that the noise spectrum precisely follow the balanced shape of the NC curves. Thus, sounds with different frequency content can have the same numeric rating, but rank differently on the basis of sound quality. With the advent of VAV systems, low-frequency content (i.e., below the 63 Hz octave band) is prevalent, and the NC rating method fails to properly address this issue (Ebbing and Blazier 1992). Consequently, if the NC method is chosen, sound levels at frequencies below 63 Hz must be evaluated by other means.

In HVAC systems that do not produce excessive low-frequency noise and strong discernible pure tones, the NC rating correlates relatively well with occupant satisfaction if sound quality is not a significant concern. NC rating is often used because of its simplicity.

RC/RC Mark II: Room Criteria Method. ASHRAE previously recommended the Room Criterion (RC) curves (beginning in Chapter 43 in the 2020 ASHRAE Handbook—HVAC Systems and Equipment; Blazier 1981a, 1981b) as an enhanced method for rating HVAC system related noise. The revised RC Mark II method is now preferred.

The RC method is a family of criterion curves and a rating procedure. The shape of these curves represents a well-balanced, bland-sounding spectrum, including two additional octave bands (16 and 31.5 Hz) to deal with excessive low-frequency noise. This rating procedure assesses background noise in spaces on the basis of its effect on speech, and on subjective sound quality. The rating value is expressed as RC followed by a number that represents the level of noise in the speech interference region of the spectrum, and a letter to indicate the quality (e.g., RC 35[N], where N denotes the desirable neutral rating). The RC method includes evaluation of the potential for noise-induced vibration from excessive airborne sound levels at and below 63 Hz.

Based on experience and ASHRAE-sponsored research (Broner 1994), the RC method was revised to the RC Mark II method (Blazier 1997). Like its predecessor, the RC Mark II method is intended for use as a diagnostic tool for analyzing noise problems in the field. The RC Mark II method is complicated, but computerized spreadsheets and HVAC system analysis programs are available to perform the calculations and graphical analysis.

The RC Mark II method has three parts: (1) a family of criterion curves (Figure 6), (2) a procedure for determining the RC numerical rating and the noise spectral balance (quality), and (3) a procedure for estimating occupant satisfaction when the spectrum does not have the shape of an RC curve (quality assessment index) (Blazier 1995). The rating is expressed as RC followed by a number and a letter (e.g., RC 35[N]). The number is the arithmetic average rounded to the nearest integer of sound pressure levels in the 500, 1000, and 2000 Hz octave bands (the main speech frequency region) and is known as the preferred speech interference level (PSIL). The letter is a qualitative descriptor that identifies the sound’s perceived character: (N) for neutral, (LF) for low-frequency rumble, (MF) for midfrequency roar, and (HF) for high-frequency hiss. There are also two subcategories of the low-frequency descriptor: (LF_B), denoting a moderate but perceptible degree of sound-induced ceiling/wall vibration, and (LF_A), denoting a noticeable degree of sound-induced vibration.

Each reference curve in Figure 6 identifies the shape of a neutral, bland-sounding spectrum, indexed to a curve number corresponding to the sound level in the 1000 Hz octave band. The shape of these curves is based on Blazier (1981a, 1981b), modified at 16 Hz following recommendations of the research in Broner (1994). Regions A and B denote levels at which sound can induce vibration in light wall and ceiling construction, which can potentially cause rattles in light fixtures, furniture, etc. Curve T is the octave band threshold of hearing as defined by ANSI Standard S12.2.

Acceptable outdoor sound levels are generally specified by local noise ordinances or other government codes, which almost always use the A-weighted noise level (dBA) as their metric. The usual metric is either L_max (maximum noise level over a period) or L_eq (average noise level over a period). The time constant (FAST or SLOW) used for L_max or L_p depends on the code.

Some communities have no ordinance and depend on state regulations that often use the day/night noise level descriptor L_DN, which is a combination of the daytime (7:00 am to 10:00 pm) and nighttime (10:00 pm to 7:00 am) average noise levels (L_eq) with a 10 dB penalty for nighttime. Other descriptors also exist; specific requirements should be identified at the outset of each project. In some cases, regulatory agencies may also impose project-specific noise conditions on the basis of community reaction and for maintaining an appropriate acoustic environment at the project vicinity.

Measurement or estimation of community noise is based on a location, often at the receiver’s property line at the approximate height of a standing individual, typically 5 ft. Measurements should also be taken 5 ft above outdoor elevated decks or outside operable windows where a receiver’s ears may be located anywhere within the property line including near the facade of the closest dwelling unit. Alternatively, the measurement may be made at the property line of the noise source.

In the absence of a local noise ordinance, county or state laws or codes or those of a similar community should be used. Even if activity noise levels do not exceed those specified by an ordinance, community acceptance is not ensured. Very low ambient levels or a noise source with an often-repeated, time-varying characteristic or strong tonal content may increase the likelihood of complaints. Without local ordinances, noise levels between 45 and 55 dBA may be considered in residential zones and 55 to 65 dBA in commercial zones. These are for outdoor use areas and, with standard building constructions, they also typically result in acceptable interior noise levels. Often, daytime noise levels (the period of daytime to be defined) are 10 dB higher than nighttime levels.

Although most ordinances are given as A-weighted pressure level, attenuation by distance, barriers, buildings, and atmosphere are all frequency dependent. Thus, A-weighted levels do not give an accurate estimation of noise levels at distances from the source. If A-weighted sound levels of sources must be determined by means other than measurement, then octave band or one-third octave band measurements of source sound pressure level at a distance, or (preferably) sound power level, must be obtained before calculating the attenuation.

Prediction of Fan Sound Power. The sound power generated by a fan performing at a given duty is best obtained from manufacturers’ test data taken under approved test conditions (AMCA Standard 300 or ASHRAE Standard 68/AMCA Standard 330). Applications of air-handling products range from stand-alone fans to systems with various modules and attachments. These appurtenances and modules can have a significant effect on air-handler sound power levels. In addition, fans of similar aerodynamic performance can have significant acoustical differences.

Predicting air-handling unit sound power from fan sound levels is difficult. Fan sound determined by tests may be quite different once the fan is installed in an air handler, which in effect creates a new acoustical environment. Proper testing to determine resulting sound power levels once a fan is installed is essential. Fan manufacturers are in the best position to supply information on their products, and should be consulted for data when evaluating the acoustic performance of fans for an air handler application. Similarly, air handler manufacturers are in the best position to supply acoustic information on air handlers.

Air handler manufacturers typically provide discharge, inlet, and casing-radiated sound power levels for their units based on one of two methods. A common method is the fan-plus-algorithm method: the fan is tested as a stand-alone item, typically using AMCA Standard 300, and an algorithm is used to predict the effect of the rest of the air-handling unit on the sound as it travels from the fan to the discharge and intake openings or is radiated through a casing with known transmission loss values. Another method is described in AHRI Standard 260, in which the entire unit is tested as an assembly, including fans, filters, coils, plenums, casing, etc., and the sound power level at the inlet and discharge openings, as well as the radiated sound power, is measured in a qualified reverberant room. Whenever possible, data obtained by the AHRI Standard 260 method should be used because it eliminates much of the uncertainty present in the fan-plus-algorithm method. For a detailed description of fan operations, see Chapter 21 in the 2020 ASHRAE Handbook—HVAC Systems and Equipment. Different fan types have different noise characteristics and within a fan type, several factors influence noise.

