Quick Summary (TL;DR)
Flow and return temperature measurements help diagnose heat-pump operation, but sensor position, attachment and calibration matter as much as the sensor specification. The flow sensor measures water leaving the heat pump; return measures water entering it. Their difference (ΔT) can help assess hydraulic operation, but temperature difference alone does not measure delivered heat or COP. Accurate heat output also requires reliable water-flow measurement and appropriate fluid properties.
Best approach: start with the manufacturer’s built-in readings for diagnostics; use professionally installed matched immersion probes for higher-quality measurements; use carefully insulated surface-mounted probes for non-invasive trends, not billing-grade heat measurement. For Home Assistant, prefer documented native telemetry or certified meters over improvised control wiring.
- Measure at the same hydraulic boundary and label sensor positions clearly.
- Compare sensors together at one stable temperature before installation to assess relative offset.
- Insulate contact sensors from room air and nearby heat sources.
- Never disturb a pressurised heating circuit or refrigerant system without a qualified professional.
- For COP, measure both electrical energy and delivered thermal energy over matching time windows.
What do flow and return temperatures tell you?
In a hydronic air-to-water heat pump, flow temperature is the leaving-water temperature and return is the entering-water temperature at the chosen measurement boundary. Under heating operation, flow is normally warmer than return. The difference changes with water-flow rate, compressor output, emitter demand, defrost cycles, and hydraulic mixing. Do not apply a universal “correct” ΔT: the manufacturer’s commissioning specification and actual operating state determine what is appropriate.
| Reading | What it can reveal | What it cannot prove |
|---|---|---|
| Flow temperature | Water temperature delivered at sensor location | Actual heat output without water flow |
| Return temperature | Water returning at sensor location | Whole-building heat loss |
| Flow minus return (ΔT) | Change in temperature across chosen boundary | COP or energy efficiency by itself |
| Heat-meter thermal kWh | Delivered heat across its metered boundary | Whole-system efficiency if electric meter excludes auxiliaries |
| Electricity meterAliExpress price kWh | Electrical energy consumed by metered circuits | Delivered heat without heat metering |
Sensor types: immersion, strap-on and built-in probes
A manufacturer-installed probe generally provides convenient diagnostics and may be used by the heat pump’s own control system. Its readings are not automatically calibrated to revenue-meter standards. An immersion probe in a correctly positioned pocket or fitting measures fluid temperature more directly, but installation involves opening or modifying the hydronic circuit. A surface-mounted contact probe is easier to retrofit on accessible pipework; its result is sensitive to contact pressure, pipe material, insulation, airflow and thermal lag. Infrared spot readings on shiny copper are especially vulnerable to emissivity errors.
| Method | Advantages | Limitations | Best use |
|---|---|---|---|
| Manufacturer native sensors | Already fitted; often visible via supported gateway | Unknown accuracy, location and update cadence | Fault diagnosis and trends |
| Immersion PT100/PT1000 | Good thermal contact; suitable with proper matched electronics | Requires approved hydraulic installation; sensor-pocket lag | Independent measurement and heat metering |
| Surface PT100/PT1000 or digital contact probe | Non-invasive where permitted; retrofit-friendly | Thermal coupling and ambient bias | Trend monitoring and comparison |
| Clamp-on commercial heat meter probes | Specified as a matched system | Meter-specific mounting rules; not interchangeable | Meter-approved heat calculations |
| IR thermometer | Quick survey with correct emissivity technique | Surface finish and distance errors | Spot checks, not permanent COP |
Where should the sensors be placed?
Choose a consistent measurement boundary. For heat-pump performance, measure flow and return as close as practical to the heat-pump hydronic ports, following manufacturer instructions. A sensor downstream of a buffer tank, hydraulic separator, mixing valve or secondary pump may describe the emitter circuit rather than the heat pump itself. These locations can produce valid readings for different questions, but they must not be mixed in the same heat-output calculation.
- Primary heat-pump circuit: best for measuring heat produced by the appliance when paired with an appropriate primary flow meterAliExpress price.
- Secondary heating circuit: useful for emitter delivery after a buffer or hydraulic separator, but measures a different boundary.
- Domestic hot-water branch: can be useful for operating-state diagnosis; avoid treating a DHW cylinder probe as instantaneous flow temperature.
- Mixed or manifold circuits: measure before and after mixing only when the intended calculation explicitly accounts for mixing.
Have the installer confirm suitable pipe sections, thermal wells, access, water pressure, isolation and any minimum straight-run requirements of the selected heat meter. Avoid locations exposed to sunlight, draughts, adjacent hot pipes or hot pump housings. Do not open live electrical enclosures or pressurised pipework for a DIY sensor retrofit.
