HVAC Noise Control — The Ductwork Mistake Ruining Your NC 30 Spec | AcousPlan

# HVAC Noise Is Ruining Your NC 30 Spec — The Ductwork Mistake in Every Building

Every acoustic consultant who has worked in commercial construction has seen it: the specification says NC 30–35, the mechanical engineer says "yes," the building opens, and the offices are NC 42–48. Nobody is surprised except the client. The mechanical engineer points to the equipment schedule and says the plant meets the specified sound power levels. He is correct. The ductwork between the plant and the space is the problem, and it was never in anyone's acoustic scope.

This is the most consistent acoustic failure in commercial construction, and it is preventable. The ductwork noise mechanism is well understood, the ASHRAE calculation methods are published, and the specification language to prevent it is straightforward. Yet most commercial projects do not include ductwork acoustic sizing in their mechanical specification, and nobody checks until the building opens.

Here is exactly what is happening in your building, why it matters, and how to prevent it.

The NC Curve: What It Is and Why It Matters

NC (Noise Criteria) curves are octave-band noise spectra used to characterise acceptable background noise levels in occupied spaces. Each NC curve number corresponds to a set of maximum octave-band sound pressure levels from 63 Hz to 8000 Hz. The NC rating of a measured spectrum is the highest NC curve tangent to the measured data.

ASHRAE Handbook of Fundamentals Table 1 (Chapter 48, 2019 edition) gives recommended NC ranges by space type:

The NC curve differs from simple dBA measurement in an important way: it is octave-band specific. A space can meet a dBA target but fail NC because a single frequency band is too prominent. A supply air diffuser generating a hiss at 2000 Hz can push a room from NC 35 to NC 42 while the overall A-weighted level changes by only 1–2 dB.

This frequency specificity matters because different frequency ranges are generated by different duct system elements, and each requires different treatment.

Where HVAC Noise Comes From: The Five Sources

1. Fan Noise at Source

Air handling unit fans generate broadband noise with tonal peaks related to blade passage frequency (BPF). A 6-blade fan rotating at 900 RPM has BPF = 6 × 900/60 = 90 Hz — in the low-frequency region where duct lining provides almost no attenuation.

Fan noise is typically attenuated by silencers installed immediately downstream of the AHU. ASHRAE recommends minimum 2m of duct liner plus a reactive silencer before the first branch. Where this is specified correctly, fan noise contribution to terminal spaces is usually minor — the silencer insertion loss (typically 20–35 dB in the 125–500 Hz bands) is usually sufficient.

Where it fails: silencers are omitted during value engineering ("the equipment is quiet, we don't need silencers"), or silencers are specified but installed in a location that allows regenerated noise from bends and transitions to bypass them.

2. Duct Self-Noise from Air Velocity — The Main Culprit

This is the primary source of HVAC noise problems in commercial construction. Turbulent airflow in ductwork generates broadband noise, particularly in the 250–4000 Hz range. The self-noise power level generated in a straight duct section is approximately:

Lw = 10 + 50 log₁₀(V) + 10 log₁₀(A)

where V is duct air velocity in m/s and A is duct cross-sectional area in m².

At V = 5 m/s (common main duct velocity in commercial systems), A = 0.06 m² (300mm × 200mm duct):

Lw = 10 + 50 log₁₀(5) + 10 log₁₀(0.06)
   = 10 + 50 × 0.699 + 10 × (-1.222)
   = 10 + 34.95 - 12.22
   = 32.7 dB re 10⁻¹² W (per octave band, centre frequency 500 Hz)

At V = 8 m/s (aggressive but not unusual commercial design):

Lw = 10 + 50 log₁₀(8) + 10 log₁₀(0.06)
   = 10 + 50 × 0.903 + (-12.22)
   = 10 + 45.15 - 12.22
   = 42.9 dB

The velocity increase from 5 to 8 m/s — 60% increase — adds 10 dB to self-noise. The relationship is V⁵ — velocity noise scales as the fifth power of velocity. This is why velocity control is so much more effective than duct lining: halving the velocity reduces self-noise by 50 log₁₀(0.5) = 15 dB. Adding 25mm of fibreglass liner reduces self-noise by 1.5–3.0 dB.