Point of Fan Operation. The point of fan operation has a major effect on acoustical output. Fan selection at the calculated point of maximum efficiency is common practice to ensure minimum power consumption. In general, for a given design, fan sound is at a minimum near the point of maximum efficiency. Noise increases as the operating point shifts to the right, as shown in Figure 8 (higher airflow and lower static pressure). Low-frequency noise can increase substantially at operating points to the left of maximum efficiency (lower airflow and higher static pressure). These operating points should be avoided.

Blade-Pass Frequency. The blade-pass frequency is represented by the number of times per second a fans impeller passes a stationary item: f_bp = (rpm × number of impeller blades)/60. All fans generate a tone at this frequency and its multiples (harmonics). Whether this tone is objectionable or barely noticeable depends on the type and design of the fan and the point of operation.

Housed Centrifugal Fans. Forward-curved (FC) fans are commonly used in a wide range of standard air-handler products. The blade-pass of FC fans is typically less prominent and is at a higher frequency than other fans. The most distinguishing acoustical concern of FC fans is the prevalent occurrence of low-frequency rumble from airflow turbulence generated at blade tips, which can be exacerbated by nonideal discharge duct conditions (less than five diameters of straight duct). FC fans are commonly thought to have 16, 31.5, and 63 Hz (full octave band) rumble, particularly when operating to the left of the maximum efficiency point.

Backward-inclined (BI) fans and airfoil (AF) fans are generally louder at the blade-pass frequency than a given FC fan selected for the same duty, but are much more energy efficient at higher pressures and airflow. The blade-pass tone generally increases in prominence with increasing fan speed and is typically in a frequency range that is difficult to attenuate. Below the blade-pass frequency, these fans generally have lower sound amplitude than FC fans and are often quieter at high frequencies.

Care should be taken with all types of housed fans to allow adequate clearance around the inlets. Also, note that belt guards and inlet screens may decrease airflow and increase sound generation.

Plenum Fans. A plenum fan has no housing around the fan impeller and discharges directly into the chamber, pressurizing the plenum, and forcing air through the attached ductwork. Air flows into the fan impeller through an inlet bell located in the chamber wall. These fans can substantially lower discharge sound power levels if the fan plenum is appropriately sized and acoustically treated with sound-absorptive material.

The plenum discharge should be located away from the fan’s air blast, because blowing directly into the duct can aggravate the blade-pass sound. Avoid obstructing the inlet or crowding the coils or filters.

Vaneaxial Fans. Generally thought to have the lowest amplitudes of low-frequency sound of any of the fan types, axial fans are often used in applications where the higher-frequency noise can be managed with attenuation devices. In the useful operating range, noise from axial fans is a strong function of the inlet airflow symmetry and blade tip speed.

Propeller Fans. Sound from propeller fans generally has a low-frequency-dominated spectrum shape; the blade-pass frequency is typically prominent and occurs in the low-frequency bands because of the small number of blades. Propeller fan blade-pass frequency noise is very sensitive to inlet obstructions. For some propeller fan designs, the shape of the fan venturi (inlet) is also a very important parameter that affects sound levels. In some applications, noise of a propeller fan is described as sounding like a helicopter. Propeller fans are most commonly used on condensers and for power exhausts.

Minimizing Fan Noise. To minimize the required air distribution system sound attenuation, proper fan selection and installation are vital. The following factors should be considered:

General Design Considerations. As in other aspects of HVAC system design, ducts for VAV systems should be designed for the lowest practical static pressure loss, especially ductwork closest to the fan or air-handling unit (AHU). High airflow velocities and convoluted duct routing with closely spaced fittings can cause turbulent airflow that results in excessive pressure drop and fan instabilities that can cause excessive noise, fan stall, or both.

Many VAV noise complaints have been traced to control problems. Although most problems are associated with improper installation, many are caused by poor design. The designer should specify high-quality fans or air handlers within their optimum ranges, not at the edge of their operation ranges where low system tolerances can lead to inaccurate fan flow capacity control. Also, in-duct static pressure sensors should be placed in duct sections having the lowest possible air turbulence (i.e., at least three equivalent duct diameters from any elbow, takeoff, transition, offset, or damper).

Balancing. VAV noise problems have also been traced to improper air balancing. For example, air balance contractors commonly balance an air distribution system by setting all damper positions without considering the possibility of reducing fan speed. The result is a duct system in which no damper is completely open and the fan delivers air at a higher static pressure than would otherwise be necessary. If the duct system is balanced with at least one balancing damper wide open, fan speed and corresponding fan noise could be reduced. Lower sound levels occur if most balancing dampers are wide open or eliminated. The specified goal should be to balance the system at the lowest static pressure required to operate the box located at the farthest point in the system.

Fan Selection. For constant-volume systems, fans should be selected to operate at maximum efficiency at design airflow. However, VAV systems must be selected to operate with efficiency and stability throughout the operating range. For example, a fan selected for peak efficiency at full output may aerodynamically stall at an operating point of 50% of full output, resulting in significantly increased low-frequency noise and unstable airflow. A stalling fan can indicate operation in the surge region, an area of operational instability where airflow reverses direction at the fan blade because of insufficient air entering the fan wheel. Similarly, a fan selected to operate most efficiently at the 50% output point may be very inefficient at full output, resulting in substantially increased fan noise at all frequencies. In general, a fan for a VAV system should be selected for peak efficiency at an operating point between 70 and 80% of the maximum required system capacity, which is where the fan will operate most of the time. This usually means selecting a fan that is one size smaller than that required for peak efficiency at 100% of maximum required system capacity (Figure 9). When the smaller fan operates at higher capacities, it produces up to 5 dB more noise. This occasional increase in sound level is usually more tolerable than stall-related sound problems that can occur with a larger fan operating at less than 100% design capacity most of the time.

Air Modulation Devices. The control method selected to vary the air capacity of a VAV system is important. Variable-capacity control methods can be divided into three general categories: (1) variable inlet vanes (sometimes called inlet guide vanes) or discharge dampers that yield a new fan system curve at each vane or damper setting, (2) variable-pitch fan blades (usually used on axial fans) that adjust the blade angle for optimum efficiency at varying capacity requirements, and (3) variable-speed motor drives in which motor speed is varied by modulation of the power line frequency or by mechanical means such as gears or continuous belt adjustment. Inlet vane and discharge damper volume controls can add noise to a fan system at reduced capacities, whereas variable-speed motor drives and variable-pitch fan blade systems are quieter at reduced air output than at full air output.

