Hydraulic Separation: Why Primary and Secondary Flow Change Water Temperature
Separate the pumps without losing track of the heat. See why a 45°C source can supply only 41.7°C water to the load when circuit flows differ.
Published by MEPFlow · Methods and limitations · Primary references
Keep temperatures connected to the water circuit
Review the room duty, water flow and hydronic layout together. Confirm source and emitter performance with manufacturer data and project-specific engineering.
In this guide: jump to a step
A heat source can deliver 45°C water while the heating circuits receive only 41.7°C. There may be no faulty sensor and no missing insulation. If the distribution pump moves more water than the source pump, some cooler return water can mix back into the distribution supply.
Hydraulic separation helps pumps work with less interference from one another. It does not make their flow rates equal, prevent all mixing or create extra heating capacity. A good design checks both the pressure interaction and the temperatures crossing the separator.
This guide follows an original, simplified water example through three operating conditions. Use the diagrams, equations and downloadable checks to understand the energy balance, then verify the actual arrangement against equipment instructions, controls and the applicable local requirements.
For the complete layout, explore MEPFlow’s hydronic design and pipe sizing software and review equipment duty, water flow, pipe sizes and circuit assumptions together.
1. Separate the pressure problem from the temperature problem
Call the circuit through the heat source the primary circuit and the circuit serving the emitters the secondary circuit. Those labels are not universal, so put arrows and equipment names on the drawing rather than relying on terminology alone. In this article, both circuits connect to a correctly arranged four-port separator: hot connections at the top and return connections at the bottom.
The pumps still need to overcome their own circuit resistance. The connection between circuits is deliberately low in pressure loss, reducing the effect that one pump has on the other. Caleffi’s multiple-circulator guidance explains this relationship between common-piping resistance and pump interaction.
Now ask a separate question: where does every litre go? Any difference between primary and secondary flow must pass internally between the supply and return regions. That movement can change a temperature even when the hydraulic separation is working as intended. Check the flow balance before assuming that the source outlet temperature is also the emitter inlet temperature.
Separation is not a universal requirement
Some systems can use a suitable direct-pumped arrangement. Do not add a separator automatically. Establish the source’s allowed flow range, distribution control strategy and manufacturer-approved piping before choosing the architecture.
2. Write down four temperatures and two circuit flows
Use source supply temperature Tps, source return temperature Tpr, load supply temperature Tss and load return temperature Tsr. Use qp for primary flow and qs for secondary flow. Express both flows in the same units. A litres-per-second value cannot be combined directly with a gallons-per-minute value in a mixing equation.
The worked examples assume steady flow of the same water throughout, negligible separator heat loss, no useful stored-energy change and ideal mixing of streams that meet. We approximate water density as 1,000 kg/m³ and specific heat as 4.18 kJ/(kg·K). These convenient teaching values are not exact fluid-property data for every temperature.
With constant density and specific heat, a volume-flow-weighted temperature balance is sufficient. With different fluids, appreciable property differences or unusual conditions, use the appropriate mass and enthalpy balance instead. A hydraulic separator is a shared-fluid connection; it does not isolate glycol from potable water or replace a required heat exchanger.
Heat rate (kW) ≈ 4.18 × flow (L/s) × temperature drop (K)
For these assumed water properties, 1 L/s corresponds to 1 kg/s. Multiply mass flow by specific heat and temperature difference. A difference of 1°C has the same magnitude as a difference of 1 K.
| Input | Value | Meaning |
|---|---|---|
| Source supply, Tps | 45°C | Water entering the separator from the source |
| Load return, Tsr | 35°C | Water entering the separator from the load |
| Water density, ρ | 1,000 kg/m³ | Constant teaching approximation |
| Water specific heat, cp | 4.18 kJ/(kg·K) | Constant teaching approximation |
| Circuit flow | 0.6 or 0.9 L/s | Stated separately for each operating point |
Keep temperatures connected to the water circuit
Review the room duty, water flow and hydronic layout together. Confirm source and emitter performance with manufacturer data and project-specific engineering.
Start free trial3. More secondary flow mixes cooler water into the supply
First set primary flow to 0.6 L/s and secondary flow to 0.9 L/s. The load asks for 0.3 L/s more than the source circuit provides. In the illustrated heating arrangement, that difference recirculates upward inside the separator from the return region to the supply region. It is not extra water entering the closed system.
The secondary supply contains 0.6 L/s at 45°C plus 0.3 L/s at 35°C. Its temperature is (0.6 × 45 + 0.3 × 35) ÷ 0.9 = 41.6667°C, rounded to 41.7°C. The source return remains 35°C under the stated ideal assumptions. This is the mixing condition described in Caleffi idronics 15, pages 40–41.
