Adding HVAC Noise Sources: A Logarithmic Decibel Calculation
Define the quantity before using the calculator
A fan path, a diffuser and nearby equipment can each contribute sound at an occupied position. Their decibel values cannot be added as ordinary numbers. A useful calculation starts by making each input comparable: the same acoustic quantity, receiver position, frequency band or weighting, operating state and averaging basis.
Use sound-pressure levels when estimating the level at a receiver. Equipment sound-power levels require a propagation calculation before they become receiver sound-pressure contributions. The OSHA Technical Manual distinguishes these quantities and gives the logarithmic summation equation. A supplier’s sound-power value and a room sound-meter reading cannot share one addition column.
Add energy-equivalent values, then convert back
For independent, uncorrelated sound contributions, L_total = 10 × log10[Σ10^(L_i/10)]. The logarithm is base ten. All L_i values must use the same reference and compatible metric. This energy-summation approach is appropriate for many broadband HVAC estimates; coherent tonal sources can require a phase-sensitive assessment.
A useful calculation check is that two equal 50 dB contributions produce 53.01 dB. A 50 dB source combined with a 40 dB source gives 50.41 dB. The quieter contribution is small but still present. These are numerical consequences of the formula; a 3 dB increase does not mean perceived loudness has doubled.
Worked example: three paths at one receiver
Assume three independent source-only sound-pressure contributions in the 500 Hz octave band at the same occupied receiver: fan duct path 52 dB, diffuser-generated sound 49 dB and nearby equipment 46 dB. Assume these have been measured or predicted without counting a shared background more than once. No room criterion is assumed in this example.
Normalize to the largest value to make the arithmetic readable: L_total = 52 + 10 × log10[1 + 10^((49 − 52)/10) + 10^((46 − 52)/10)]. The relative energy terms are 1.000, 0.5012 and 0.2512. Their sum is 1.7524, giving 54.44 dB in that octave band.
The arithmetic sum, 147 dB, has no physical meaning for these contributions. Their arithmetic average, 49 dB, also fails as a combined level. Keeping the calculation band-by-band makes the result useful for comparing with a frequency-based room criterion or locating a dominant frequency region.
Estimate the value of treating one source
Now assume an effective treatment reduces the fan-path contribution by 6 dB at this receiver, from 52 to 46 dB, while the other two contributions remain unchanged. The revised total is 10 × log10[10^(46/10) + 10^(49/10) + 10^(46/10)] = 52.02 dB.
The room-band improvement is 54.44 − 52.02 = 2.42 dB. A 6 dB improvement to one path therefore produces a smaller change in the combined result because the remaining paths still contribute. This example assumes the treatment truly delivers that receiver-path reduction; a catalogue attenuation figure alone does not demonstrate it.
For comparison, reducing only the quietest contribution from 46 to 36 dB gives a total of 53.84 dB, an improvement of about 0.60 dB. Use this sensitivity check before spending effort on a minor source. Recalculate each affected frequency band if the proposed treatment has a frequency-dependent effect.
A 6 dB reduction in one source contribution produces a 2.42 dB reduction in this combined octave-band result.
Subtract background on the same linear basis
A measurement taken with one item running may contain unrelated room noise. If the source-plus-background level is L_on and representative background alone is L_background, the source-only estimate is 10 × log10[10^(L_on/10) − 10^(L_background/10)]. Both observations must use the same position, metric and representative conditions.
For a separate illustrative measurement, 55 dB with equipment on and 49 dB background gives 53.74 dB for the equipment contribution. Subtracting 49 directly from 55 would produce a 6 dB difference between readings, not the source-only sound level.
Brüel & Kjær’s Environmental Noise guide explains this correction and warns that background too close to the total makes a reliable source estimate difficult. If the on and background levels are nearly equal, small measurement changes dominate the subtraction. Apply the governing test method’s background criteria; do not force a precise result from unsuitable data.
Avoid mixing bands, weighting and time
An overall A-weighted level is a different result from an individual unweighted octave-band level. To derive an A-weighted estimate from bands, apply the appropriate band corrections and then sum the corrected band energies. Do not add an overall dBA figure to one octave-band value from another source.
The example combines sources operating together. Equipment that operates at different times requires a time-based assessment when the target metric is an equivalent continuous level. Document the duty cycle rather than assuming all sources are simultaneous or quietly averaging decibel readings.
Turn the calculation into a field decision
Prepare a receiver plan, source-state log and frequency-band worksheet. Use appropriate calibrated instruments and retain the background observations. Where safe and operationally permitted, controlled source-on/source-off tests can help separate contributions; note any control response that changes the other sources during those tests.
Compare the verified combined spectrum with the actual project requirement. A single calculated band or overall number cannot establish every acoustic criterion. The useful deliverable is a clear ranking of contributions, a realistic predicted benefit from treatment, and repeat measurements showing what changed.
Sources and further reading
Continue your engineering workflow
Continue with the HVAC Duct Design and Sizing Handbook for a broader duct-design workflow.
Comments