Variable-Inlet Vanes and Discharge Dampers. Variable-inlet vanes vary airflow capacity by changing inlet airflow to a fan wheel. This type of air modulation varies the total air volume and pressure at the fan, but fan speed remains constant. Although fan pressure and air volume reductions at the fan reduce duct system noise by reducing air velocities and pressures in the duct work, there is an associated increase in fan noise caused by airflow turbulence and flow distortions at the inlet vanes. Fan manufacturers’ test data show that, on airfoil centrifugal fans, as vanes mounted inside the fan inlet (nested inlet vanes) close, the sound level at the blade-pass frequency of the fan increases by 2 to 8 dB, depending on the percent of total air volume restricted. For externally mounted inlet vanes, the increase is on the order of 2 to 3 dB. The increase for forward-curved fan wheels with inlet vanes is about 1 to 2 dB less than that for airfoil fan wheels. In-line axial fans with inlet vanes generate increased noise levels of 2 to 8 dB in the low-frequency octave bands for a 25 to 50% closed vane position.

Discharge dampers, typically located immediately downstream of the supply air fan, reduce airflow and increase pressure drop across the fan while fan speed remains constant. Because of air turbulence and flow distortions created by the high pressure drop across discharge dampers, there is a high probability of duct rumble near the damper location. If the dampers are throttled to a very low flow, a stall condition can occur at the fan, resulting in an increase in low-frequency noise.

Variable-Pitch Fans for Capacity Control. Variable-pitch fan blade controls vary the fan blade angle to reduce airflow. This type of system is predominantly used in axial fans. As air volume and pressure are reduced at the fan, there is a corresponding noise reduction. In the 125 to 4000 Hz octave bands, this reduction usually varies between 2 to 5 dB for a 20% reduction in air volume, and between 8 to 12 dB for a 60% reduction in air volume.

Variable-Speed-Motor-Controlled Fan. Three types of electronic variable-speed control units are used with fans: (1) current source inverter, (2) voltage source inverter, and (3) pulse-width modulation (PWM). The current source inverter and third-generation PWM control units are usually the quietest of the three controls. In all three types, matching motors to control units and the quality of the motor windings determine the motor’s noise output. The motor typically emits a pure tone with an amplitude that depends on the smoothness of the waveform from the line current. The frequency of the motor tone depends on the motor type, windings, and speed, but is typically at the drive’s switching frequency. Some drives allow adjustment to a higher frequency that does not carry as well, but at a cost of lower drive efficiency. Both inverter control units and motors should be enclosed in areas, such as mechanical rooms or electrical rooms, where the noise effect on surrounding rooms is minimal. The primary acoustic advantage of variable-speed fans is reduction of fan speed, which translates into reduced noise; dB reduction is approximately equal to 50 × log (higher speed/lower speed). Because this speed reduction generally follows the fan system curve, a fan selected at optimum efficiency initially (lowest noise) does not lose efficiency as the speed is reduced. When using variable-speed controllers,

Terminal Units. Fans and pressure-reducing valves in VAV units should have manufacturer-published sound data indicating sound power levels that (1) are discharged from the low-pressure end of the unit and (2) radiate from the exterior shell of the unit. These sound power levels vary as a function of valve position and fan point of operation. Sound data for VAV units should be obtained according to the procedures specified by the latest AHRI Standard 880. In critical situations, a mock-up test should be conducted of a production terminal box under project conditions and space finishes. The test is required because minor changes in box motor, fan, or valve components can affect the noise generated by such equipment.

If the VAV unit is located in noncritical areas (e.g., above a storeroom or corridor), sound radiated from the shell of the unit may be of no concern. If, however, the unit is located above a critical space and separated from the space by a ceiling with little or no sound transmission loss at low frequencies, sound radiated from the shell into the space below may exceed the desired noise criterion. In this case, it may be necessary to relocate the unit to a noncritical area or to enclose it with a high-transmission-loss construction. Room sound levels can be estimated using attenuation factors detailed in AHRI Standard 885. In general, fan-powered VAV units should not be placed above or near any room with a required sound criterion rating of less than RC 40(N) (Schaffer 2005). For further information, see the section on Indoor Sound Criteria.

Full shutoff of VAV units can produce excessive duct system pressure at low flow, sometimes causing a fan to go into stall, resulting in accompanying roar, rumble, and surge. Systems providing more than 30% of their air to VAV devices should be provided with a means of static pressure control. Variable-frequency drives are preferred, but in the case of constant-volume air handlers, some means of bypass pressure control should be used to relieve system pressure as VAV devices close down (Schaffer 2005).

Rooftop air handlers can have unique noise control requirements because these units are often integrated into a lightweight roof construction. Large roof openings are often required for supply and return air duct connections. These ducts run directly from noise-generating rooftop air handlers to the building interior. Generally, there is insufficient space or distance between roof-mounted equipment and the closest occupied spaces below the roof to apply standard sound control treatments. Rooftop units should be located above spaces that are not acoustically sensitive and should be placed as far as possible from the nearest occupied space. This measure can reduce the amount of sound control treatment necessary to achieve an acoustically acceptable installation.

The common sound transmission paths associated with rooftop air handlers (Figure 10) are

Flanking-path noise can enter through lightweight roof structures, adjacent walls, and windows. Avoid placing rooftop units on light structure over sensitive spaces or close to higher sidewalls with windows or other lightly constructed building elements. If it is necessary to place the rooftop unit over sensitive spaces or lightly constructed walls, then lagging with additional layers of gypsum board or other similar material may be required in these areas.

Using proper vibration isolation can minimize structureborne sound and vibration from vibrating equipment in a rooftop unit. Special curb mounting bases are available to support and provide vibration isolation for rooftop units. For roofs constructed with open web joists, thin long-span slabs, wooden construction, and any unusually light construction, evaluate all equipment weighing more than 300 lb to determine the additional deflection of the structure at mounting points caused by the equipment. Isolator deflection should be a minimum of 10 times the additional static deflection. If the required spring isolator deflection exceeds commercially available products, stiffen the supporting structure or change the equipment location.

Airborne paths are associated with casing-radiated sound that passes through the air handler enclosure and roof structure to the spaces below. Airborne sound can result from air handler noise or from other equipment components in the rooftop unit. Rooftop units should not be placed on open curbs or over a large opening in the roof structure through which both supply and return air ducts pass. Roof penetrations should be limited to two openings sized to accommodate only the supply and return air ducts. These openings should be properly sealed after installation of the ducts. If a large single opening exists under the rooftop unit, it should be structurally, acoustically, and flexibly sealed with one or more layers of gypsum board or other similar material around the supply and return air ducts. Airborne sound transmission to spaces below a rooftop unit can be greatly reduced by placing the rooftop unit on a structural support extending above the roof structure, and running supply and return air ducts horizontally along the roof for several duct diameters before the ducts turn to penetrate the roof. The roof deck/ceiling system below the unit can be constructed to adequately attenuate sound radiated from the bottom of the unit.

