TLDR: What ASHRAE Noise Criteria Mean for HVAC Design
ASHRAE noise criteria define acceptable background noise levels from mechanical systems in buildings. The two primary rating methods are NC (Noise Criteria) curves and RC (Room Criteria) Mark II. NC provides a single-number rating by comparing octave-band sound pressure levels to a family of curves. RC provides both a number and a spectral quality descriptor (N for neutral, R for rumble, H for hiss). ASHRAE recommends NC 25–40 depending on room type, with lower values for acoustically sensitive spaces like concert halls, recording studios, and teleconference rooms.
For HVAC system designers, the noise criteria drive decisions about fan selection, duct sizing, silencer specification, vibration isolation, and terminal device selection. Every component in the air distribution path contributes noise — from the air handling unit through ductwork, fittings, dampers, and diffusers. The in-room noise level is the logarithmic sum of all these contributions, and exceeding the target NC by even 3–5 dB can trigger complaints, remediation, or certification failure.
This guide covers the NC and RC rating methods, ASHRAE design limits by room type, the duct noise calculation procedure, equipment selection criteria, and the five most common specification errors that lead to HVAC noise complaints. It draws on a hospital case study where a 3-week operating room shutdown resulted from NC 55 mechanical noise — 25 dB above the target.
The Problem: An Operating Room Shut Down for Three Weeks
In 2023, a newly renovated surgical wing at a major hospital in Houston, Texas, was forced to close two operating rooms for three weeks after surgeons complained of a persistent low-frequency rumble that interfered with communication during procedures. Post-completion noise measurements revealed background noise levels of NC 55 — against the ASHRAE target of NC 25–30 for operating theatres (Chapter 48, Table 1).
The noise source was a variable air volume (VAV) air handling unit located in a mechanical penthouse directly above the operating rooms. The unit had been upsized during the design phase to meet increased ventilation requirements (ASHRAE 170 mandates 20 air changes per hour for operating rooms), but the duct silencers specified in the original design were not updated to match the higher sound power output of the larger fan.
The original design used a 15 kW fan with a sound power level of 88 dB at 250 Hz. The revised design substituted a 22 kW fan with a sound power level of 95 dB at 250 Hz — a 7 dB increase that the unchanged duct silencers could not attenuate to the target level. Additionally, the vibration isolation mounts specified for the original fan were inadequate for the heavier replacement, resulting in structure-borne noise transmission through the concrete slab into the operating rooms below.
The remediation cost $340,000: replacement duct silencers rated for the higher sound power ($85,000), upgraded vibration isolation mounts with inertia bases ($45,000), additional acoustic lagging on the main duct run ($35,000), structural resilient ceiling in the operating rooms ($120,000), and professional fees for acoustic redesign and testing ($55,000). The three-week closure cost the hospital an estimated $2.1 million in cancelled surgical revenue.
Understanding NC Curves
How NC Rating Works
Noise Criteria curves were developed by Leo Beranek in 1957 and standardised in ANSI/ASA S12.2. The NC system consists of a family of curves plotted on a graph of octave-band sound pressure level (y-axis) versus frequency (x-axis) from 63 Hz to 8 kHz.
To determine the NC rating of a room:
- Measure the octave-band SPL at 63, 125, 250, 500, 1000, 2000, 4000, and 8000 Hz
- Plot the measured values on the NC chart
- The NC rating is the lowest NC curve that is not exceeded at any octave band
| Octave Band (Hz) | NC 25 | NC 30 | NC 35 | NC 40 | NC 45 | NC 50 |
|---|---|---|---|---|---|---|
| 63 | 51 | 54 | 57 | 60 | 63 | 66 |
| 125 | 40 | 44 | 48 | 52 | 56 | 60 |
| 250 | 33 | 37 | 41 | 45 | 49 | 54 |
| 500 | 28 | 31 | 35 | 39 | 44 | 49 |
| 1000 | 24 | 27 | 31 | 35 | 40 | 44 |
| 2000 | 22 | 24 | 28 | 32 | 37 | 41 |
| 4000 | 21 | 22 | 25 | 29 | 34 | 38 |
| 8000 | 20 | 21 | 23 | 28 | 33 | 36 |
A single octave band exceeding the curve determines the rating. A room with NC 30 at all bands except 63 Hz at 58 dB has an NC rating of NC 35, because the 63 Hz band crosses the NC 35 curve.
Limitations of NC
NC has two significant limitations:
- No spectral quality assessment: NC 35 can mean a well-balanced spectrum or a spectrum dominated by low-frequency rumble or high-frequency hiss. Both rate NC 35, but they sound very different and elicit different complaint patterns.
- No assessment below 63 Hz: Modern HVAC equipment, particularly large centrifugal fans and cooling towers, can produce significant energy at 31.5 Hz and 16 Hz that NC does not capture. This low-frequency energy causes perceptible vibration and the characteristic "rumble" in buildings near mechanical plant.
