From Sound Power to the Neighbour's Window: Judging Heat Pump Noise Under MCS 020
Few domestic retrofit decisions turn on a quantity as small as three or four decibels. Heat pump noise is one of them. A margin of that size, assessed at the correct position in the correct conditions, is frequently the difference between permitted development proceeding without friction and a project stalling for months while a boundary dispute runs its course. The engineering is not complicated, but the vocabulary is frequently muddled — and muddled vocabulary is what turns a straightforward assessment into an argument.
Sound power and sound pressure are different currencies
The figure quoted in a manufacturer's literature is almost always the sound power level, expressed as dB(A) and typically in the range of the mid-forties to low sixties for a domestic air-source unit. Sound power describes the total acoustic energy the machine radiates. It is a property of the equipment, not of the garden, and it is measured in a laboratory under controlled conditions.
Sound pressure is what a person actually hears at a given point in space. It depends on distance, on the acoustic character of the surroundings, on the height of the receiver, and on how much of the sound is reflected back off walls, paving and fencing before it reaches a window. The gap between the two numbers is routinely eight to fifteen decibels in a domestic setting, and in a tight courtyard it can be considerably more.
Confusing the pair is the single most common source of misplaced confidence in this area — on both sides of the boundary. A homeowner reading "58 dB(A) sound power" may conclude the neighbours will hear 58 decibels. They will not; they will hear something lower, and the assessment exists to establish how much lower.
What the assessment position represents
The MCS 020 methodology does not attempt to model every listener everywhere. It selects a defined assessment position — conventionally one metre from the façade of the nearest neighbouring dwelling, at the elevation most exposed to the unit — and predicts the sound pressure level there, having accounted for distance and for the reflective surfaces between source and receiver.
The result is then compared against the threshold the methodology applies, most commonly expressed as 42 dB(A). Where the existing background noise at that position is already higher than the threshold, a comparison against the measured background becomes the relevant test, since a heat pump that is quieter than the ambient noise of a busy road is unlikely to be the source of a nuisance.
The output is a worksheet: unit model, declared sound power, distances and angles, reflection corrections, predicted level at the assessment position, background evidence where used, and a clear statement of compliance. That document is what a planning authority or an enforcement officer will ask for first, and it is worth keeping a signed copy with the commissioning pack rather than in a drawer of loose paperwork.
Why a compliant prediction and an irritated neighbour can coexist
The methodology is sound but it is deliberately simple, and simplicity has limits. Three factors routinely explain the divergence between a passing worksheet and a genuine complaint.
The first is tonal character. A broadband number averages across the spectrum, but the ear is not an averaging instrument. Compressor hum in the low hundreds of hertz and fan blade tones at higher frequencies can be distinctly audible even where the overall level sits comfortably below the threshold. A sound that is tonal and continuous is noticed in a way that a marginally louder but more neutral sound is not.
The second is intermittency. Heat pumps do not run at a constant level. They modulate, they ramp, they stop, and they start again. The change is itself a stimulus, and a unit cycling quietly through the night can be more intrusive than a steady background drone.
The third is timing. Eleven at night in a suburban cul-de-sac is a far quieter environment than two in the afternoon, and the same unit that is inaudible against daytime traffic can be perfectly clear once the road goes quiet. Bedroom windows are open in July and August; that is precisely when a neighbour is most likely to be listening.
Defrost, modulation and the hour that matters
Cold, humid weather produces defrost cycles, and defrost cycles are loud. The unit reverses, the fan frequently runs at higher speed, and the vapour plume is visible from the boundary. A handful of minutes several times a day in January is unlikely to breach a standard written around a five-minute average, but it is memorable — and complaints are made from memory.
Modulation has a subtler effect. A well-set heat pump running at low output for long periods is generally preferable to a unit short-cycling at maximum. However, quiet modes that simply cap the fan speed can reduce noise at the expense of capacity, pushing the unit into longer runtimes or higher flow temperatures. The sensible approach is to commission weather compensation properly so the unit settles at modest output by design, rather than leaning on a quiet-mode button as a remedy after the fact.
A night-time setback of two or three decibels, scheduled to match the hours when the neighbourhood is asleep and the bedrooms are occupied, is a legitimate and effective measure. It should be set at commissioning, evidenced in the controller settings, and explained to the neighbour before it is needed.
Siting: the cheapest decibels available
Distance remains the most reliable attenuator. In free-field conditions the level falls by roughly six decibels for every doubling of distance from source to receiver; against a reflective façade the fall is less generous, but the principle holds. Moving a unit two metres further from a boundary is often worth more than any accessory.
Corner positions are the enemy. A unit tucked into the internal angle between two walls radiates into a reflective trap, and the sound that would otherwise disperse is directed towards the one place it should not go. Siting beneath a window — the installer's convenience, not the acoustic optimum — deserves to be challenged at survey stage.
Discharge direction matters as much as position. An airstream aimed at a close-boarded fence, a neighbouring wall or a hard paved area will reflect and carry. Where a screen is used, it needs mass, no gaps at the base, and proximity to the source; a lightweight panel five metres away is decorative rather than acoustic.
When a complaint arrives
Respond promptly and in person. A measured, non-defensive conversation at the boundary resolves more cases than any technical document. Offer to visit at the time of day the noise is most noticeable, and listen from the complainant's own window rather than inviting them to listen from yours.
Keep the compliance file to hand, but do not lead with it. The evidence establishes that you have met the standard; it does not establish that your neighbour is mistaken about being disturbed, and treating the two as the same question tends to harden positions. Where disagreement persists, an independent acoustic survey is inexpensive relative to an enforcement process, and its findings carry weight with a local authority either way.
It is worth remembering that planning compliance and statutory nuisance are separate regimes. Meeting the permitted development noise condition does not immunise an installation against action under the Environmental Protection Act 1990 if a local authority concludes that a nuisance exists in fact. Courts and councils look at duration, frequency, intensity and character, not solely at a worksheet.
Retrofit measures that work, and those that do not
Anti-vibration mounts address structure-borne transmission through brackets and walls; they do nothing for airborne noise, and they are frequently oversold. Compressor acoustic blankets, where approved by the manufacturer and warranty-safe, can take the edge off tonal content. Relocation, re-aiming the discharge, and a properly built close-proximity screen are the measures that consistently deliver. Adding a fence on the boundary line, at distance from the unit, generally achieves little beyond changing which room is affected.
None of this is a reason to avoid heat pumps. It is a reason to treat the noise assessment as design work rather than a form to be completed after the unit has been ordered — and to keep the paperwork, the settings and the neighbour informed from the outset.