Ductborne transmission of sound through the supply air duct consists of two components: sound transmitted from the air handler through the supply air duct system to occupied areas, and sound transmitted via duct breakout through a section or sections of the supply air duct close to the air handler to occupied areas. Sound transmission below 250 Hz through duct breakout is often a major acoustical limitation for many rooftop installations. Excessive low-frequency noise associated with fan noise and air turbulence in the region of the discharge section of the fan (or air handler) and the first duct elbow results in duct rumble, which is difficult to attenuate. This problem is often worsened by the presence of a high-aspect-ratio duct at the discharge section of the fan (or air handler). Rectangular ducts with duct lagging are often ineffective in reducing duct breakout noise. Using either a single- or dual-wall round duct with a radiused elbow coming off the discharge section of the fan can reduce duct breakout. If space does not allow for the use of a single duct, the duct can be split into several parallel round ducts. Another effective method is using an acoustic plenum chamber constructed of a minimum 2 in. thick, dual-wall plenum panel, lined with fiberglass and with a perforated inner liner, at the discharge section of the fan. Either round or rectangular ducts can be taken off the plenum as necessary for the rest of the supply air distribution system. Table 7 shows 12 possible rooftop discharge duct configurations with their associated low-frequency noise reduction potential (Beatty 1987; Harold 1986, 1991).

Ductborne transmission of sound through the return air duct of a rooftop unit is often a problem because there is generally only one short return air duct section between the plenum space above a ceiling and the return air section of the air handler. This does not allow for adequate sound attenuation between the fan inlet and spaces below the air handler. Sound attenuation through the return air duct system can be improved by adding at least one (more if possible) branch division where the return air duct is split into two sections that extend several duct diameters before they terminate into the plenum space above the ceiling. The inside surfaces of all return air ducts should be lined with a minimum of 1 in. thick duct liner. If conditions permit, duct silencers in duct branches or an acoustic plenum chamber at the air-handler inlet section give better sound conditions.

Aerodynamic sound is generated when airflow turbulence occurs at duct elements such as duct fittings, dampers, air modulation units, sound attenuators, and room air devices. For details on air modulation units and sound attenuators, see the sections on Variable-Air-Volume (VAV) Systems and Duct Silencers.

Although fans are a major source of sound in HVAC systems, aerodynamically generated sound can often exceed fan sound be-cause of close proximity to the receiver. When making octave-band fan sound calculations using a source-path-receiver analysis, aerodynamically generated sound must be added in the path sound calculations at the location of the element.

Duct Velocities. The extent of aerodynamic sound is related to the airflow turbulence and velocity through the duct element. The sound amplitude of aerodynamically generated sound in ducts is proportional to the fifth, sixth, and seventh power of the duct airflow velocity in the vicinity of a duct element (Bullock 1970; Ingard et al. 1968). Therefore, reducing duct airflow velocity significantly reduces flow-generated noise. Tables 8 (Schaffer 2005) and 9 (Egan 1988) give guidelines for recommended airflow velocities in duct sections and duct outlets to avoid problems associated with aerodynamically generated sound in ducts.

Fixed Duct Fittings. Fixed duct fittings include elbows, tees, transitions, fixed dampers, and branch takeoffs. In all cases, less generated air turbulence and lower airflow velocities result in less aerodynamic sound. Figures 11 and 12 show typical frequency spectra for specific sizes of elbows and transitions. Data in these figures are based on empirical data obtained from ASHRAE RP-37 (Ingard et al. 1968). Normalized data from ASHRAE RP-37 and others, which can apply to all types of duct fittings and dampers, have been published (Bullock 1970) and presented in ASHRAE RP-265 (Ver 1983a). When multiple duct fittings are installed adjacent to each other, aerodynamic sound can increase significantly because of the added air turbulence and increased velocity pressures. Note that the magnitude of the field-measured static pressure drop across fixed duct fittings does not relate to the aerodynamic generated sound. However, total pressure drop across a duct fitting, which includes the velocity pressure change resulting from air turbulence, does affect aerodynamically generated sound.

Operable Volume Dampers. Operable damper aerodynamic sound is created because the damper is an obstacle in the airstream, and air turbulence increases as the damper closes. Because total pressure drop across the damper also increases with closure, the aerodynamic sound is related to the total pressure drop. Both single-blade and multiblade dampers, used to balance and control the airflow in a duct system and at room air devices, have similar frequency spectra. Figure 13 shows the frequency spectrum for a 45° damper in a 24 by 24 in. duct (Ingard et al. 1968).

Depending on its location relative to a room air device, a damper can generate sound that is transmitted down the duct to the room air device, or radiate sound through the ceiling space into the occupied space below. When an operable control damper is installed close to an air device to achieve system balance, the acoustic performance of the air outlet must be based not only on the air volume handled, but also on the magnitude of the air turbulence generated at the damper. The sound level produced by closing the damper is accounted for by adding a correction to the air device sound rating. As the damper is modulated for air balance, this quantity is proportional to the pressure ratio (PR), that is, the throttled total pressure drop across the damper divided by the minimum total pressure drop across the damper. Table 10 provides decibel corrections to determine the effect of damper location on linear diffuser sound ratings.

Volume dampers in sound-critical spaces should always be a minimum of 5 to 10 duct diameters from the air device, with an acoustically lined duct between the damper and air device. Acoustically lined plenums may also be used between the damper and room air device to reduce damper sound. Linear air devices with a round duct connected to an insulated plenum have been successfully used for damper sound control. However, acoustical lining in this type of plenum does not minimize the sound generated by air flowing through a short section of the linear air device. If multiple inlets/outlets are used to spread airflow uniformly over the lined plenum and air device, then the linear slot generates less sound.

Proper air balancing of a fan/duct system directly affects aerodynamically generated sound even in a correctly designed and installed duct system. Primary volume dampers in the longest duct from a fan should always be nearly wide open. If the primary damper in the longest duct run is more than 20% closed, the duct system has not been properly air balanced, and the fan may operate at a higher speed than required for the duct system. The result is an increase in air velocities and turbulence throughout the entire duct system, with excessive aerodynamic sound generated at all duct elements.

Room Air Devices (Grilles, Registers, Diffusers). Manufacturers’ test data should be obtained in accordance with ASHRAE Standard 70 for room air devices such as grilles, registers, diffusers, air-handling light fixtures, and air-handling suspension bars. Devices should be selected to meet the noise criterion required or specified for the room. However, the manufacturer’s sound power rating is obtained with a uniform velocity distribution throughout the air device neck or grille collar; this is often not met in practice when a duct turn, sharp transition, or balancing damper immediately precedes the entrance to the diffuser. In these cases, airflow is turbulent and noise generated by the device can be substantially higher than the manufacturer’s published data (by as much as 12 dB). In some cases, placing an equalizer grid in the neck of the air device can greatly reduce this turbulence by helping to provide a uniform velocity gradient in the neck of the device, so the sound power generated in the field is closer to that listed in the manufacturer’s catalog.

At present, air devices are rated by manufacturers in terms of noise criterion (NC) levels, which usually includes a receiver room effect sound correction of 10 dB. The NC ratings may be useful for comparison between different air devices, but are not helpful for source-path-receiver calculations in terms of octave bands. For a complete analysis, the designer should request the component sound power level data in octave bands from the manufacturer. Whether using NC levels or sound power levels, the designer should also correct manufacturer’s data for actual room effect, location of air devices, and number of air devices used in a specific design. The acoustical room effect is the reduction in sound level caused by distance from the sound source (e.g., air outlet); the room volume and amount of acoustical absorption present also affect the value. For example, in a small room with an actual calculated room effect of 6 dB, and given a manufacturer’s room effect correction of 10 dB, the discrepancy (in this case, 4 dB) must be added to the manufacturer’s data. When an air device is located at the intersection of the ceiling and vertical wall, 6 dB should be added, and in the corner of a room, 9 dB should be added to manufacturer’s data. When multiple room air devices are located in a small room or grouped together in a large room, the sound of air devices is additive by up to 10 × log(number of air devices). For more information, see the section on Receiver Room Sound Correction.