Check the heat on both sides using unrounded temperatures. The source transfers 4.18 × 0.6 × (45 − 35) = 25.08 kW. The distribution transfers 4.18 × 0.9 × (41.6667 − 35) = 25.08 kW. The higher distribution flow carries the same heat with a smaller water-temperature drop.
That lower supply temperature matters when selecting a radiator, coil or underfloor circuit. Do not size the emitter at 45°C inlet if the design arrangement supplies 41.7°C. The low-temperature radiator guide explains why a catalogue output needs the actual operating-temperature basis.
Original simplified heating schematics. Source supply 45°C; load return 35°C. Arrows show net flow, not pipe size or installation details. On small screens, scroll each diagram horizontally or use the table below.
Load flow greater
Source 0.6 L/s · Load 0.9 L/s
Flows equal
Source 0.6 L/s · Load 0.6 L/s
Source flow greater
Source 0.9 L/s · Load 0.6 L/s
Independent ideal examples, each transferring 25.08 kW. A real separator has finite pressure loss; this drawing omits pumps, valves, expansion provisions, insulation and safety devices.
When qs > qp: Tss = [qp × Tps + (qs − qp) × Tsr] ÷ qs
The formula applies to the stated shared-fluid, ideal heating arrangement. Source return equals load return in this case. It calculates a mixing temperature, not the actual emitter output or the final control response.
4. More primary flow raises the source return temperature
Now use primary flow of 0.9 L/s and secondary flow of 0.6 L/s, keeping the stated source supply and load return temperatures. The source circuit provides 0.3 L/s more than the load circuit takes. This excess moves downward inside the separator and blends with the load return.
The load receives 45°C water. The source return becomes (0.3 × 45 + 0.6 × 35) ÷ 0.9 = 38.3333°C, or 38.3°C rounded. The load-side heat rate is 4.18 × 0.6 × 10 = 25.08 kW. The source-side result is 4.18 × 0.9 × (45 − 38.3333) = 25.08 kW. Energy still balances.
A return sensor at the source therefore need not read the same temperature as a sensor on the distribution return. Avoid interpreting their difference as an unexplained load gain. First compare the flows and the sensor locations. The implications for boiler condensing performance, heat-pump operation or cycling depend on the equipment and its controls; this example does not predict an efficiency change.
When both circuits instead carry 0.6 L/s, the ideal internal transfer between supply and return regions is zero. The load supply is 45°C and the source return is 35°C. Real devices are not perfect mathematical junctions, so installation and operating data still govern.
Source flow held at 0.6 L/s, source supply at 45°C and load return at 35°C. This is a fixed-boundary sensitivity, not a prediction of emitter or control behaviour. Temperature axis starts at 30°C.
When qp > qs: Tpr = [(qp − qs) × Tps + qs × Tsr] ÷ qp
For this arrangement, load supply equals source supply. When qp = qs, the ideal model has no net supply-to-return transfer inside the separator.
| Primary / secondary flow | Internal transfer | Load supply | Source return | Heat rate |
|---|---|---|---|---|
| 0.6 / 0.9 L/s | 0.3 L/s return → supply | 41.7°C | 35.0°C | 25.08 kW |
| 0.6 / 0.6 L/s | Zero net transfer | 45.0°C | 35.0°C | 25.08 kW |
| 0.9 / 0.6 L/s | 0.3 L/s supply → return | 45.0°C | 38.3°C | 25.08 kW |
Do not mistake fixed inputs for a predicted system response
The three cases hold source supply at 45°C and load return at 35°C to isolate mixing. Changing pump speed in a real installation can change emitter output, return temperature, source modulation and controls. Solve or measure those interactions before making a design or operating decision.
5. Check operating combinations before selecting the separator
List the credible operating states: all zones open, a small active zone, source staging, setback recovery and relevant cooling or defrost modes. Record the flow limits specified by the source manufacturer. A design-day balance alone cannot demonstrate acceptable operation when valves close or another pump starts.
For the cited Caleffi Series 548, the selection guidance uses the greater primary or secondary maximum flow, not their sum. In our largest illustrated state, that would be 0.9 L/s, or 3.24 m³/h, as the flow value to take into that product family’s selection process. This is not a final model recommendation.
Check the exact separator’s permitted fluid, working temperature and pressure, connection arrangement, flow rating, maintenance access and installation orientation. An air or dirt removal function is product-specific, not something every low-loss header automatically provides. A compact separator also does not prove that the system has enough active water volume for the source’s minimum run time.
