What hydronic flow rate actually means
A hydronic circuit moves heat by moving mass. Every pound of fluid that passes through a boiler, a coil or a radiant panel picks up or gives up heat in proportion to how far its temperature changes, and the rate at which heat moves is simply the mass flow multiplied by the specific heat multiplied by the temperature change. Flow in gallons per minute is that same statement written in the units a pump is sold in.
Flow and temperature difference trade off against each other exactly. For a fixed load, halving the design temperature difference doubles the required flow. That is the single most consequential decision in a hydronic layout, because pipe size follows flow and pumping power follows roughly the cube of flow in a fixed piping system. A 100,000 BTU/h load at a 20 °F drop needs 10.0 GPM; the same load at a 10 °F drop needs 20.0 GPM, which will not fit in the same pipe at an acceptable velocity.
The number also runs in reverse, and that is how you diagnose an existing system. Measure the supply and return temperatures across a working circuit, know the flow, and you have the load it is actually carrying. Measure supply, return and load, and you have the flow without a flow meter. This calculator does both directions from the same constant, and the HVAC delta-T calculator covers the air-side version of the same idea.
Where the constant 500 comes from, and when it is wrong
Almost every hydronic text writes the relationship as q = 500 × GPM × ΔT. That 500 is not a fundamental constant; it is three unit conversions and two water properties multiplied together, and it stops being right the moment the fluid stops being water.
Build it up. There are 60 minutes in an hour, which turns gallons per minute into gallons per hour. Water weighs 8.33 pounds per gallon near 60 °F — the exact figure is 8.337 lb/gal, from a density of 62.37 lb/ft³ divided by 7.4805 gal/ft³ — which turns gallons per hour into pounds per hour. Water's specific heat is 1.00 Btu per pound per °F, which turns pounds per hour and a temperature difference into BTU per hour. Multiply: 60 × 8.33 × 1.00 = 499.8, rounded in the trade to 500, a difference of 0.04% that no field measurement can resolve.
For a glycol solution, two of those three factors change and they move in opposite directions. Aqueous propylene and ethylene glycol solutions are denser than water, which raises the constant, and lower in specific heat, which lowers it. The specific-heat effect is much larger, so the constant falls and the required flow rises. Entering a specific heat of 0.90 and a specific gravity of 1.04 gives 60 × 8.33 × 1.04 × 0.90 = 467.8, which is 6.4% below water's 499.8 and therefore calls for 6.8% more flow to move the same heat.
Get the properties from the fluid manufacturer at your average circuit temperature, not from a rule of thumb. Both specific heat and density vary with concentration and with temperature, and a 30% mix and a 50% mix are meaningfully different fluids. If all you know is that the system is glycol-protected, a 30–50% solution generally sits somewhere near 0.85–0.93 for specific heat and 1.02–1.06 for specific gravity at hydronic temperatures, and the manufacturer's data sheet will give you the exact pair.
Worked example: a 100,000 BTU/h zone at a 20 °F drop
Take a boiler serving a 100 MBH heating zone designed for 180 °F supply and 160 °F return, so ΔT is 20 °F. The circuit is plain water.
- Build the constant. C = 60 × 8.33 × 1.000 × 1.000 = 499.8 BTU/h per GPM per °F.
- Divide the load by the constant and the temperature difference. GPM = 100,000 ÷ (499.8 × 20) = 100,000 ÷ 9,996 = 10.004 GPM. Using the rounded constant 500 gives exactly 10.00 GPM, so the rounding is worth four thousandths of a gallon a minute.
- Express it per unit of load. 10.004 ÷ 100 MBH = 0.10004 GPM per MBH, which is the familiar rule that 1 GPM carries about 10,000 BTU/h at a 20 °F drop.
- Check against the pump you have. Suppose the circulator on its selected curve delivers only 8 GPM into this circuit's head. Then ΔT = 100,000 ÷ (499.8 × 8) = 25.01 °F. With a 180 °F supply the return comes back at 155 °F rather than 160 °F.
- Convert if you need metric. 10.004 GPM × 0.0630902 = 0.631 L/s.
Now redo step 2 for a 30% propylene glycol fill with a specific heat of 0.90 and a specific gravity of 1.04. The constant becomes 60 × 8.33 × 1.04 × 0.90 = 467.81, and the flow becomes 100,000 ÷ (467.81 × 20) = 10.688 GPM. That is 10.688 ÷ 10.004 − 1 = 6.8% more flow for the same load, and it is exactly the ratio of the two constants, 499.8 ÷ 467.81 = 1.0684.
Choosing the design temperature difference
The temperature difference is a design choice, not a property of the load, and it is where the money is. Pick it first, then let the flow follow.
20 °F is the traditional heating default because it fits copper and steel pipe at reasonable velocities and because cast-iron and fin-tube emitters were catalogued at it. It remains a sound starting point for fin-tube baseboard, where the baseboard heater length calculator shows how output falls along the run as the water cools.
Condensing boilers reward a wider difference. Return water temperature governs whether flue gas condenses at all, and a 30 °F or 40 °F design difference brings the return low enough to condense for more of the season. Wider differences also cut flow, pipe size and pump power. The cost is that emitters at the end of the circuit see cooler fluid and must be sized for it, and low-temperature emitters such as radiant floors are the natural fit — see the radiant floor loop length calculator.
Chilled water conventionally uses 10 °F, which is why 2.4 GPM per ton is quoted so often: 12,000 ÷ (499.8 × 10) = 2.401 GPM. Larger plants push to 12–16 °F to cut pumping energy, at the cost of coils with more rows. The chiller tonnage calculator works the same equation from measured flow and temperatures.