A flexible duct connection between a branch air duct and an air device provides a convenient means to align the air device with the ceiling grid. The resulting misalignment in this connection, as shown in Figure 14, can cause as much as 12 to 15 dB higher sound levels in the air device’s aerodynamically generated sound.

Avoiding Aerodynamically Generated Noise. Aerodynamic noise in duct systems can be avoided by

Chillers and air-cooled condensers have components such as compressors, motors, gears, and fans that can produce significant amounts of both broadband and tonal noise. The broadband noise is typically caused by flows of refrigerant, water, and air, whereas the tonal noise is caused by rotation of compressors, motors, gears, and fans (in fan-cooled equipment). Chiller and condenser noise is significant in the octave bands from 63 to 4000 Hz and depends primarily on the type of compressor used.

Noise from Compressors and Chillers. All compressors produce tonal noise to varying degrees. Acoustical differences among compressors relate in large part to their tonal content:

The noise levels of indoor chillers are used primarily for determining compliance with occupational noise exposure in the workplace (in accordance with U.S. Occupational Safety and Health Administration [OSHA] regulations), and secondarily for determining equipment room transmission loss requirements to ensure that the desired sound levels in adjacent or remote spaces are achieved. The noise levels of outdoor chillers and condensing units are primarily used to determine compliance with local noise ordinances at property lines and to predict sound levels inside adjacent or nearby buildings and residences.

Indoor Water-Cooled Chillers. The dominant noise sources in water-cooled chillers are the compressor (most often centrifugal or screw) and condenser. As physical sizes of chillers increase, their radiated sound power levels also increase.

Factory-provided sound data for indoor chillers should be obtained in accordance with AHRI Standard 1280, which determines the equipment’s octave band sound power levels (L_w) at the following four load conditions: 100, 75, 50, and 25%. Note that equipment sound power levels do not change if the unit is placed in a different equipment room. If AHRI 1280 sound power data are not available, AHRI Standard 575 provides A-weighted and octave band sound pressure level (L_p) values at many locations 3.28 ft from the chiller and 4.92 ft above the floor. AHRI Standard 575 sound pressure levels are ideally reported at operating points of 25, 50, and 100% of a chiller’s nominal full capacity, but partial-load testing is not required by AHRI 575. The average A-weighted sound pressure levels can be used directly along with exposure times to determine OSHA compliance (OSHA 2008) in the machinery room. An example of the ranges of AHRI Standard 575 values for a typical centrifugal and screw chiller are shown in Figures 15 and 16, respectively. The sound levels depend on manufacturer design, type and size of compressor, and operating conditions.

AHRI Standard 575 measurements for factory-provided ratings are often made in very large rooms with an anechoic or minimally reverberant environment rather than that typically found in mechanical equipment rooms. For that reason, assessment of sound pressure levels in situ should typically be adjusted for each chiller installation to account for the mechanical room’s size and surface treatment. For a given chiller at a given operating point, a small equipment room (or one with mostly hard surface finishes) has higher L_p values than one that is large or has sound-absorbing treatments on its ceiling and walls. Figure 17 shows maximum typical adjustment factors that should be added to factory-provided AHRI Standard 575 values to estimate the L_p values in specific installations caused by reverberant (reflective) sound effects. The adjustment for each octave band requires knowing the size of an imaginary box that is circumscribed 3.28 ft away from the top and sides of the chiller (the AHRI Standard 575 measurement surface), the dimensions of the equipment room, and the average sound absorption coefficient of the room surfaces. The adjustment in each octave band depends on the ratio of the areas of the equipment room and the imaginary box as well as the average sound absorption of the room finishes. Each curve in Figure 17 is for a different value of the average sound absorption, with the higher curves being for lower values.

The approximate reverberant L_p values in the last line of Example 2 can be used with sound transmission loss data of the construction to estimate transmitted L_p values in rooms adjacent to a chiller room.

An alternative approach to this method for estimating L_p in adjacent rooms is to use an estimate of the sound power levels of the chiller from the factory-provided AHRI Standard 575 values (Stabley 2006) in conjunction with sound transmission loss data. A conversion factor (CF) is determined and used to convert the AHRI Standard 575 sound pressure values to sound power L_w values. The conversion factor is calculated using

where

S = area of measurement parallelepiped (excluding the top) used in AHRI Standard 575

= 2(L × H) + 2(W × H)

L, W, and H = length, width, and height of measurement parallelepiped, ft

So = 10.72 ft2

and

where

Lw = sound power level (A-weighted or octave band)

Lp = sound pressure levels per AHRI Standard 575

This approach assumes that the factory-provided data were obtained in a free-field environment.

Indoor chillers are often offered with various types of factory noise-reduction options. These options can include variable-speed drives, or variable-geometry diffusers on centrifugal compressors, to reduce the strength of the noise sources inside the machine. They may also include various external noise-attenuation devices ranging from compressor, refrigerant line, and heat exchanger blankets (typically providing overall noise reduction of 2 to 6 dBA), to complete enclosures with sound-absorbing inner surfaces (which may reduce the overall noise by as much as 18 dBA). The amount of compressor noise reduction achieved by external attenuation approaches is usually limited by structureborne transmission of compressor vibration into the equipment frame and heat exchanger shells, which act as sounding boards. Attenuation options for chiller noise control vary widely, depending on the application and the type of compressor used. Typically, they either reduce the sound radiating from the source (using acoustic enclosures or blankets) or reduce the internal sound-generating mechanisms of the source (using variable-speed drives on compressors and variable-geometry diffusers for centrifugal compressors). The effectiveness of each approach is affected by variables such as the type of compressor and its behavior with load, heat exchanger design, and type of prime mover used.

Field-installed noise-control options include full-sized sheet metal housings with specially treated openings for piping, electrical conduit, and ventilation. This option may require upgraded building construction. For more information, refer to the section on Mechanical Equipment Room Sound Isolation.

Outdoor Air-Cooled Chillers and Condensers. Outdoor units often use either reciprocating, scroll, or screw compressors. They are also used as the chiller portion of rooftop packaged units. The dominant noise sources in outdoor air-cooled chillers are the compressors and the condenser fans, which are typically low-cost, high-speed propeller fans. For air-cooled condensing units, propeller fans are the only significant noise source.

Factory sound data for outdoor equipment are obtained in accordance with AHRI Standard 370, which requires determination of the equipment’s octave band sound power levels (L_w), A-weighted overall sound power level (L_wa), and tone-adjusted A-weighted overall sound power level (L_wat). Because AHRI Standard 370 is a sound power measurement technique, it provides certifiable sound data that can be compared across chiller manufacturers with greater certainty than is possible using the sound-pressure-based AHRI Standard 575. An example of the range of AHRI Standard 370 L_w values for outdoor chillers in the 20 to 380 ton range is given in Figure 18.