3. Turbulence at Fittings: Bends, Transitions, and Branches

Every fitting in the duct system — elbow, tee, transition, flex connection — generates additional turbulence and noise above the straight-duct self-noise level. ASHRAE provides velocity pressure correction factors for each fitting type. The key values:

A mitered elbow without turning vanes 600mm upstream of a diffuser is one of the most common sources of high-velocity duct noise in commercial HVAC systems. The turbulence from the elbow feeds directly into the diffuser neck, where it is re-radiated into the room. This single configuration accounts for a significant fraction of "loud HVAC" complaints in office buildings.

The fix is simple: minimum 3–5 duct diameters of straight duct between the last fitting and any terminal device. For a 300mm diameter neck, this means 900mm–1500mm of straight duct before the diffuser. This constraint is almost never shown on duct shop drawings because it is not a default requirement — it must be explicitly specified.

4. Terminal Unit Noise (VAV Boxes)

Variable air volume terminal units are the single largest source of duct noise variation in commercial HVAC systems. A VAV box rated for 800 m³/h maximum has a sound power level at 500 Hz of approximately 42–48 dB at maximum flow. At minimum flow (20% of design), this drops to 28–34 dB. The difference of 14 dB means the space is quiet when the VAV is nearly closed and loud when it is fully open.

Spaces specified for NC 35 often achieve NC 35 in winter (heating mode, VAV mostly open, but supply temperature warm so low velocity) and NC 45 in summer peak cooling (full cooling flow, VAV wide open, high velocity through the terminal unit).

The mistake is designing for average or minimum conditions rather than peak operating conditions. The acoustic specification must state: "HVAC noise criteria to be met under all operating conditions from minimum to maximum design airflow." This forces the mechanical engineer to size terminal units so that NC compliance is maintained at peak flow, which typically means selecting a larger unit running at lower velocity than the minimum-cost option.

5. Diffuser Noise at the Terminal End

Supply diffusers generate noise proportional to flow rate and throw velocity. Manufacturers publish sound power data for diffusers at different flow rates. A standard 600mm × 600mm four-way ceiling diffuser at 250 m³/h generates approximately NC 35 at 2m distance. At 400 m³/h the same diffuser generates NC 42–45.

The common mistake is using the diffuser noise data from the manufacturer's catalogue at nominal flow rate and not checking the noise at design peak flow. The ASHRAE calculation chain is:

1. Establish design supply airflow per diffuser from cooling load calculation
2. Select diffuser based on throw characteristics (not noise)
3. Look up diffuser sound power from manufacturer data at design flow
4. Calculate room NC using ASHRAE path-correction method

The ASHRAE Ductwork Noise Calculation: How to Do It Correctly

ASHRAE Handbook HVAC Applications Chapter 48 (2019) provides the method for calculating room NC from ductwork system noise. The calculation proceeds octave band by octave band (63 Hz through 8000 Hz):

Lp = Lw + 10 log₁₀(Q/(4πr²) + 4/R) + 10

where Lp is sound pressure level in the room (dB), Lw is the total sound power entering the room from the duct system (dB), Q is source directivity, r is distance from the diffuser to the receiver, and R is room constant (m²). The +10 term converts power reference.

For a typical 9m × 6m × 3m office with suspended ceiling and moderate absorption (R ≈ 40 m² at 500 Hz), receiver position 3m from the diffuser (Q = 2 for ceiling diffuser):

Lp = Lw + 10 log₁₀(2/(4π×9) + 4/40) + 10
   = Lw + 10 log₁₀(0.0177 + 0.100) + 10
   = Lw + 10 log₁₀(0.1177) + 10
   = Lw + (-9.29) + 10
   = Lw + 0.71

So for this room at 500 Hz, room sound pressure level ≈ Lw + 0.7 dB. If the target is NC 35 (Lp ≤ 42 dB at 500 Hz per NC 35 curve), then maximum allowable Lw = 41.3 dB at 500 Hz entering the room from the diffuser.

Checking against the four-way diffuser at peak flow of 400 m³/h (Lw ≈ 50 dB at 500 Hz from manufacturer data), the diffuser is generating 8.7 dB more noise than the room can accept. The NC in the room will be approximately NC 43.

To meet NC 35 with this room configuration, the diffuser flow must be reduced to approximately 230 m³/h, or a larger, quieter diffuser selected, or the duct velocity upstream reduced to lower self-noise contribution.