Keep hydraulic separation, buffer volume and fluid isolation as distinct requirements. They can sometimes be combined in selected equipment, but each requirement still needs its own check. Likewise, a separator does not balance unequal emitter circuits. The direct-return and reverse-return guide addresses that separate distribution problem.
1.Define the source envelope
Record minimum and maximum flow, permissible temperatures and manufacturer-approved system arrangements.
2.Define distribution states
Identify simultaneous zones, variable-flow behavior and the expected return-temperature range.
3.Review the connection
Check flow rating, port arrangement, hydraulic resistance, mixing and required service functions.
4.Review controls and volume
Confirm sensor locations, source enable logic, modulation and any minimum active-volume requirement.
6. Commission the temperatures at the right locations
Label all four ports on the as-built schematic. Measure or trend their temperatures and the two circuit flows during an identified operating state. Record which zones and sources are enabled. A snapshot taken while equipment is starting or a valve is moving may not satisfy a steady-state balance.
If the load supply is unexpectedly cool, compare measured secondary flow with primary flow before increasing the source setpoint. Check the required emitter temperatures, source limits, control sequence and whether the actual piping matches the intended arrangement. A higher setpoint may mask a mixing problem rather than resolve the underlying design mismatch.
Use suitable calibrated instruments and account for sensor lag, mounting and insulation. Two nearby readings can differ because of measurement error, while two distant readings can differ because of real heat loss. A source-temperature display alone cannot establish the temperature delivered to every load.
With reliable measurements, calculate heat rates on both circuits. A large mismatch is a prompt to investigate transient storage, heat loss, fluid-property assumptions, measurement quality or unrecorded flow paths. It is not proof of a particular faulty component. Keep safe-access and manufacturer commissioning procedures in force.
| Record | Why it matters |
|---|---|
| Time, active sources and zones | Makes the operating state repeatable |
| Primary and secondary flow | Identifies the possible internal transfer |
| All four port temperatures | Distinguishes supply mixing from return mixing |
| Fluid and property basis | Keeps the heat-rate calculation consistent |
| Source limits and controller targets | Checks that acceptable temperatures also satisfy equipment requirements |
| Instrument method and uncertainty | Prevents over-interpreting small temperature differences |
7. Keep the calculation attached to the design
Download the checked flow and temperature examples and the operating-state review schedule. The example file records fixed teaching inputs, not a system simulation. Complete the review schedule with project-specific equipment limits, measurements and evidence.
Carry the secondary supply temperature into emitter selection, not just the source outlet temperature. Carry the actual circuit flow into pipe and pump checks. Keep the source return-temperature assumption visible when reviewing equipment performance. These handoffs are where a correct individual calculation can become an incorrect system selection.
MEPFlow provides a browser-based workspace for hydronic layout and sizing alongside other mechanical-design work. Use the platform’s documented workflow for the tasks it supports, and retain the manufacturer selections, temperature-balance checks and commissioning requirements with the project.
Frequently asked questions
Does hydraulic separation mean primary and secondary flows must be equal?
No. Reducing pump interaction does not require equal circuit flows. Unequal flows create an internal transfer that can mix either the distribution supply or the source return in the illustrated four-port arrangement.
Why is my load supply cooler than the heat-source outlet?
One possible cause is secondary flow greater than primary flow. Cooler return water can then recirculate into the load supply. Confirm the piping, flows, sensor locations and operating state before diagnosing the installation.
Will increasing the distribution pump speed increase available heat?
Not by itself. Available heat depends on the source, emitters and operating temperatures. Higher secondary flow can increase return-water mixing when primary flow does not rise. The worked example isolates this balance; it is not a pump-control recommendation.
Should I add the primary and secondary flows to size the separator?
Not for the cited Series 548 guidance, which uses the larger of the two maximum circuit flows. Always check the exact manufacturer’s selection rules and all other pressure, temperature, fluid and installation limits.
Is a hydraulic separator the same as a buffer tank?
No. Hydraulic separation concerns interaction between circuits. Buffer sizing concerns usable thermal storage and source operation. Some arrangements provide both functions, but sufficient storage volume must be demonstrated separately.
Can I use these equations for glycol?
The water example’s 4.18 heat-rate factor is not a glycol value. Use the mixture’s properties at the applicable concentration and temperature. For mixing, use a suitable mass and enthalpy balance when constant-property assumptions are not adequate.
Continue the engineering workflow
Keep temperatures connected to the water circuit
Review the room duty, water flow and hydronic layout together. Confirm source and emitter performance with manufacturer data and project-specific engineering.
Start free trialPrimary references
Sources and further reading for the methods discussed in this guide.