Factory-supplied noise reduction options for outdoor equipment include compressor enclosures, component sound blankets, oversized condenser fans, and variable-speed condenser fans. Because air-cooled equipment needs a free flow of cooling air, full enclosures are not feasible. However, strategically placed barriers can help reduce noise propagation on a selective basis. For more information, see the section on Sound Control for Outdoor Equipment.

Emergency or standby generators create very high sound levels and require special consideration, especially if used inside an occupied building. The primary noise sources include the engine casing, radiator, and engine exhaust, and must be considered separately if the generator is installed indoors. Sound power levels for these units depend on the power rating, fuel type, engine speed, exhaust muffler design, and radiator system. Overall sound power levels can be as high as 130 dBA (re 10^–12 W) for larger (1.5 to 2.0 MW) diesel-powered units with standard mufflers. Noise levels inside generator rooms almost always exceed 100 dBA if the power rating of the unit is greater than 50 kW.

Noise from the generator casing is broadband with a relatively uniform spectrum. Octave band noise data are usually available from the generator manufacturer. Casing noise does not vary much with load. Conversely, exhaust noise typically contains strong tones at the engine shaft’s running frequency and at the engine firing rate. Standard engine exhaust mufflers reduce exhaust noise by 20 to 25 dBA (compared to unsilenced exhaust), but even with this noise reduction, strong tones still radiate from the exhaust outlet in most cases. High-performance (critical- and supercritical-grade) mufflers are available but are larger and more expensive than standard units. Exhaust pipes should be routed away from noise-sensitive areas, and the exhaust outlet should be located and oriented to ensure that community noise levels are not excessive. In occupied buildings, the entire exhaust pipe should be suspended from the structure above with spring hangers.

Engine casing noise is best controlled by enclosing the generator in a sound-rated enclosure. The biggest problem with generator room design is finding adequate space for ventilation air. Generators require a substantial volume of air for engine cooling, and controlling engine noise transmission using air intake and exhaust paths can be difficult. In most cases, air intake and exhaust openings require sound attenuators. Because engine radiators usually use propeller fans to move air across the radiator core, the ventilation system cannot always handle the added pressure drop created by sound attenuators. In some cases, auxiliary fans are needed to draw fresh air into the generator room through the intake silencers. Sound attenuators at the discharge opening should be located between the radiator and exhaust louver. A smooth, slowly expanding transition duct is required between the radiator and the discharge louver. Nonsymmetrical transitions and transitions with expansion angles greater than 15° usually result in nonuniform airflow through the sound attenuators, causing a higher-than-expected pressure drop and reduced airflow. Unless careful space planning is done early in the design phase, there is often not enough space in the generator room to fit the sound attenuators with a proper transition fitting.

Emergency generator radiator fans usually make as much or more noise than the engine casing. In noise-sensitive installations, consider using a remote radiator. Most manufacturers offer remote radiators as an option, available in a wide variety of designs and noise levels. Lower sound levels with remote radiators are usually achieved by using larger fans running at lower speeds. Using a remote radiator can often save substantial expense in noise-sensitive applications because airflow requirements in the generator room are dramatically reduced.

A major transmission path of noise from mechanical equipment is through air distribution ductwork. Duct elements and concepts covered in this section include plenums, unlined rectangular ducts, acoustically lined rectangular ducts, unlined round ducts, acoustically lined round ducts, elbows, acoustically lined round radiused elbows, duct silencers, duct branch power division, duct end reflection loss, and terminal volume regulation units. Simplified tabular procedures for obtaining the sound attenuation associated with these elements are presented.

Plenums. Plenums are often placed between a fan and main air distribution ducts to smooth turbulent airflow. They are typically lined with acoustically absorbent material to reduce fan and other mechanical noise. Plenums are usually large rectangular enclosures with an inlet and one or more outlets.

Based on experience, ASHRAE-sponsored research (Mouratidis and Becker 2004), and earlier work (Wells 1958), transmission loss associated with a plenum can be expressed using the following considerations:

At frequencies above the cutoff frequency, as defined by the plenum’s inlet duct dimensions, the wavelength of sound is small compared to the characteristic dimensions of the plenum. Plane wave propagation in a duct exists at frequencies below the cutoff, creating a need to consider two frequency ranges, where

where

fc = cutoff frequency, Hz

c = speed of sound in air, ft/s

a = larger cross-sectional dimension of rectangular duct, ft

d = diameter of round duct, ft

The cutoff frequency f_c is the frequency above which plane waves no longer propagate in a duct. At these higher frequencies, waves that propagate in the duct create cross or spinning modes. The transmission loss (TL) in this higher frequency range may be predicted using the following relationship:

where

TL = transmission loss, dB

b = 3.505

n = –0.359

Sout = area of plenum outlet, ft2

S = total inside surface area of plenum minus inlet and outlet areas, ft2

r = distance between centers of inlet and outlet of plenum, ft

Q = directivity factor; taken as 2 for opening near center of wall, or 4 for opening near corner of plenum

αa = average absorption coefficient of plenum lining (see Equation [8])

OAE = offset angle effect; additional attenuation found in Tables 14 and 15, which tabulate frequency-dependent sound transmission properties that are manifested when inlet and outlet of plenum are not in a direct line; 90° angle is referred to as elbow effect

The average absorption coefficient α_a of plenum lining is given by

where

α1 = sound absorption coefficient of any bare or unlined inside surfaces of plenum

S1 = surface area of any bare or unlined inside surfaces of plenum, ft2

α2 = sound absorption coefficient of acoustically lined inside surfaces of plenum

S2 = surface area of acoustically lined inside surfaces of plenum, ft2

In many situations, inner surfaces of a plenum chamber are lined with a sound-absorbing material. For these situations, α_a = α₂. Table 12 gives sound absorption coefficients for selected common plenum materials.

Note: transmission loss (TL) of a plenum is the difference between the duct sound power level at the outlet and inlet of the plenum, unlike insertion loss (IL) ratings for silencers, which represent the difference (at a downstream measurement location) between the duct sound pressure levels with the silencer and with no silencer (replaced with an empty duct). For purposes here, both TL and IL are interpreted as attenuation, or the net reduction in propagating duct sound power.

For frequencies that correspond to plane wave propagation in the duct (below the cutoff frequency), the following relationship applies, with a lower frequency limit of 50 Hz:

where

Af = surface area coefficient, dB/ft2 (see Table 13 for small and large plenum size ranges)

We = wall effect, dB (see Table 13 for common HVAC plenum wall types)

OAE = offset angle effect

The maximum TL predicted by Equation (7) should be limited to 20 dB.

For an end-in/end-out plenum configuration, where the openings are not in-line, the offset angle θ must be considered in the TL calculation. The value of θ is obtained from the following relationship:

where (refer to Figure 19)

θ = offset angle representing r to long axis l of duct

l = length of plenum, ft

rv = vertical offset between axes of plenum inlet and outlet, ft

rh = horizontal offset between axes of plenum inlet and outlet, ft

For a given offset angle, apply the applicable effects on TL (decibel addition or subtraction) for angles up to 45° (Table 14).

