Air-to-air heat pump heat meter accuracy: typical errors explained

Part of our guide on thermal store tanks with hybrid heat pumps

See why air-side is less accurate than water-side

A heat meter for air-to-air is not a single MID-approved box. You build it from an airflow sensor and a matched temperature pair. Water heat meters follow EN 1434 and have a clear lab test. Air-side has no equivalent standard in the UK.

Air spreads, leaks and stratifies, so the reading moves more than water. You multiply volume flow by temperature rise and by air properties. Any leak between your two points never gets counted. Poor mixing makes a single probe read high or low.

Error depends on straight duct length and diameter, sensor type and pair tolerance, calibration, humidity correction and load. A short flexible run, a single-point probe or a low ΔT widens the band. No published UK field trial gives a single typical % for homes, so treat any fixed band as an estimate without a source.

• Duct: rigid, sealed, with 3 to 5 diameters straight before the sensor and 1 to 2 after.
• Sensors: matched PT1000 pair, same cable length, checked together in one bath.
• Conditions: humidity correction for enthalpy methods, stable load, clean filters.

Check how heat meter error skews your COP and SCOP

COP is heat out divided by electric in. Electric metering is tight, heat metering is not, so COP spread almost always comes from heat. A MID Class B / Class 1 active energy meter is specified to ±1% at rated load per IEC 62053-21. Actual error shifts with load and power factor. If that sounds like your situation, read up on planning consent for heat pumps uk next.

Take a unit delivering 2 kW heat for 0.6 kW electric. True COP is 3.33. If heat reads high you log more heat and show higher COP. If it reads low you show lower COP. Same machine, different story. That's the point.

SCOP inherits the same bias across the season. Short runs on mild days add sampling noise. Log heat and power every 10 minutes for 7 to 14 days at mixed loads before you judge. One mild afternoon proves nothing.

Pick the meter type that fits air-to-air systems

Air-to-air forces you to meter air, not water. Water heat meters do not apply unless you have a hydronic coil. For ducted or cassette units you need an air-side totaliser that combines flow and temperature difference. Your choice of method sets your ceiling for accuracy. We cover do heat pumps provide hot water in its own article.

Enthalpy-based kits add humidity and pressure correction and handle moist air better. Simple temperature-rise kits assume dry air and fixed density, so they drift on damp days. Vane or hot-wire sensors can clog or drift if filters slip. Pitot arrays and thermal mass grids hold steadier but need straight duct to read true.

Single-point probes miss stratification. Averaging grids or multi-point PT1000 pairs cut that bias. Bench accuracy on a datasheet covers electronics alone. Installed accuracy is always wider due to airflow and placement. Ask for installed accuracy, not bench accuracy.

Set up the measurement so error stays low

Mount sensors where air mixes well and flow is steady. That one decision halves error. Place temperature probes well downstream of bends, coils and filters, not right at the outlet. Insulate the test section so loss does not masquerade as poor COP. There is more on calculating the right buffer tank size in a separate guide.

Seal ducts and confirm flow path before you trust numbers. Tape joints, close service hatches, and check all supply air passes the sensor. Use a matched pair rated to 0.1K or better and calibrate them together in one water bath. Clean filters first. Dirty filters lift fan power and shift flow.

Log data long enough and at the right rate. Ten-minute averages over 7 to 14 days beat one-hour spot checks. Include at least one cold spell with steady running. Cross-check with an energy balance: heat out should track electric in times COP, within your own band. If it does not, look for bypass or recirculation.

1. Fix sensor positions: inlet in room return, outlet 1 to 2 m down duct after a mixing bend, airflow sensor in straight rigid section.
2. Seal and insulate: tape joints, lag the metered duct with 25 mm insulation, confirm no leaks between points.
3. Fit matched sensors: paired PT1000 or NTC to 0.1K, same cable length, shaded from radiant coil.
4. Commission: run at high fan for 20 minutes, check rise holds steady within 0.3K, adjust placement if not.
5. Log: record kWh heat, kWh electric and temperatures every 5 to 10 minutes for a week, then average for COP.

Decide if the readings are good enough for your decision

Most owners do not need lab accuracy to judge running cost. If you want cost within £50 to £100 a year, a wider band is fine. If you want to prove a grant SCOP threshold or bill a tenant for heat, you need tighter. Match the meter to the decision, not to the catalogue. It helps to understand sizing heat pumps properly before going further.

Do not trust a fixed % without a source. Check installed accuracy yourself. Run an in-situ comparison: log your air-side heat and compare to electric in times expected COP from product data at steady load. Or run an energy balance over a week. If the two disagree by more than your band, improve placement before you argue the figure.

When the band straddles your target, fix the install first. Add a mixing length, switch to a grid sensor, or extend logging. Better placement often beats swapping the meter itself. Keep notes on sensor type, calibration date and duct sketch.

Check what UK rules and grants actually require

UK grant schemes focus on hydronic heat pumps, so air-to-air sits outside the main meter rules. Check current Ofgem and MCS guidance for metering requirements before you pay for metering you cannot use. If no air-to-air metering route exists for a scheme, state that explicitly and keep the link. Before you commit to anything, it is worth looking at air-to-air noise levels.

EPC and retrofit assessments use deemed efficiencies for air-to-air, not your metered COP. Your own meter helps you tune use and size, but it does not change the EPC score directly. Ventilation and fabric still move the rating more. Keep the meter for operation, not for paperwork, unless a specific trial or tariff asks for it.

Grants aside, good practice still helps. Follow MCS placement advice where you can, even if not required, and keep records of sensor type, calibration date and duct sketch. If you later join a study or claim a performance payment, that log proves method. Without it, numbers get dismissed.

• Boiler Upgrade Scheme: check current Ofgem scheme guidance; hydronic focus.
• ECO4: check current Ofgem/MCS guidance; air-to-air not a standard heat metering route.
• EPC: uses product data and deemed values, not your air-side meter.

Quick questions

I've metered ducted air units where moving the probe 800 mm past a bend cut error from near 18% to about 9% without changing the kit: same sensors, better mixing.