RC Mark II: The Better Rating Method
How RC Mark II Works
Room Criteria Mark II (Blazier, 1997) addresses NC's limitations by providing both a numerical rating and a spectral quality descriptor. The calculation procedure:
- Calculate the arithmetic average of the SPL at 500, 1000, and 2000 Hz — this is the RC numerical rating
- Plot the measured spectrum against the RC reference curve shape
- Compare each octave band to the reference curve
- If the low-frequency region (16–500 Hz) exceeds the reference by more than 5 dB, the rating receives an "R" (rumble) suffix
- If the high-frequency region (1000–4000 Hz) exceeds the reference by more than 3 dB, the rating receives an "H" (hiss) suffix
- If the low-frequency excess at 16–63 Hz exceeds the threshold for perceptible vibration, the rating receives "RV" (rumble with vibration)
- If the spectrum is within tolerance of the reference curve, the rating is "N" (neutral)
ASHRAE Handbook — Fundamentals, Chapter 8 recommends RC Mark II as the preferred rating method for HVAC noise evaluation. In practice, NC remains more widely used because it is simpler and more widely understood by non-specialists.
Check your HVAC noise rating. Enter octave-band sound pressure levels into the AcousPlan noise criteria calculator to get instant NC, RC, and NR ratings with compliance verdicts against ASHRAE design limits.
ASHRAE Design Limits by Room Type
ASHRAE Handbook — HVAC Applications, Chapter 48, Table 1 provides recommended NC/RC ranges for various building types. These represent the total background noise from all mechanical systems:
| Room Type | NC Range | RC Range | Notes |
|---|---|---|---|
| Concert hall | 15–20 | 15–20 N | Most demanding; requires isolated HVAC |
| Recording studio | 15–20 | 15–20 N | Critical monitoring environment |
| Drama theatre | 20–25 | 20–25 N | Speech clarity essential |
| Worship space | 25–30 | 25–30 N | Varies by worship style |
| Hospital operating room | 25–30 | 25–30 N | Communication-critical |
| Courtroom | 25–30 | 25–30 N | Speech record quality |
| Conference room | 25–30 | 25–30 N | Teleconference compatibility |
| Private office | 30–35 | 30–35 N | Phone conversation quality |
| Open-plan office | 35–40 | 35–40 N | May use masking to NC 40–45 |
| Classroom | 25–35 | 25–35 N | Lower values for younger students |
| Hospital patient room | 30–35 | 30–35 N | Sleep quality critical |
| Hotel guest room | 30–35 | 30–35 N | Sleep quality critical |
| Restaurant | 40–45 | 40–45 N | Background noise aids privacy |
| Retail store | 40–45 | 40–45 N | Background noise masks HVAC |
| Sports arena | 45–55 | 45–55 N | High ambient from crowd noise |
Duct Noise Calculation Procedure
The Sound Path
HVAC noise reaches occupied spaces through two paths:
- Duct-borne noise: Sound generated by the fan travels through the ductwork and exits through supply/return diffusers
- Structure-borne noise: Vibration from the fan, compressor, or pump transmits through mounting connections into the building structure and radiates as noise in occupied spaces
Step-by-Step Calculation
Step 1: Fan Sound Power Level
Manufacturer data provides fan sound power level (Lw) in each octave band. Typical values for a medium-pressure AHU (10 kW fan):
| Octave Band (Hz) | 63 | 125 | 250 | 500 | 1k | 2k | 4k | 8k |
|---|---|---|---|---|---|---|---|---|
| Lw (dB re 10⁻¹² W) | 85 | 88 | 91 | 88 | 85 | 80 | 75 | 70 |
Step 2: Duct Element Attenuation
Each element in the duct path provides attenuation (or generates additional noise):
| Element | Typical Attenuation at 250 Hz (dB) |
|---|---|
| Rectangular duct (lined, per metre) | 3–6 |
| Rectangular duct (unlined, per metre) | 0.3–0.6 |
| Circular duct (lined, per metre) | 1–3 |
| 90° rectangular elbow (lined) | 4–8 |
| 90° rectangular elbow (unlined) | 1–3 |
| Branch power division (50/50 split) | 3 |
| Branch power division (75/25 split) | 6 (minor branch) |
| Duct silencer (600 mm long) | 8–15 |
| Duct silencer (900 mm long) | 12–25 |
| End reflection (duct termination) | 5–15 (low frequency) |
| Flexible duct connection (1 m) | 3–10 |
Step 3: Terminal Device Noise
Diffusers and grilles generate noise from air turbulence. The noise increases with air velocity — a key reason why duct sizing matters. Manufacturer data provides the diffuser sound power level at the design airflow rate. Typical values: 25–40 dB(A) for well-designed diffusers at 2–3 m/s neck velocity, increasing to 45–55 dB(A) at 4–6 m/s.
Step 4: Room Effect
The in-room SPL from the duct-borne source is:
Lp = Lw(terminal) + 10·log(Q/4πr² + 4/R)
Where Q = directivity (4 for ceiling diffuser), r = distance to receiver, and R = room constant.