Worked Example: Open-Plan Office Floor

A 500 m² open-plan office floor in a commercial tower. Mechanical specification: NC 40 maximum. Actual HVAC system as installed:

• Main duct velocity: 7.5 m/s (1500 fpm)
• Branch duct velocity: 5.5 m/s (1100 fpm)
• VAV terminal boxes: sized for NC 38 at maximum flow (NC 35 compliant at design flow, NC 38 at peak)
• Supply diffusers: 600mm × 600mm four-way, 12 units at 350 m³/h each
• Duct liner: 25mm fibreglass, 1200mm downstream of each VAV box
• No straight duct length before diffusers (diffusers on 300mm flex connection directly off tee branch)

Combined Lp at 500 Hz:

Lp_combined = 10 log₁₀(10^4.67 + 10^4.47 + 10^3.87 + 10^4.17)
            = 10 log₁₀(46773 + 29512 + 7413 + 14791)
            = 10 log₁₀(98489)
            = 49.9 dB

The NC 40 criterion allows maximum Lp of 46 dB at 500 Hz. The system is generating ~50 dB — NC ≈ 44. Specification missed by 4 NC units despite every individual component appearing to be within reasonable tolerance.

After redesign:

• Main duct velocity reduced to 5.0 m/s (requires 12% larger duct cross-sections, costs approximately $8,000 in additional ductwork)
• Branch velocity reduced to 3.5 m/s
• 600mm straight duct added before each diffuser
• Diffuser flow reduced to 280 m³/h (15 diffusers instead of 12 to maintain total supply airflow)
• VAV boxes upsized by one size (NC 35 at design flow, NC 37 at peak)

The redesign cost: approximately $22,000 additional ductwork and diffuser cost on a $2.5 million fit-out. Less than 1% of project cost. The cost of rectification after handover (replacing 12 diffusers, rerouting ductwork through an occupied building): approximately $85,000.

The Specification Language Nobody Uses

Add this to your acoustic specification, and require the mechanical engineer to respond:

"Ductwork system shall be designed to achieve [NC 35/NC 40] under all operating

conditions from minimum to maximum design airflow. Maximum main duct velocity

shall not exceed [5.0 m/s] and maximum branch duct velocity shall not exceed

[3.5 m/s]. Minimum straight duct length between last fitting and any terminal

device shall be 5 equivalent duct diameters. VAV terminal units shall be

selected and scheduled with sound power data at maximum design airflow,

demonstrating NC compliance at maximum flow condition.

Post-installation noise measurement per ASHRAE Handbook HVAC Applications §48.15

shall be conducted at peak cooling load condition. Measurements to demonstrate

NC ≤ [35/40] in all occupied areas."

The phrase "peak cooling load condition" is critical. Most commissioning measurements are done in mild weather. The building is accepted at NC 35. Summer arrives, cooling load increases, VAV boxes open fully, and the office floor is suddenly NC 45. Without the peak-load measurement requirement, this will not be caught at commissioning.

The Backlink-Bait Finding: Building Refresh Rates Create Noise Spikes

Modern buildings with increased ventilation rates (post-COVID design guidelines recommending 10 L/s/person in offices versus the pre-2020 standard of 7 L/s/person) require higher airflow volumes through the same duct infrastructure. Buildings designed pre-2020 that have been upgraded to higher ventilation rates are running their duct systems at 30–40% higher velocity than the acoustic design anticipated. This is why post-refurbishment noise complaints in older buildings are so common — the acoustic design was for a lower airflow requirement, and nobody checked whether the acoustic performance was maintained at the new flow rate.

Any building ventilation upgrade project should include an acoustic reassessment of duct velocities and terminal unit performance before increasing design airflow. The HVAC engineer who designed the original system calculated minimum duct sizes for thermal performance, not acoustic performance — and increasing flow through his minimum-sized ducts pushes noise up by (50 log₁₀(V_new/V_old)) dB.

Run Your HVAC Noise Calculation

Use the acoustic simulation tool to model your room with HVAC noise as a sound source and verify NC compliance before mechanical design is finalised. Cross-reference with NC and NR curve explained for a guide to reading and interpreting noise criteria curves.

Related reading: Noise Criteria NR, NC, RC Explained · Acoustic Design for Architects · Hospital Acoustic Design Guide