Accuracy, resolution and matching: why a 0.5°C offset matters
When ΔT is small, errors in two independent temperature readings can be a large fraction of the calculated difference. If actual flow and return are 35.0°C and 32.0°C, ΔT is 3.0 K. If one probe reads +0.5°C and the other −0.5°C, the displayed ΔT becomes 4.0 K, an error of about 33%. Matching and calibration across the operating range are therefore critical for heat calculations. A sensor’s display resolution of 0.1°C is not proof of ±0.1°C accuracy.
| Actual flow / return | True ΔT | Reported flow / return | Reported ΔT | ΔT error |
|---|---|---|---|---|
| 35.0 / 32.0°C | 3.0 K | 35.5 / 31.5°C | 4.0 K | +33.3% |
| 40.0 / 35.0°C | 5.0 K | 40.3 / 34.7°C | 5.6 K | +12.0% |
| 30.0 / 28.0°C | 2.0 K | 29.8 / 28.2°C | 1.6 K | −20.0% |
For comparative calibration, put both probes in the same stable thermal environment using an appropriate safe test method, log their difference and repeat at more than one representative temperature. A single offset correction at room temperature does not guarantee accurate matching at 45°C. Calibration of a commercial heat-meter pair should follow its certified procedure rather than a generic Home Assistant template.
PT100 vs PT1000 vs DS18B20 for Home Assistant
Platinum resistance temperature detectors (RTDs) such as PT100 and PT1000 are widely used for accurate industrial temperature sensing when combined with suitable lead compensation, signal conditioning and calibration. DS18B20AliExpress price digital probes can be useful for general trends but are not a substitute for a calibrated heat-meter temperature pair. Encapsulation, waterproofing and cable installation are product-specific. An ESP32 interface must use appropriately isolated, low-voltage sensor circuitry and should never be wired into a manufacturer’s proprietary safety or BMS control bus.
| Sensor | Interface | Practical consideration | Recommendation |
|---|---|---|---|
| PT100 | Resistance RTD; suitable RTD converter | Lead resistance important; 3/4-wire techniques help | Professional-grade measurement with calibrated electronics |
| PT1000 | Resistance RTD; suitable RTD converter | Lower relative lead-wire influence than PT100 | Good candidate for independent trends |
| DS18B20AliExpress price | 1-Wire digital | Probe packaging and accuracy vary; bus wiring needs care | Budget trends, not certified heat metering |
| Native heat-pump probe | Vendor-specific | May not expose raw value or accuracy specification | First choice for monitoring without invasive work |
| Certified heat meter pair | Matched sensors and flow meterAliExpress price | Must follow approved installation and calibration | Preferred for defensible thermal kWh |
Calculating heat output correctly
For water near normal heating temperatures, an illustrative instantaneous thermal-power equation is P(kW) ≈ 1.163 × flow(m³/h) × ΔT(K). The coefficient assumes water properties approximated over a limited temperature range. Antifreeze/glycol mixtures require different properties and compatible meter configuration. For example, 0.80 m³/h and a measured 4 K temperature difference give approximately 3.72 kW. This is an instantaneous estimate, not seasonal heat delivery.
| Water flow | ΔT | Illustrative thermal power |
|---|---|---|
| 0.50 m³/h | 3 K | 1.74 kW |
| 0.80 m³/h | 4 K | 3.72 kW |
| 1.20 m³/h | 5 K | 6.98 kW |
| 1.00 m³/h | 2 K | 2.33 kW |
For a proper thermal-energy total, integrate measured heat output over time or use a certified heat meter. A nominal pump-speed setting is not a measured flow rate. Changes in filter restriction, zone valves, bypass flow and pump control can alter flow substantially. Do not calculate COP by dividing heat output inferred from ΔT alone by compressor power.
How to use the readings in Home Assistant
Prefer a manufacturer-supported Home Assistant integration or a supported Modbus gateway exposing flow, return, operating mode and fault state. When using independent sensors, name them by physical location and measurement boundary, not just “flow” and “return”. Keep timestamps aligned and reject stale or unavailable values before deriving ΔT. Record compressor state and defrost periods, which can reverse or distort the apparent temperature relationship.
A basic Home Assistant template for an illustrative difference is shown below. Replace the entity IDs with your own, and treat it as a diagnostic indicator rather than a safety control or certified energy measurement.
template:
- sensor:
- name: "Heat pump primary delta T"
unit_of_measurement: "K"
state_class: measurement
availability: >
{{ has_value('sensor.hp_primary_flow_temperature') and
has_value('sensor.hp_primary_return_temperature') }}
state: >
{{ (states('sensor.hp_primary_flow_temperature') | float -
states('sensor.hp_primary_return_temperature') | float) | round(2) }}A template can be mathematically valid while using stale measurements. If the integration does not expose sample timestamps or quality flags, avoid using this derived sensor for automated plant decisions. The heat pump’s native frost, high-pressure, anti-cycling and flow protections must remain independent.
Surface sensor installation quality: what to check
- Use a manufacturer-approved pipe-surface sensor mounting method and ensure sound, repeatable contact.