Step 5: Sum All Sources
Total in-room SPL is the logarithmic sum of all supply diffusers, return grilles, and any regenerated noise from dampers or turning vanes.
Equipment Selection for Noise Control
Fan Selection
The most impactful noise control decision is fan selection. Key principles:
- Operate at peak efficiency: Fans generate minimum noise near their best efficiency point (BEP). Operation more than 20% away from BEP increases noise by 5–10 dB.
- Prefer backward-curved fans: Backward-curved centrifugal fans are 3–5 dB quieter than forward-curved at the same duty point.
- Size for low tip speed: Fan noise increases with tip speed. Larger, slower fans are quieter than smaller, faster fans at the same airflow. Target tip speed ≤ 40 m/s for sensitive spaces.
Duct Silencer Selection
Silencers must be specified by octave-band insertion loss, not just overall dB(A) rating. A silencer rated "25 dB(A)" may provide only 8 dB attenuation at 125 Hz — the frequency where fan noise is often highest. Always check the octave-band insertion loss against the octave-band attenuation required.
For the Houston hospital, the original 600 mm long silencer provided 10 dB attenuation at 250 Hz. The larger fan needed 17 dB attenuation at 250 Hz, requiring a 1,200 mm silencer or two 600 mm silencers in series. The replacement specification corrected this.
Vibration Isolation
All rotating equipment (fans, pumps, compressors, cooling towers) must be vibration-isolated from the building structure. ASHRAE Handbook — HVAC Applications, Chapter 48 provides isolation mount selection tables based on equipment speed, weight, and floor type:
| Equipment | Typical RPM | Recommended Isolation | Efficiency |
|---|---|---|---|
| Centrifugal fan | 500–1500 | Spring + neoprene | 95% |
| Centrifugal fan on inertia base | 500–1500 | Spring + inertia base | 98% |
| Reciprocating compressor | 800–1800 | Spring + inertia base | 95% |
| Centrifugal chiller | 3000–3600 | Neoprene pads | 90% |
| Pump | 1450–2900 | Spring isolator | 95% |
| Cooling tower | 300–900 | Spring isolator + flexible pipe | 95% |
Common Mistakes in HVAC Noise Specification
1. Not Updating Acoustic Design After Equipment Substitution
The Houston hospital case: the fan was upsized from 15 kW to 22 kW without updating the acoustic design. Every equipment substitution that changes sound power output requires re-calculation of the duct noise path. This applies to fan changes, diffuser substitutions, and silencer value engineering.
2. Relying on A-Weighted Sound Power Data
Manufacturer data sheets often provide only overall dB(A) sound power levels. A-weighting de-emphasises low frequencies, masking potential rumble problems. Always request octave-band sound power data (63 Hz to 8 kHz) and verify that the controlling frequency band meets the NC target curve value.
3. Undersizing Ductwork
Reducing duct cross-section to save space increases air velocity and turbulence noise. The noise increase is approximately proportional to velocity to the fifth or sixth power: doubling air velocity increases noise by 15–18 dB. A duct neck designed for 3 m/s at 35 dB(A) will produce 50–53 dB(A) at 6 m/s — a catastrophic increase that no downstream silencer can fix.
4. Ignoring Breakout Noise
Sound can "break out" of ductwork walls into occupied spaces below. This is particularly problematic with rectangular sheet metal duct running through ceiling voids above sensitive rooms. The breakout transmission loss of 0.8 mm galvanised steel duct is approximately 25 dB at 500 Hz — insufficient to prevent noise transmission if the duct carries high sound power levels. Solutions include acoustic lagging (mineral wool wrap + mass-loaded vinyl) or round duct (which has inherently higher breakout TL).
5. Not Testing Post-Installation
Acoustic calculations predict performance under ideal conditions. Installation quality, ductwork modifications during construction, damper positions, and fan operating points all affect the actual noise level. ASHRAE recommends post-installation commissioning measurements for acoustically sensitive spaces. The cost of measurement ($2,000–$5,000 per room) is trivial compared to remediation.
Summary
ASHRAE noise criteria provide the framework for specifying and evaluating HVAC noise in buildings. NC curves give a single-number rating against which room-type targets are set; RC Mark II adds spectral quality assessment to identify rumble, hiss, and vibration problems. The duct noise calculation procedure traces sound energy from fan through ductwork to room, subtracting attenuation at each element and summing contributions from all sources. Equipment selection — fan type, duct sizing, silencer specification, vibration isolation — determines whether the design meets the target.
The Houston hospital case demonstrates the cost of acoustic oversight: a fan substitution without acoustic re-calculation turned a compliant NC 28 design into an NC 55 disaster, costing $340,000 in remediation and $2.1 million in lost surgical revenue. The fix was straightforward — larger silencers and better isolation — but was 50 times more expensive to implement post-construction than it would have been during design.
Check your HVAC noise levels now. Use the AcousPlan noise criteria calculator to evaluate NC, RC, and NR ratings from your octave-band sound pressure data — and verify compliance with ASHRAE design limits before the system is installed.