For an end-in/side-out plenum configuration, where openings are perpendicular to each other, the elbow effect must be considered in the TL calculation. For any plenum configuration involving an elbow condition, apply the applicable effects on TL (decibel addition or subtraction) for the two frequency ranges, both above and below the cutoff (Table 15).

For plenum applications within a practical size envelope of 20 to 1100 ft³ volume or 50 to 650 ft² surface area, using duct sizes in the range 12 < d < 48 in., this model may be applied with an anticipated standard deviation of ±3.5 to 5.0 dB for 50 Hz < f ≤ f_c and ±1.5 to 3.0 for f_c < f ≤ 5000 Hz. Use caution when applying these prediction methods for plenum configurations where either the width or height dimension is <1.5d. In this case, the plenum may not perform as an expansion chamber, thus changing its broadband TL characteristics significantly.

Unlined Rectangular Sheet Metal Ducts. Straight, unlined rectangular sheet metal ducts provide a fairly significant amount of low-frequency sound attenuation. Table 16 shows the results of selected unlined rectangular sheet metal ducts (Cummings 1983; Reynolds and Bledsoe 1989a; Ver 1978; Woods Fan Division 1973). Attenuation values in Table 16 apply only to rectangular sheet metal ducts with the lightest gages allowed by Sheet Metal and Air Conditioning Contractors’ National Association, Inc. (SMACNA) HVAC duct construction standards. Attenuation for lengths greater than 10 ft is not well documented.

Sound energy attenuated at low frequencies in rectangular ducts may manifest itself as breakout noise along the duct. Low-frequency breakout noise should therefore be checked. For additional information on breakout noise, see the section on Sound Radiation Through Duct Walls.

Acoustically Lined Rectangular Sheet Metal Ducts. Internal duct lining for rectangular sheet metal ducts can be used to provide both thermal insulation and sound attenuation. The thickness of duct linings for thermal insulation usually varies from 0.5 to 2 in.; the density of the internal lining usually varies between 1.5 and 3.0 lb/ft³, but may be as low as 0.75 lb/ft³. For duct lining to attenuate sound effectively, it should have a minimum thickness of 1 in. At low frequencies, thicker lining performs slightly better than thinner lining. Tables 17 and 18 give attenuation values of selected rectangular sheet metal ducts for 1 and 2 in. thick glass fiber duct lining, respectively (Herrin et al. 2017; Kuntz 1986; Kuntz and Hoover 1987; Machen and Haines 1983; Reynolds and Albright 2018; Reynolds and Bledsoe 1989a). Note that attenuation values for ducts with a cross-sectional area less than 1 ft² or greater than 16 ft² in these tables are taken from previous editions of the ASHRAE Handbook and based on laboratory tests using 10 ft lengths of lined duct. Insertion loss values for all other duct sizes in these tables were obtained in ASHRAE research project RP-1408 (Reynolds et al. 2018). This research project included lengths of uninterrupted straight ducts from 3 to 40 ft with varying sizes and lengths of round and rectangular ducts with fiberglass duct liner, and also captured data at 63 and 8000 Hz not reported in earlier research. The liner used in this testing was 1.5 lb/ft³ for the rectangular ducts.

Though previous editions of this table indicated values of dB/ft, research projects have shown that attenuation does not vary exactly linearly with length. Therefore, these updated tables show attenuation for lengths as actually tested. For interpolation to various duct lengths and cross sections, using the duct liner attenuation spreadsheet (available as an extra feature for the ASHRAE Handbook Online version of this chapter) is recommended. Note that the total attenuated noise will never be below the flow-generated noise level in the duct, and, because of typical flows and flanking noise in practice, attenuation values are capped at 50 dB.

Insertion loss values in Tables 17 and 18 represent the difference in the space-average sound pressure level measured in a reverberation room with sound propagating through a straight section of unlined rectangular duct minus the corresponding sound pressure level measured when the unlined section of rectangular duct is replaced with a similar section of acoustically lined rectangular duct. The net result is the attenuation resulting from adding duct liner to a sheet metal duct.

Insertion loss values discussed in this section apply only to straight rectangular sheet metal ducts made with the lightest gages allowed by the SMACNA HVAC duct construction standards. Insertion loss values for heavier gage ductwork are expected to be slightly less than the tabulated values. Insertion loss values in octave and 1/3-octave bands are available in the online version of the Handbook. Also available in the online version is a calculator which provides octave band and 1/3-octave band insertion loss values for any duct cross section larger than 12 × 12 in. and smaller than 48 × 48 in. in lengths ranging from 3 to 40 ft.

Insertion loss and attenuation values discussed in this section apply only to rectangular sheet metal ducts made with the lightest gages allowed by SMACNA HVAC duct construction standards. Attenuation for lengths greater than 10 ft is not well documented.

Unlined Round Sheet Metal Ducts. As with unlined rectangular ducts, unlined round ducts provide some natural sound attenuation that should be taken into account when designing a duct system. Compared to rectangular ducts, round ducts are much more rigid and thus breakout, especially at low frequencies, is significantly less than that from other shapes. Though this reduces attenuation values for low-frequency sound propagating down the duct, it better protects the surrounding space from breakout noise. Table 19 lists sound attenuation values for unlined round ducts (Kuntz and Hoover 1987; Woods Fan Division 1973).

Acoustically Lined Round Sheet Metal Ducts. ASHRAE research project RP-1408 (Reynolds et al. 2018) measured the insertion loss of round sheet metal ducts with 1 and 2 in. thick glass fiber duct liner and diameters of 12 to 48 in. Tables 20 and 21 present the measured insertion loss of acoustically lined round sheet metal ducts as a function of frequency and length of duct lining for a variety of common duct sizes. Insertion loss values for diameters smaller than 12 in. and larger than 48 in. were taken from previous editions of the ASHRAE Handbook. The liner was 4.0 lb/ft³.

Though previous editions of this table indicated values of dB/ft, research projects have shown that attenuation does not vary exactly linearly with length. Therefore, these updated tables show attenuation for lengths as actually tested. For interpolation to various duct lengths and cross sections, using the duct liner attenuation spreadsheet (available as an extra feature for the ASHRAE Handbook Online version of this chapter) is recommended. Note that the total attenuated noise will never be below the flow-generated noise level in the duct, and, because of typical flows and flanking noise in practice, attenuation values are capped at 50 dB.

Rectangular Sheet Metal Duct Elbows. Table 22 displays insertion loss values for unlined and lined square elbows without turning vanes (Beranek 1960). For lined square elbows, duct lining must extend at least two duct widths w beyond the elbow. Table 22 applies only where the duct is lined before and after the elbow. Table 23 gives insertion loss values for unlined radiused elbows. Table 24 gives insertion loss values for unlined and lined square elbows with turning vanes. The quantity fw in Tables 22 to 24 is the midfrequency of the octave band times the width of the elbow (Figure 20) (Beranek 1960; Ver 1983b).