Sanity-check the flow against pipe velocity before accepting it. The usual working range in occupied buildings is 2 to 4 ft/s: below about 2 ft/s air will not be swept out of the pipe, and above about 4 ft/s noise and erosion become concerns in copper. Once the flow is settled, size the circulator against the circuit's pressure drop with the hydronic pump head calculator.
Required water flow in GPM by load and design temperature difference
| Load | ΔT 10 °F | 15 °F | 20 °F | 25 °F | 30 °F | 40 °F |
|---|---|---|---|---|---|---|
| 20 MBH | 4.00 | 2.67 | 2.00 | 1.60 | 1.33 | 1.00 |
| 40 MBH | 8.00 | 5.34 | 4.00 | 3.20 | 2.67 | 2.00 |
| 60 MBH | 12.01 | 8.00 | 6.00 | 4.80 | 4.00 | 3.00 |
| 80 MBH | 16.01 | 10.67 | 8.00 | 6.40 | 5.34 | 4.00 |
| 100 MBH | 20.01 | 13.34 | 10.00 | 8.00 | 6.67 | 5.00 |
| 150 MBH | 30.01 | 20.01 | 15.01 | 12.01 | 10.00 | 7.50 |
| 200 MBH | 40.02 | 26.68 | 20.01 | 16.01 | 13.34 | 10.00 |
Every row scales linearly with load and inversely with ΔT, so a load not shown follows directly from the formula: 125 MBH at 20 °F is 125,000 ÷ (499.8 × 20) = 12.51 GPM.
Glycol changes the pump selection twice over
A glycol fill raises the flow you need because the fluid carries less heat per pound, and it separately raises the head that flow costs because the fluid is more viscous — an effect this calculator does not model. Sizing the circulator on the water flow rate and then adding glycol leaves the circuit short on both counts. Take the corrected flow from here, then apply the manufacturer's viscosity correction to the pressure drop, and check the pump's own derating: many wet-rotor circulators are rated for a maximum glycol concentration. Also confirm the concentration by refractometer rather than by the label on the drum, because a system topped up with water over several years is not the mix it started as.
Mistakes that show up as a system that will not heat
- Sizing flow on boiler input rather than output. An 80% efficient boiler with a 100 MBH input delivers 80 MBH to the water. Use the output, or the flow is 25% high.
- Using water properties on a glycol system. At a specific heat of 0.90 and specific gravity of 1.04 the constant falls from 499.8 to 467.8 and the required flow rises 6.8%. A circuit sized for water runs a wider temperature difference than designed.
- Designing the ΔT without checking the emitters. Fin-tube output is catalogued against average water temperature. Widen the design difference and the average temperature falls, so the same length of baseboard puts out less heat.
- Applying the whole-system flow to a single branch. Each branch carries its own load at its own temperature difference. Size the branch on the branch load, then sum branch flows at the mains.
- Ignoring pipe velocity. A flow that is arithmetically correct can still be unusable. Keep occupied-space velocity roughly between 2 and 4 ft/s and re-choose the design ΔT if the pipe size that results is impractical.
- Forgetting that this equation is sensible heat only. It does not apply to a steam circuit, where the heat moves as latent heat of condensation and the fluid does not change temperature at all.
Key terms
- MBH
- One thousand BTU per hour. A 100 MBH boiler output is 100,000 BTU/h. Not to be confused with MMBH or therms.
- Design ΔT
- The supply-minus-return temperature difference the circuit is intended to run at when the load is at design conditions. At part load the actual difference narrows unless the flow is modulated.
- Specific heat
- Heat needed to raise one pound of fluid by one degree Fahrenheit. Water is 1.00 Btu/lb·°F, which is unusually high, and it is why water is the working fluid of choice.
- Specific gravity
- Density of the fluid divided by the density of water at a reference temperature, here 60 °F. Multiplying by 8.33 lb/gal gives the fluid's own density in pounds per gallon.
- Primary-secondary piping
- An arrangement in which a boiler loop and a distribution loop are hydraulically separated by closely spaced tees, so each can be sized for its own flow without one imposing on the other.
Where this calculation sits in a hydronic design
Flow is the second step of a hydronic design, not the first. The sequence is: establish the load, choose the fluid and the design temperature difference, calculate the flow here, size the pipe to keep velocity in range, add up the pressure drop through pipe, fittings and terminal units, and only then select a circulator that meets that flow at that head.
The load itself has to come from somewhere defensible. For heating that is a room-by-room heat loss calculation and a boiler output; for cooling it is a coil capacity at design entering conditions. Sizing a boiler from a rule of thumb and then computing flow to three decimals from it is precision on top of a guess — the boiler size calculator is the better starting point.
This equation carries only sensible heat, and only for a single-phase liquid. Steam moves almost all its heat as latent heat, so a steam circuit's mass flow follows the latent heat of vaporisation at the operating pressure rather than a temperature difference. Refrigerant circuits are the same story with different numbers, which is why refrigeration capacity is worked from enthalpy difference across the evaporator rather than from a ΔT.
Two effects sit outside this model and both matter in practice. Fluid properties drift with temperature — water's specific heat is not exactly 1.000 at 180 °F, though it is within a percent — and the pressure drop a given flow costs depends strongly on viscosity, which glycol raises sharply at low temperatures. Neither changes the flow you need; both change the pump that will deliver it. Take the flow from here and the head from a pressure-drop calculation done with the same fluid at the same temperature.