- Keep the contact location consistent and document the distance from the heat-pump ports.
- Fit suitable thermal insulation over contact probes where approved to reduce ambient bias.
- Keep sensor cables away from hot surfaces, abrasion, moisture ingress and mains wiring as specified by the manufacturer.
- Verify the correct electrical sensor type and converter configuration; a PT1000 configured as PT100 produces grossly incorrect readings.
- Use identical sampling and filtering windows for flow and return readings.
- Check how the system behaves during DHW, space heating, standby and defrost.
Troubleshooting: readings that do not make sense
| Symptom | Possible reason | Safe check |
|---|---|---|
| Negative ΔT during steady heating | Sensors reversed, defrost or wrong pipe boundary | Check operating mode and physical labels |
| Flow and return identical for hours | Stale data, poor sensor contact or no circulation | Check timestamps and manufacturer diagnostics |
| Very large ΔT | Low flow, offset error or mixed measurement points | Compare native readings; request hydraulic assessment |
| Temperature spikes | Intermittent sensor connection or sampling artefact | Review logs and wiring condition without opening mains enclosures |
| COP looks implausibly high | Missing electrical auxiliaries or overstated heat output | Audit both meter boundaries and units |
| Surface sensor differs from built-in probe | Different positions, insulation, response time | Compare under stable conditions and document locations |
Buying checklist: select the measurement system, not just the probe
- Confirm whether your heat pump already exports flow and return temperatures through an officially supported integration.
- Decide whether you need diagnostic trends, hydraulic commissioning data or defensible heat-energy measurement.
- For accurate heat output, budget for a compatible heat meter with flow measurement and matched temperature sensors.
- For non-invasive trends, choose appropriately rated probes, proper contact mounting and insulation, with a compatible low-voltage gateway.
- Check installation access, IP rating, cable length, sensor range, sampling frequency and calibration documentation.
- Keep heating plant control and protective functions in the manufacturer-approved controller.
Frequently asked questions
What is the ideal heat pump flow-return temperature difference?
There is no universal value. It depends on the manufacturer, flow rate, operating mode, emitter system and load. Use the installation and commissioning specifications for the exact unit.
Can I calculate COP using only flow and return temperatures?
No. You also need measured water flow to estimate heat output and measured electrical energy over the same period.
Is PT1000 better than DS18B20?
A properly installed, calibrated PT1000 system is often better suited to precision measurement. A DS18B20 can still be useful for inexpensive trend monitoring.
Can I tape a sensor to copper pipe?
A contact sensor can show useful trends when mounted with suitable thermal coupling and insulation, but the accuracy depends heavily on the method and is not automatically equivalent to an immersion probe.
Where do sensors go if I have a buffer tank?
It depends on whether you want heat-pump output or secondary heating-circuit delivery. The two boundaries can differ; measure and label them separately.
Does Home Assistant measure heat output automatically?
No. Home Assistant can display manufacturer estimates or integrate suitable meter readings, but it cannot infer accurate thermal energy from two temperatures alone.
Why is return temperature higher than flow?
This may occur during defrost, cooling, sensor reversal, stale readings or a wrongly defined measurement boundary. Check operating mode before assuming a fault.
Do I need a certified heat meter?
For trustworthy thermal-energy totals or performance verification, a suitable calibrated heat meter is preferable to improvised sensors and assumed flow rates.
Can I use the heat pump built-in sensors?
Yes for many diagnostic tasks, provided the model exposes them. Their accuracy and placement should not be assumed to match a certified meter.
What about glycol in the circuit?
Glycol changes fluid density and specific heat. Use meter settings and calculation coefficients appropriate to the actual mixture.
Can I connect an ESP32 directly to the heat pump temperature sensor?
Do not interfere with the manufacturer’s control sensor circuit. Use independent, appropriately isolated low-voltage probes or documented interfaces.
How often should readings update?
Choose a rate suited to the task and the available interface. Fast sampling does not improve a poorly mounted probe or a slowly updated vendor sensor.
Final recommendation
For most Home Assistant users, begin with native heat-pump telemetry and add independent pipe sensors only when they answer a specific diagnostic question. For credible efficiency calculations, commission a properly matched heat meter and an electricity meter with clearly defined boundaries. Correct placement and measurement provenance matter more than a large number of decimal places.
References and further reading
- Home Assistant Template integration — creating derived diagnostic sensors.
- Home Assistant Modbus integration — supported data acquisition.
- ESPHome MAX31865 RTD component — PT100/PT1000 interface reference.
- ESPHome Dallas temperature sensor — 1-Wire temperature telemetry.
- BIPM Consultative Committee for Mass and Related Quantities — measurement metrology context.
- MAX31865 with ESP32 and ESPHome — relevant sensor project.
- Heat pump monitoring: native vs external meter — electrical measurement boundary.