Nonmetallic Insulated Flexible Ducts. Nonmetallic insulated flexible ducts can significantly reduce airborne noise. Insertion loss values for specified duct diameters and lengths are given in Table 25 and in Appendix D of ARI Standard 885. Recommended duct lengths are normally 3 to 6 ft. Take care to keep flexible ducts straight; bends should have as long a radius as possible. Although an abrupt bend may provide some additional insertion loss, the noise associated with airflow in the bend may be unacceptably high. Because of potentially high breakout sound levels associated with flexible ducts, take care when using flexible ducts above sound-sensitive spaces.

Duct Branch Sound Power Division. When sound traveling in a duct encounters a junction, the sound power contained in the incident sound waves in the main feeder duct is distributed between the branches associated with the junction (Ver 1982, 1983b). This division of sound power is called branch sound power division. The corresponding attenuation of sound power transmitted down each branch of the junction is comprised of two components. The first is associated with reflection of the incident sound wave if the sum of the cross-sectional areas of individual branches ΣS_Bi differs from the cross-sectional area of the main feeder duct. The second and more dominant component is associated with energy division according to the ratio of the cross-sectional area of an individual branch S_Bi divided by ΣS_Bi. Values for the attenuation of sound power ΔL_Bi are given in Table 26.

Duct Silencers. Silencers, sometimes called sound attenuators, sound traps, or mufflers, are designed to reduce the noise transmitted from a source to a receiver. For HVAC applications, the most common silencers are duct silencers, installed on the intake and/or discharge side of a fan or air handler. Also, they may be used on the receiver side of other noise generators such as terminal boxes, valves, dampers, etc.

Duct silencers are available in varying shapes and sizes to fit project ductwork. Generally, a duct silencer’s outer appearance is similar to a piece of ductwork. It consists of a sheet metal casing with length commonly ranging from 3 to 10 ft. Common shapes include rectangular, round, elbow, tee, and transitional. Figure 21 shows some duct silencer configurations.

All silencers can be rated for (1) insertion loss, (2) dynamic insertion loss, (3) pressure drop, and (4) self-generated noise in accordance with ASTM Standard E477 test standards. As such, the performance is under rather ideal conditions as seen in Figure 22.

Insertion loss is the reduction in the sound power level at the receiver after the silencer is installed (“inserted”) in the system. Insertion loss is measured as a function of frequency and commonly published in full octave bands ranging from 63 to 8000 Hz.

Dynamic insertion loss is insertion loss with given airflow direction and velocity. A silencer’s insertion loss varies depending on whether sound is traveling in the same or opposite direction as airflow. Silencer performance changes with absolute duct velocity. However, airflow velocity generally does not significantly affect silencers giving a pressure drop of 0.35 in. of water or less, including system effects.

Pressure drop is measured across the silencer at a given velocity. Good flow conditions are required for accurate measurements at both the inlet and discharge of the silencer. The measuring points are usually 2.5 to 5 duct diameters upstream and downstream of the silencer to avoid turbulent flow areas near the silencer and to allow for any static pressure regain. For nonideal installations, with duct elbows or transitions closer than 2.5 to 5 duct diameters, the total system effect will be larger than the laboratory test data.

Airflow-generated self noise is the sound power generated on the receiving side by the silencer when quiet air flows through it. This represents the noise floor, or the lowest level achievable regardless of high insertion loss values. A silencer’s self-generated noise is a function of frequency and internal geometry, and is referenced to specific velocities and airflow direction (forward or reverse). The airflow-generated sound power of the silencer is logarithmically proportional to silencer cross-sectional area. Self noise generally does not vary with silencer length.

There are three types of HVAC duct silencers: dissipative (with acoustic media), fiber-free reactive (no media), and active.

Dissipative Silencers. Dissipative silencers use sound-absorptive media such as fiberglass as the primary means of attenuating sound; mineral wool can be used in high-temperature applications but may contain too much contamination (“shot”) for commercial HVAC applications. Usually, the absorptive medium is covered by perforated metal to protect it from erosion by airflow. If internal silencer velocities are high (faster than 4500 fpm), media erosion may be further reduced by a layer of material such as fiberglass cloth or polymer film liner placed between the absorptive media and the perforated metal. Dissipative silencers may be supplied as hospital-grade or as film-lined silencers that include special polymer film linings to prevent contamination of the airstream by acoustical media fibers and prevent particles from the airstream from getting into the media. These silencers are commonly used in hospitals, pharmaceutical facilities, cleanrooms, and other places where indoor air quality is of paramount concern. Consult manufacturers for construction and testing performance details.

Dissipative silencer performance is primarily a function of silencer length; airflow constriction; number, thickness, and shape of splitters or centerbodies; and type and density of absorptive media. The absorptive media allows dissipative silencers to provide significant insertion loss performance over a wide frequency range.

Insertion loss performance does not necessarily increase linearly with silencer length; for a given length, silencer designs can produce varying insertion loss and pressure drop data. Even at the same pressure drop and length, silencers can be configured to provide varying insertion loss performance across the frequency spectrum.

Reactive Silencers. Reactive silencers are constructed only of metal, both solid and perforated, with chambers of specially designed shapes and sizes behind the perforated metal that are tuned as resonators to react with and reduce sound power at certain frequencies. The outward appearance of reactive silencers is similar to that of their dissipative counterparts. However, because of tuning, insertion loss over a wide frequency range is more difficult to achieve. Longer lengths may be required to achieve similar insertion loss performance as dissipative silencers. Airflow generally increases the insertion loss of reactive silencers.

Figure 23 compares insertion loss of dissipative silencers, with and without protective film materials, against that of a reactive silencer for the same pressure drop.

Active Silencers. Active duct silencers, sometimes called noise canceling systems, produce inverse sound waves that cancel noise primarily at low frequencies. An input microphone measures noise in the duct and converts it to electrical signals, which are processed digitally to generate opposite, mirror-image sound signals of equal amplitude. A secondary noise source destructively interferes with the primary noise and cancels a significant portion of it. An error microphone measures residual sound beyond the silencer and provides feedback to adjust the computer model for improved performance.

Because components are mounted outside the airflow, there is no pressure loss or airflow-generated noise. Performance is limited, however, if excessive turbulence is detected by the microphones. Manufacturers recommend using active silencers where duct velocities are less than 1500 fpm and where duct configurations are conducive to smooth, evenly distributed airflow.

Active silencers have significant low-frequency insertion loss, and are self-regulating because, if fan noise levels increase, an active silencer can increase performance to compensate for the increased source noise. Mid- and high-frequency insertion loss is minimal, however, so if required, combinations of active (for low-frequency components) and passive (for mid- and high-frequency components) can be used to achieve insertion loss over a wide frequency range.

Test Standard. Data for dissipative and reactive silencers should be obtained from tests consistent with the procedures outlined in ASTM Standard E477. (This standard has not been verified for determining performance of active silencers.) Because insertion loss measurements use a substitution technique, reasonably precise insertion loss values can be achieved . Airflow-generated noise values can be obtained with similar accuracy. Round-robin tests conducted from 2015 to 2016 from five ASTM E477 participating laboratories have shown that the revised test method (2020 and later) will yield insertion loss and self-generated noise reproducibility standard deviations of less than 3 dB in all eight octave bands. (For normal distribution, uncertainty with a 95% confidence interval is about two standard deviations.)

Silencer Selection Issues. When selecting a duct silencer, consider the following: